ANCIENT CEYLON

'f !?ANCIENT CEYLON

An Account of the Aborigines

and of Part of the Early

Civilisation

By

H. PARKER ,

. ? ,;'?/.,,

Late of the Irrigation Department, Ceylon ' ^ ^ ., ; "/^Vy

WITH ILLUSTRATIONS BY THE AUTHOR

LONDON

LUZAC & CO.

PUBLISHERS xo THE INDIA OFFICE

1909 [AH

"' ,*(

, " '.. '\"BUTLER & TANNER, THE &EI,WOO0 PimrOTG WO

FROME, AND LONDON.PREFAC'E

IN the -last thirty years our acquaintance with the interior of Ceylon, a country four-fifths of the size of Ireland, has made great advances. The researches of members of various Government Departments have extended throughout the whole island, until it may now be said that there is no part of it which has not been investigated.

During this period, however, little new information regarding it has been published in England otherwise than in the Journals of various Societies, with the exception of some excellent studies of its natural history; a work by Professor Rhys Davids on the Ancient Coins and Measures; and two books prepared for the Government, one by Mr. Smither, the former Government Architect, containing an architectural description of the dagabas at Anuradhapura, and the otbef^ by Dr. Edward Miiller, giving a first account of the aftc$ieft&' inscriptions. "?'',; -/,

Evidently the time has arrived when part of the 0tlfe0f, recently obtained knowledge of the country should be1 presented to the world. My employment in the Irrigation Department from the middle of 1873 to the end of 1904 having given me opportunities of acquiring some information of the interior of the island, I have therefore prepared the present work, which describes some phases of the early civilisation, beginning with the history, life, and religion of the aborigines, and ending, as regards local matters, with the village games. Although the subjects included in it are dealt with in a disconnected manner, it will be seen that they advance from the primitive stages to more recent times.

The character of such a work must naturally render it more useful to students of the subjects treated of than attractive to the public. For reason it has been my en-VI

PREFACE

deavour as far as possible to furnish accurate and detailed information rather than generalities among which the student might search in vain for the particulars he requires. I may be permitted to express a hope that my critics will deal leniently with the errors which must be inseparable from such a publication.

In transliterations I have followed Dr. E. Muller in indicating by & the vowel which appears as e in publications of the Ceylon Government. The form accepted by me, when pronounced as a diphthong as in the Oxford Dictionary, both gives the sound of the letter and is historically accurate, the letter having been in most cases derived from an ancient a.

The consonant which is often expressed by v has been represented by either v or w, so as to be in general agreement with its local sound. In Ceylon it is a w, and any one who pronounced it otherwise in nearly all words would make himself ridiculous. In the case of Pali words, especially the names of places and books, I have used only the letter v, in order to avoid confusion through being in disagreement with other works. I adhere in general to the Pali forms of names.

1 have to express my obligations to the Secretary of State for the Colonies for his readily granted authorisation to reproduce some of Mr. Smither's drawings of the dagabas ; and to my friends Mr. H. T. S. Ward, the recent Director of Irrigation, and Mr. H. C. P. Bell, the Archaeological Commissioner, the former for permission to copy and utilise the drawings of irrigation works in his office, and the latter for allowing me to include in this worfc a description of some early coins in the possession of the Ceylon Government, without which the account of the first local coinage would have been incomplete.

In the various chapters in which it has been utilised I

have acknowledged the information furnished by several

kind friends in Ceylon, and by Mr. C. H. Read of the British

and Dr. C. G. Seligmann, to all of whom it is a

return to my grateful thanks.

Mtssr*. H. B. Andris and Co. of Kandy were so good as to bring about the publication of a Sinhalese work on the KG-PREFACE vil

hoxnba Yaka in order that it might be available for me, and to the kindness of Mr. H. W. Codrington, of the Civil Service, I am indebted for native accounts of this deity compiled in various provinces. To my friend the late Dr. Paul Gold-schmidt I owe my interest in the early inscriptions.

With regard to the scales of the drawings, which are usually expressed in fractions, the denominator divided by twelve gives the number of feet equal to one inch.

Through an inadvertence the word Vyddha appears in some places as Vy&da.CONTENTS

PAGE

PREFACE .......... v

LIST OF ILLUSTRATIONS ....... xi

PART I. THE ABORIGINES ........ i

PART II. STRUCTURAL WORKS ...... 207

PART III. ARTS, IMPLEMENTS AND GAMES .... 413.

THE ORIGIN AND SIGNIFICATION OF THE CROSS %

AND SWASTIKA ...... 643.

ADDENDA ......... 667

APPENDIX. TABLE OF MEASURED BRICKS .... f36g

INDEX ........... 675

JxLIST OF ILLUSTRATIONS AND PLATES

PAGE

The Swastika of Ceylon . Cover

2. The God of the Rock .... 'Frontispiece'

3. Map of Ceylon ..... Facing i

4. Rakshasa as Guardian .... ,, 5.

5. Vibhisana, his wife, and Lakshmana . ,, 7

6. A Modern Rakshasa ...... 7

7. Ritigala ........ 8

8. Nagas as Guardians .... Facing 15,

9. A Yaksha . . . . . . .17

10-19. Weapons and Utensils of Vaeddas . , . . 55,

20-34. Stone Implements . . . . 65

35. Skanda and Valliyamma . . . Facing 115

36. Mohim ....... ,, 136

37. Ayiyanar as Guardian . . . . ,, 148

38. Ganesa, Vibhisana, and Pattini . . ,, 148

39. Ayiyanar on his Elephant . . ,, 148

40. Kokka-gala . . . . . . . .178

41. Vaedda Temple of the Gale-Yaka . . . . 182.

42. Rock Temple of the Gale-Deviya . . . .183

43. The Nirammulla Dewalaya . . . . .185,

44-50. Dancing Rocks of the Gale-Deviya . . 193

51-53. Ancient Utensils of the Gale-Deviya . . .199

54-56. Small Cup-holes ....... 223,

57-62. Large Cup-holes . . . . . . .227

63. The Giant's Tank ....... 248

64. The Thupararna Dagaba.....263

65-69. The Thuparama Dagaba. Elevation and Details . 267

70. Map of Anuradhapura and its Tanks . . .27!

71. The Maha Sagya, Mihintale . . 276-xii LIST OF ILLUSTRATIONS AND PLATES

FIG. PAGE

72. King Duttha-Gamini .... Facing 279

73. Tlie Ruwanwaeli Dagaba . . . . .281 74-79. The Ruwanwaeli Dagaba. Plan and Details . . 287 80-83. The Ruwanwaeli Dagaba. Restored Elevation . 292

84. Southern Wahalkada, Miriswaeti Dagaba Facing 295

85. The Abhaya-giri Dagaba. Elevation . . . 306

86. The Jetavana Dagaba ...... 308

87. Pillars at Wahalkada, Jetavana Dagaba Facing 310

88. The Mahiyangana Dagaba . . . . .316

89. The Kaelaniya Dagaba . . . . . 317

90. The Idikatu Dagaba . . . . . . 319

91. The Ambatthala Dagaba . . . . -321

92. The Mahanaga Dagaba ...... 325

93-110. Articles deposited in Dagabas .... 329

in. Statue of Prince Sali .... Facing 333

112. Miniature Stone Dagaba . . . ,,341

113. The Ottappuwa Dagaba . . . . .344

114-115. Panda-waewa ....... 354

116. Basawak-kulam ....... 360

117. Pool in Tissa-waewa ...... 363

118-119. Vavunik-kulam ....... 368

120-125. Pavat-kulam .... . . .372

126. Pavat-kulam. Northern Bis5kotuwa . . . 375

127-129. Sangili-kanadara Tank . . . . . .384

130. Map of Tissa and its Tanks . . . . -387

131. Tissa-waewa ........ 388

132. Destruction of a Dam ...... 391

w

133. Direct and Oblique Dams ..... 391

134. Batalagoda Tank. Plan ..... 398

135. Batalagoda Tank . . . . . . . 399

136. Nuwara-waewa. Plan and Section of L. L. Sluice 402

137. Nuwara-waewa. Plan and Section of H. L. Sluice 402

138. Basawak-kulam. Section of Bank .... 402

139. Tissa-waewa. Section of Bank .... 402

140. Nuwara-waewa. Remains of Dam.... 406

141. Nuwara-waewa. Bridge over Channel ,. . . 406LIST OF ILLUSTRATIONS AND PLATES riii

FIG. 142-146,

147-

148. 149. 150,

151-

152.

*53. 154. 155. 156-

157-158.

159.

160.

161-169.

170-171.

172-178.

179. 180. 181. 182.

183.

184.

185-200.

201-216.

217,

218,

219.

220-239.

240-241.

242-243,

244.

245.

246.

247,

Naccaduwa Tank ......

The Allekattu Dam

Tevandan Puliyan-kulam Hill

Tevandan Puliyan-kulam. Inscribed Boulders.

PAGE 406

4II

417 417

The Earliest Inscription (No. 2) . . Facing 418 Facsimiles of Inscriptions . . . . .421

Cave Temple, Kaccatkodi . . . . * 435

Facsimiles of Inscriptions . . . « . 447

Mulleittlvu and Tissa Coins . . . Facing 469

Anuradhapura Coins .... ? 482

Seal from Yatthala Dagaba ... ? 495

Guard-Stone, Anuradhapura ... ? 499

Durga, as Kali, destroying the Asuras . ? 501

Relief of Building. Anuradhapura. . . . 508

Bhairava as Guardian . Facing 515

Swords and Clubs. . . . * . . 527

Soldiers in Panels, Ridi Wihara . . . . 529

Kandian Knives ....... 533

.Dagger. ... ... . . . . 533

Shield engraved on Rock ..... 533

Dagger engraved on Rock ..... 533

Knife engraved on Rock . .... 533

Waved Spear Head . . . . . 533

Sinhalese Marteau . . . , '533

Sinhalese Weapons ,,,... 535

Sinhalese Weapons .... , ? 537

Side of Duttha-Gamini's Crown . . . . 538

Pillar at the Giant's Tank . * . . . 539

Rock Carving at Isurumuniya . . . . 545

Sinhalese Tools ..... . . 553

The Pump Drill....... 550

The Spinning-Wheel «'. . . . . . 563

The Cotton Gin . , , ? . ... 564

Nerenchi Diagram . . .. . . 577

Indian Diviyan-keliya Diagram . . . . . 581

Diagram for H£w£kam Diviyan-keliya ? - 582XIV

LIST OF ILLUSTRATIONS AND PLATES

FIG.

248-255.

256.

257-259.

260.

261-262.

263.

264-269.

270.

271.

272.

273.

274,

275.

Olinda (Mancala) Boards . .. Mancaia Holes at Third Pyramid, Gizeh Saturankarn Diagram .... Siga Diagram .... . Pancha-keliya Diagram .... Asi-keliya Diagram . . . . Pachis Diagram . . . . . Arasadi Diagram . . . . . Tattu-keliya Diagram . . .

Pattini and her Husband . . .

. . 588

Facing 590

.. 605

. . 607

. . 609 . -615.

. . 619

. 625

. . 627

Facing 631

Masons' Diagrams on Roof of Kurna Temple . . 644 Flower Altar near Vammiyadi . . . .658 Yantra-gala, Anuradhapura ..... 658CEYLON

\Tofette p. i.

Part I THE ABORIGINES

PAGE

I. THE FIRST INHABITANTS . .... 3 II. THE ANCIENT VAEDDAS...... 23

III, THE MODERN VAEDDAS . . . . . - , 35

THE MODERN VAEDDAS. SOCIAL DIVISIONS AND ? CUSTOMS,.......:;; .. -'J$3.-.

IV. THE RELIGION OF THE VAEDDAS . ? - * ^

V. THE PRIMITIVE DEITY OF CEYLON . / ,;. -'/:"-'vijrtr'"

[To /« ;« £./f^C^/lA/plA^l

^Q^JjrO^^ XW x$

^oL*sJZa£li\

I &?JLJ^. X1?,^/.?

THE FIRST INHABITANTS

T Tf 7"HEN the first Aryan invaders entered India they VV brought with them an exaggerated belief in the existence of various classes of evil beings, among whom those termed Rakshasas occupied the most prominent place. These demons were thought to be especially active and powerful during the darkness of the night, when, though invisible in their true shapes, they acted in many objectionable ways in opposition to the new settlers ; and most of the ills which beset the Aryans were attributed to their malevolence. Every mysterious sound heard during the night, and especially the weird calls of the forest owls, showed them to be then in the immediate neighbourhood of the villages or encampments, but with the first gleams of sunrise they vanished ; the spear-like rays of the mighty Sun-god had annihilated them, or at the least had driven them away into the obscurity of the trackless forests. Being thus powerful during the nocturnal hours, it was naturally believed to be they who inspired the night attacks of the aboriginal tribes, the constant enemies of the Aryan settlers ; and many and fervent were the prayers addressed to Agni, the Fire-god, and Incfra, the God of the Firmament, the Lord of the Thunder ^nd the Controller of the Heavenly Fires, to arise and disperse and overwhelm them* In the fourth hymn of Book iv of the Rig-Veda (Griffiths' translation) the prayer runs :?

Rise, Agni, drive off those who fight against us: make manifest thine

own celestial vigour.

Slacken the strong bows of the demon-driven. . . . Destroy the cursing R&kshasas.

As the Aryans advanced further into the country their belief in the existence of these demons of the night remained4 ANCIENT CEYLON

firmly impressed on their minds. They afflicted both man and beast, and were devourers of raw flesh. Sometimes they appeared bodily?not in their true forms, but in the shape of dogs, owls, and other birds?and obstructed the sacrifices of the Aryans in various ways, and especially by the pollution of their presence. In the hymn 104 of Book vii, the Maruts? the Gods of the Storm-winds?and Indra are appealed to :?

She, too, who wanders like an owl at night-time, hiding her body in

her guilt and malice, May she fall downward into endless caverns. May press-stones with

loud ring destroy the demons. Spread out, ye Maruts, search among the people: seize ye and grind

the Rakshasas to pieces, Who fly abroad, transformed to birds at night-time, or sully ancl

pollute our holy worship. Indra hath ever been the fiends' destroyer who spoil oblations of the

Gods* invokers: Yea, Sakra, like an axe that splits the timber, attacks and smashes

them like earthen vessels. Destroy the fiend shaped like an owl or owlet, destroy him in the form

of dog or cuckoo. Destroy him shaped as eagle or as vulture: as with a stone, O Indra,,

crush the demon.

They were considered to be especially-malignant sorcerers.

The same hymn continues : * Slay the male demon, Indra f Slay the female, joying and triumphing in arts and magic/ It concludes with the prayer, f Indra and Soma, watch ye

well Cast forth your weapon at the fiends : against the sorcerers hurl your bolt/

The hymn 87 of Book x is entirely devoted to denunciations of these demons, and appeals to Agni to destroy them :?

Where now thoii seest, Agni Jatavedas, one of these demons standing

still or roaming. Or flying on those paths in air's mid-region, sharpen the shaft and as

an archer pierce Mm. The fiend who smears himsell with flesh of cattle, with flesh of horses

and of human bodies, Who steak the milch-cow's milk away, O Agni?tear off the heads

of with fiery fury.

Agni, from days of old then slayest demons: never shall Rakshasas

in fight o'ercome thea Bum up the foolish ones, the flesh devourers ; let none of them escape

thine heavenly arrow.

I;IG. 4. Rakshasa as Guardian, Tanjore Temple.,

To /i'ti' p. 5-..THE RAKSHASAS 5

In the Sanaa-Veda (Stevenson's translation) the Rakshasas are said to be indomitable (Adhyaya xii, 2), and to be all around (Prapathaka vi, 6).

In the hymns of the Atharva-Veda (Bloomfield) we learn that the Rakshasas robbed people of their senses (vi, 3), and * possessed ' them (ii, 9), and that errors made in the prescribed ritual of the sacrifice were also sometimes due to their malicious interference (vii, 70). They were unable to face Indra ; * Indra forced the demons into the nethermost darkness ' (ix, 2).

Such were some of the earliest ideas of the Aryans concerning the Rakshasas, in the second or third millenium before Christ. In the first half of the pre-Christian millenium, the Ordinances of Manu confirm the statement that the Rakshasas were flesh-eating demons, and that night was the special time of demons' activity; they also place them in a position of high respectability after the Gods and Manes, along with other classes of supernatural beings. In the Sutta-Nipata (Fausboll's translation, S.B.E., p. 51) we find the Rakshasas uniting with the Gods in reprobating the slaughter of cows.

When the Indian epic poem, the Ramayana, was composed, the Rakshasas had developed into beings who constantly made their appearance before men, in their own or other forms which they took at will They were first described as wandering malignant demons of the great Vindhya forest, which extended far to the south in India; and afterwards, in the later portions of that work, they were represented as occupying all Ceylon, then (and still) denominated Lanka, under the rule of their own king, Ravana. The Maha-Bharata has the same tradition.

The latest account of them in these works is as follows l: When Brahma created the Waters he formed Rakshasas to guard them.2 Visvakarman, the general architect and builder of the Gods, erected a city termed Lankapura for them in Ceylon, on the top of the mountain Trikuta, * Three Peaks/ on the shore of the southern ocean. Three of their princes

1 Main Original Sanskrit Textst Vol. iv, pp. 414 ff.

2 Two Rakshasas are carved in relief as guards In Fig, 159. I know of no other representation of them In the Sinhalese carvings.6 ANCIENT CEYLON

performed intense austerities for which they were rewarded by the grant of long life and a certain amount of invincibility. They made use of these gifts to oppress the Gods and sages, and at last prepared to attack heaven itself. The Sama-Veda mentions another Rakshasa called Kravi, who had previously got heaven and earth into his power and desolated them (Adhyaya xiii, 8). They were defeated by Vishnu, and driven back to Ceylon, and afterwards to the underworld, Patala, as stated also in the Atharva-Veda, where the deed is attributed to Indra (see above).

Kuvera, the God of Wealth, with his attendants the Yak-shas, who were demons of another type, in some respects not much better than the Rakshasas, but of a higher rank, then took up his residence in Ceylon, at Lankapura. Eventually, his half-brother Ravana, the Rakshasa king, by means of thousands of years of austerity obtained from Brahma the boon of indestructibility by all beings of a higher class than man. This enabled him to re-occupy Ceylon, which once more became the headquarters of the Rakshasas. He also conquered Kuvera, whose magic car he took, Yama, the God of Death, and Indra, and generally made the lives of the Gods extremely unpleasant. ' The Gods then addressed a word to Brahma, the Creator of the world: " A Rakshasa named Ravana having obtained a boon from thee, 0 Brahma, in his pride harasses us all. Obedient to thy words, we endure everything at his hands. . . . We are therefore in great fear of this Rakshasa of horrible aspect" * (Muir, O.S.T. iv, p. 140).

The Ramayana recounts at great length how these truculent demons interfered with or polluted the sacrifices of the anchorites in the Vindhya forest, and even devoured those holy men. The situation was evidently insupportable. In the meantime, the Gods had a' rod in pickle for the demons. Vishnu, the younger brother of Indra,- had acceded to the unanimous request of the- deities, .and'become., .incarnate as Rama, the son of Dasaratha, the king of Ayodhya or Oudh. Rama, who was suitably provided with .magic weapons, first destroyed the Rakshasas in India,on.account of their crimesTHE RAKSHASAS

there; and then, assisted by an immense army of monkeys and bears, proceeded to attack and kill Ravana in Ceylon, after the demon king had carried off his wife Sita to Lanka-pura. He then returned to Oudh with Sita. According to the Rajavaliya, one of the Sinhalese historical works, the date of this event was 1844 years before Gotama Buddha entered on his mission, that is, about 2370 B.C.

Although he had promised to do it, Rama did not exterminate the Rakshasas. Vibhisana, the younger brother of Ravana, a good and devout worshipper of Vishnu, who had joined Rama's forces in the war against the Rakshasas, was appointed the sovereign of the survivors in Ceylon, in the place of Ravana ; and there the story ends so far as it concerns Ceylon. The Rakshasas also vanish from history, with the exception of an occasional appearance of a fever- or ophthalmia-causing demon who is termed a Rakshasa in the Sinhalese chronicles. They are found, however, in early times and down to the present day in the folk-stories of the villagers, both in India and Ceylon. In Ceylon they have degenerated into mere man-eating ogres of the European Jack - and - the - Beanstalk type/ who are much more powerful than the Yakshas?according to one story four Yakshas took to

FIG. 6. A Modern Rakshasa. (After a Native Fainting.)

1 The reader may remember the striking description of one in the Third Voyage of Sindbad the Seaman :?-' A huge creature in the likeness of a man, black of colour, tall and big of bulk, as he were a great date-tree, with eyes like coals of fire and eye-teeth like boar's tusks and a vast big gape like the month of a well. Moreover, he had long loose lips like camels", hanging down upon his breast, and ears like two Janus [a kind of barge] falling over his shoulder-blades and the nails of his hands were like the claws of a lion/1?-4ralzaw Xights, Lady Burton's Ed , Ml, p. 485.

8

ANCIENT CEYLON

flight when opposed by one Rakshasa?but are outwitted by clever girls and men. The Rig-Veda had already termed them foolish.

Although there is nothing in this legend of the Ramayana to indicate that the composer of even the last section possessed more than the slightest knowledge of Ceylon, most of the geographical outlines referring to the island are accurately pour-trayed. He knew that Ceylon was an island near the southern coast of India, and tied to it, as it were, by a chain of islands or sandbanks. He was aware that the country was

FIG. 7. Ritigala, from the South.

about 100 leagues in length?the actual distance is about 266 miles 1?and that there are mountains in the -southern part of it. He had also learnt that on the side of the ancient highway leading from the end of Mannar to the southern districts, the traveller passed a Mil termed Arishtha, the Arittha of the Pali histories of Ceylon, now called Ritigala, near the foot of which the high road certainly ran in historic times. The name Suvela, which is also mentioned as that of a hill, cannot be identified as such, but may be a reference to the land round the town called Uruvela, In the northern

, l The earliest Sinhalese history, the Dipavansa, p. 196, says that it is 32 yojanas ; at 8J- miles per ydjana this is 272 miles.THE RAKSHASAS 9

part of the Kandian hill-country there are also three very conspicuous peaks on one of the higher mountains, when viewed from the northern low country, from which the idea of the mountain Trikiita may have been derived.

It is evident that before this knowledge of the interior of Ceylon could be available in India, the island must have been thoroughly explored by intelligent travellers. This could only be done in a settled and peaceable country such as we find under the Sinhalese kings, and there is no probability that it was ever feasible at an earlier period. As European scholars now agree, the whole account of the invasion of Ceylon by Rama must therefore have been invented during historic times, and it thus becomes simply and -purely a poetic fiction, an improvement of the original story without any basis whatever in fact. Even such a slight foundation for it as the spread of the Hindu religion, or Aryan civilisation, among the tribes of the south must be swept away so far as Ceylon is concerned, since the descendants of the original inhabitants of the island, the Vaeddas of the interior, have never adopted the worship of the Hindu gods, nor, until historic times, the civilisation of the Aryans.

We now come to the Sinhalese annals, and here we.soon begin to feel our feet on firmer ground.1 Of these histories, the two most important ones are written in the Pali language ?the Dipavansa and the Mahavansa. The former, which ends with the death of King Maha-Sena (277-304 A.D.), and appears to have been completed not later than the beginning of the fifth century A.D., and possibly nearly a century earlier, is believed by its translator, Dr. H. Oldenberg, to consist chiefly of extracts from histories or chronicles of much earlier date.

The Mahavansa was written at various times, and has been continued to the end of the eighteenth century. My references will be to the English translation made for the Ceylon ?Government by the late Mr. L. C. Wijesinhe. There is no doubt that the author of the first part of it was a Buddhist

1 The Rt. Rev. Dr. Caldwell termed the writers * on the whole, the most truthful and accurate of oriental annalists.* (Dravzdtan ?mar, Introduction, p. 121.)A

10 ANCIENT CEYLON

monk who bore the title Mahanama, and was the uncle of King Dhatu-Sena (463-479 A.D.) ; and that most probably soon after the death of that king he completed the book up to his own day. It is recorded in the Tika, or ' Commentary' on the Mahavansa, a work of somewhat later date, that he derived his materials from Chronicles written long before in Sinhalese, one of which owed its authorship to the monks of the Uttara Piriwena (the Northern Monastic residence) at the Maha Wihara, the great Buddhist temple founded at Anuradha-pura in the middle of the third century B.C.

It is expressly mentioned that several histories were extant in his time, and were consulted by him. , Some of them were also termed Mahavansas. In the Commentary it is stated.: ' Thus the title "Mahavansa" is adopted in imitation of the history composed by the fraternity of the Maha Wihara. . . . In case it should be asked in this particular place, " Why, while there are Mahavansas composed by ancient authors in the Sinhalese language, this author has written/" etc.1 Mahanama himself insists on the accuracy with which he adheres to the accounts of the early chroniclers. At the beginning, he states : ' Having bowed down to the supreme Buddha, immaculate in purity, illustrious in descent; without suppression or exaggeration I celebrate the Mahavansa.5 It can hardly be doubted, from the amount and accuracy of the details which Mahanama gives in his work, that at least one of these prior Chronicles was begun in the third century B.C., and certainly not later than the second century B.C.

It is important to understand clearly that as regards the pre-Christian and early post-Christian details which are found in the Mahavansa we have got, not the opinions or fancies of a monk who lived 500 years after Christ, but a work carefully compiled from annals that were committed to writing in the second or third century before Christ, and continued without a break up to the time of the reverend author. With respect to the information to be collected from the work regarding the earliest rulers, we have at least the opinions of

1 Turnout. The-Makawan$Q Introduction, pp, xxxi, xxxILTHE YAKKHAS n

annalists, or traditions recorded by them, dating from a time that was perhaps only a century and a half later than the earliest local events of which they preserved the story. Some of these early chroniclers may have seen, or have known persons who had seen, the great king Pandukabhaya, the record of whose reign is of the utmost value for the light it throws on the position occupied by the aborigines in the third and fourth centuries before Christ.

There are other historical works of subsequent date, nearly all written in the Sinhalese language. Occasionally they contain supplementary details of the early period which are not found in these two first books, thus showing that their composers had also access to some manuscripts that are now lost. Among such works may be noted the RajavaJiya, the Rajaratnakara, the PujavaJiya, the Thupavansa, and the Dhatuvansaya.

It has been already mentioned that the later parts of the Ramayana and the Maha-Bharata contain the statement that Ceylon was once occupied by a class of beings termed Yakshas, under their sovereign Kirvera or Vaisravana, the God of Wealth, the Wessawana of the Sinhalese. The Ramayana also incidentally adds that some Yakshas dwelt on the Arishtha Mil at the period of the mythical invasion by Rama, and on the mountain Mahendra?at the southern end of the Vindhya chain, the Western Ghats?on the opposite coast of India, It is possible that the person who composed that part of the epic had heard of the stories related by Indian traders regarding the first settlement of the Sinhalese in Ceylon.

Apparently, at the time when the first Magadhese traders 1 came to Ceylon from the lower part of the Ganges valley, they described the inhabitants whom they found occupying the central and southern forests as beings who were scarcely

1 The way of the tradesman [is the occupation] * of M&gadhas. Ordinances of Manu, Translation by Bumell and Hopkins, xf 47. The translators state that the Commentator Medh£tithi specifies * the way * as referring to both land and water.

* Throughout work, the words In square brackets are Inserted by me.12 ANCIENT CEYLON

*

human, a custom of many later travellers when delineating aborigines. They may have exaggerated and embellished their accounts of them with a view to deterring others from venturing into Ceylon, so as to enable them to retain a lucrative trade in their own hands. However this may be, the chronicles of their descendants, the Sinhalese, applied the Pali term Yakkha, ' demon/ to the beings whom they found in the island, but described them as devoid of most of the supernatural attributes of the Yakshas of the early Indian works. They were no longer beings of a semi-divine nature, but were looked down upon as approaching much more nearly to the class of evil demons, just as the references to the aboriginal Dasyu of Vedic times are often couched in terms "that might equally describe the characteristics of demons. They no longer possessed the power of aerial flight and of passing through the water.

The historical works of Ceylon contain a mythical story of three visits that were supposed to have been paid to the island by the last Buddha, Gdtama, as well as by the three previous Bttddhas. It is not found in the canonical works, and is therefore not accepted by the more intelligent Buddhists in the island, whether monks or laymen; but it is credited as an article of faith by the less-instructed classes, and it has had the effect of greatly enhancing the prestige of the Buddhist remains at Anuradhapura and Kaelaniya, the sites of two of the supposed visits.

In them an account is related of the miraculous expulsion of the Yakkhas from the island at the last Buddha's first visit, in the ninth month after he attained Buddhahood, in order to render it habitable by the Gangetic settlers who were about to occupy it after his death. The Dipavansa gives the story as follows (i, pp. 46 ff., Oldenberg's translation) : * At that time the ground of Lanka was covered with great forests, and full of horrors: frightful, cruel, blood-thirsty Yakkhas of various kinds, and savage, furious, and pernicious Pisachas [a lower form of demon] of various shapes and full of various (wicked) thoughts, had all assembled together. [The Teacher thought] " I shall go there, in their midst; I shall dispel theTHE YAKKHAS 13

Rakkhasas and put away the Pisachas ; men shall be masters (of the island)."

He came through the air from the Andtatta Lake in the Himalayas, and alighted at Mahiyangana, on the eastern side of the Central mountains. There he first sent down ' rain, cold winds, and darkness/ and afterwards intense heat, to escape from which the unfortunate Yakkhas could merely stand on the shore.

In the end he permitted the Yakkhas and Rakshasas (who are suddenly introduced into the story) to escape to an island called Giridipa, ' the Island of Hills/ a name which may possibly indicate Malayalam, ' the Mountain Region/ The Raja-valiya terms the place Yak-giri-duwa, * the Island of Demon Hills/ This place is described as ' beautifully adorned by rivers, mountains, and lakes . . . full of excellent food and rich grain, with a well-tempered climate, a green, grassy land . . . adorned by gardens and forests; there were trees full of blossoms and fruits/ It was situated * in the great sea, in the midst of the ocean and the deep waters, where the waves incessantly break; around it there was a chain of mountains, towering, difficult to pass/

The second visit of the Buddha is stated to have been paid in the fifth year of his mission. In this case he visited the Nagas, a class of beings entirely different from the Yakkhas, who were engaged in a civil war in Northern Ceylon.1 He first cowed them in the manner which had proved so effective with the Yakkhas, by means of a * deep terrifying darkness/ and then reconciled them and converted great numbers to Buddhism. On this occasion he was accompanied by Indra as his attendant, who brought with him a large KiripaJu tree (Buchanania angustifolia) in which he commonly resided, and held it as a sunshade over his illustrious1 master, finally planting it in northern Ceylon as an object for the Nagas to worship. The third visit was made in his eighth year. On the full-rnoon day of Wesak (April-May), accompanied by 500

1 The Raj avaliya fixes the incident at Kaelaniya, and states that lie then remained three days in Ceylon. It omits his visit to that place on the occasion of his third journey.14 ANCIENT CEYLON

monks, he is represented as going to Kaelaniya, on the western side of Ceylon, near Colombo, at the invitation of Mani-Akkhika the Naga king of Kaelaniya, who had undertaken a journey to India in order to invite him to come. Mani-Akkhika, who is stated in the Dhatuvansa to have been the maternal uncle or father-in-law of Mahodara, one of the kings who was at war on his former visit, is described as a devout Buddhist, having been converted at the Buddha's first visit to the Yak-khas. The Naga king erected a highly-decorated pavilion .for the reception of the distinguished visitors, and distributed a great donation to the monks. After this, the Buddha is believed to have first left the impression of his foot on the Sumana Kuta mountain (Adam's Peak), and to have afterwards proceeded to the site of the future Dighavapi, on the eastern side of Ceylon, and finally to Anuradhapura, where he visited the sites subsequently occupied by the celebrated Bo-tree and three dagabas.

According to these accounts, the Nagas were apparently considered to be a comparatively civilised race. The incident of the planting of the Rajayatana (Kiripalum Sinhalese) tree of Indra in their country Nagadipa,' the Island of the Nagas/ plainly shows that they belonged to the older faith of India, and were worshippers of Indra, and not of Siva. They were ruled by their own kings, and had a settled and regular form of government. They seem to have been confined to the western and especially the northern part of Ceylon, this latter tract being invariably referred to in the histories for many centuries as Nagadlpa. In these works the expression * island ' is often applied to a tract of land only partly surrounded or bordered by water. Similarly, in the Sinhalese histories India is always known as Jambudwlpa or Dambadiva, * the Island of Jambu (trees)/

Nagas are generally understood to be a form of nondescript beings with the bodies of serpents attached to the upper parts of human beings; but they are never represented in this manner in Sinhalese carvings, nor at Bharhut and Amaravati in India, In the Bharhut carvings they resemble human beings in all respects, and can be recognised as Nagas onlyTHE NAGAS 15

by the addition of this descriptive title to their names. In the reliefs at Anuradhapura and Amaravati, Naga princes and princesses are only distinguishable from human beings by means of the cobras' heads with outspread hoods which appear behind or at the side of their heads. The Pujavaliya mentions dancers among North Indian Nagas, and refers to the arms of the Naga raja, Aravala. The old notion regarding them appears to have been that they had two forms which they could assume at will?either a human shape or that of a cobra.

Just as the Rakshasas disappear from history after the events described in the Ramayana, so the Nagas of Ceylon are never mentioned again as inhabiting the island after their supposed partial conversion by the last Buddha. Yet the fact that the only name for the northern portion of Ceylon was ' the Island of Nagas/ must be held to prove that some beings designated Nagas once inhabited it.

The word Naga may be applied either to human beings-there are still people of this name In north-eastern India?or cobras, or elephants, or to the class of supernatural beings referred to above, whose home was in the water, or below Mount Meru, the centre of the universe. The latter were especially beings of the water, as the Yakshas were beings of the land. We may venture in these days to leave such creatures out of consideration, and to assume that the early occupiers of Northern Ceylon were human beings, as the account of them in the histories indicates.

The original home of such a race must evidently be looked for in the most southern part of India. In such a case, I think we must naturally turn first to the people of an identical name in Southern India, the Nayars, who still occupy practically the extreme south-west part of the country. Their situation itself renders it In every way likely that Northern and Western Ceylon might be colonised by a branch of this race. There is no direct proof of the occurrence of such an Immigration, but some evidence of It may be found in the fact that it would provide an explanation of the existence among the Kandian Sinhalese, who are a more or less mixed race, of some social features resembling those of the Indian Nayars. Among thesej6 ANCIENT CEYLON

may be especially noted (i) the practice of polyandry ; (2) the elasticity, or rather the slenderness, of the marriage tie, which permits the discarding, without any disgrace being attached to it, of undesirable husbands or wives ; (3) the re-marriage of such wives, and of widows, with others, as a universal national custom; and (4) the absence of * Sati/ or widow immolation. These are all customs that with perhaps the exception of the last, apparently cannot have been brought to Ceylon by the settlers who came from the valley of the Ganges ; but they are still maintained by the Nayars and the Kandian Sinhalese, Neither Sati nor the first three practices are found among the Vaeddas, the wild inhabitants of the inland forest tracts, and the three social customs must therefore have been introduced by others. It would be difficult to account for their presence in Ceylon by any other probable hypothesis than a Nayar connexion of early date, since in historical times there has been no special intercourse between the island and Malayalam, beyond the enlistment of a few mercenary soldiers who were natives of the latter country. I suggest, therefore, that the Nagas who occupied Northern Ceylon long before the arrival of the Gangetic settlers were actual Indian immigrants, and were an offshoot of the Nayars of Southern India.

During the reign of the first king of Ceylon we find a town to the north of Anuradhapura, on the Kadamba river, which may have been then, as it is now, the boundary of the Dra-vidian territory, that is, of Nagadipa, specially referred to by the annalists as the seat of * the Brahmanical Upatissa/ Thus it may possibly have been a town or settlement of early Dravidian colonists.

Returning to the Yakshas, the Yakkhas of the Pali works, who evidently occupied the portion of Ceylon which was not included in Nagadipa, we find that in addition to Mahiyangana, which is stated to have been the scene of one of their battles (Mah. i» p. 4), they are more than once mentioned as being present in north-central Ceylon, They are expressly said to have been numerous * in the south/ where the Indian prince Wijaya, the future ruler of the island, and Ms party from theTHE YAKKHAS 17

Ganges valley are reported to have landed; one of their

capitals, Sirivattha, or the headquarters of one of their chiefs, was near this landing-place.

FIG. 9. A Yaksna (WiMra Painting). '

Notwithstanding their supposed previous removal from the island about forty-five years before his-arrival (according to the statement that he came in the year of Buddha's death) we are told that Wijaya found the country still occupied by the Yakkhas. This is explained by the Rajavaliya, which states that some Yakkhas had concealed themselves in the midst of the forest, and thus escaped banishment. According to the Mahavansa, Wijaya married a Yakkha princess, called Kuweni, and with her advice and assistance succeeded in overcoming her countrymen and making himself master of at any rate a considerable part of Ceylon. A great part of the story of Wijaya's exile from his father's realm, and his journey to the island appears to be fictitious; but the whole account is valuable as indicating the early beliefs current in pre-Christian times regarding the aborigines.

In the Jataka tales, or instructive incidents in the former

ci8 ANCIENT CEYLON

lives of the last Buddha, Gotama?the most recent stories of which are at any rate of earlier date than the period of the compilation of the Dipavansa, while others date from the fourth or fifth century B.C.?some interesting evidence is forthcoming regarding the tract inhabited by the Yakkhas.

After the usual introductory remarks, the Valahassa Jataka (No. 196) begins as follows: ' Once upon a time, there was in the island of Ceylon a goblin town called Sinsavatthu, peopled by she-goblins. When a ship is wrecked these adorn and deck themselves, and taking rice and gruel, with trains of slaves, and their children on their hip, they come up to the merchants/ The story relates how they entice the traders to accompany them to the goblin city; ' then, if they have any others already caught, they bind these [other men] with magic chains, and cast them into the house of torment. And if they find no shipwrecked men in the place where they dwell, they scour the coast as far as the river Kalyani [Kaelaniya, which enters the sea at Colombo] on the one side and the island ?of Nagadlpa on the other. This is their way/ l Then follows an account of the ensnaring of five hundred shipwrecked merchants in this manner, and the escape of two hundred and fifty of them by the aid of the Bodhisattva [Gotama Buddha, in this former life], who assumed the shape of a wonderful flying horse which carried them back to India. When some new men were entrapped the Yakkhas are described as killing and eating the two hundred and fifty who were left behind.

This anecdote implies that the Yakkhas occupied all the coast districts outside the limits of Nagadlpa and Kaelaniya.

1 The ' goblins' were Yakkhas. It is to be regretted that the translators of these stories, as well as other translators, decided to transform the appellations of the various inferior supernatural beings who are mentioned in them, into words that are assumed to be their English equivalents, but in reality belong, in some cases, to beings of different characteristics. The word 'goblin,* for instance, would never mean to the ordinary reader both a being, Yafcsha, who was sometimes ranked in India close to the Gods?in the Atharva-Veda Yakshas precede the Rishis and the Fathers?and also a ghiJ, VMala, an eater of dead bodies. * Demon * and ' fiend * are used to designate stich different beings as DSaavas, Daityas, R&kshasas, Yatudhanas Pisachas,THE VAEDDAS 19

Taken with the information gleaned from the histories, this Jataka story renders it clear that the old authors believed them to have held the southern two-thirds of the island, including one-third of the western coast. The fact that the Nagas are described as being in possession of two-thirds of the western coast districts tells very strongly in favour of their coming from some part of the Malayalam tracts.

There is good reason to suppose that the accounts which the early writers have given respecting the Yakkhas have some foundation in fact. If so, they must necessarily refer, not to any supernatural beings who had made Ceylon their home, but to the aborigines, who In any case must have been driven out of the northern districts of the island by the intrusion of the Nagas. It is the general consensus of opinion that they are now represented by the Vaeddas, the hunting and fishing tribe who at one time occupied all the central forests as well as the southern coasts.

The late Mr. H. Nevill, of the Ceylon Civil Service, and others, have traced the identification of the Vaeddas with the Yakkhas, by the old authors, to a similarity of the names of the two classes of beings. According to this view, the Pali expression Yakkha was wrongly applied to the aborigines because of its resemblance to a title which is supposed to have been given to them as descriptive of their calling as hunters. It is believed by these writers that they were known as * Arrow-persons *; this would be expressed by the word iya, * arrow/ plus the personal suffix ka forming the word Iyaka which in sound is sufficiently close to Yakkha for such a confusion to arise. Although the arrow is certainly given a very prominent place in the ceremonies and worship of the Vaeddas, there appears to be no other evidence in favour of this derivation of the name applied to them by the ancient authors*

On the other hand, we have unmistakable evidence that they were known In pre-Christian times by the name which they still bear. The statement of the Mahavansa that in the fourth century B.C. King Pan£uk5bhaya provided a site at Anuradhapura for the Vyada-Deva, * the Vaedda deity/ and erected special dwellings for the Vyadas there, appears

20 ANCIENT CEYLON

to prove conclusively that at that early date the aborigines were known as Vyadas or ' hunters;' that is, Vaeddas, and not lyakas. In the Mahavansa they are also once termed Pulin-das, that is, savages or barbarians, a name applied by Indian writers to the Bblls ; and in place-names they are Sabaras, a word with the same meaning. It was probably due to exaggerated tales about these hunters, which the primitive Indian traders told their credulous countrymen on their return from their long and arduous expeditions to Ceylon, that the aborigines came to be denominated Yakkhas, that is, demons or goblins.

For the original home of these first comers we must search in the nearest aboriginal tracts of the adjoining continent, the hills of Southern India, or ttieir neighbourhood. It has been already noted that the Ramayana mentions the existence of Yakshas on them. Professor R. Virchow has shown that the character of the skulls of the present Vaeddas indicates a race with an affinity to some of the South Indian hill tribes. In several respects their customs incline to those of other South Indian hill-men, and their supreme deity is the Hill-God, whose cult prevails throughout the Western arid Southern Ghats. Perhaps the strongest evidence of the country of their origin is their own tradition that this deity came to Ceylon from Malawara-desa, * the Country of the Hill-region/ that is the Malayalam hills. It remains to be seen whether any affinities can be recognised between their dialect?which is practically a compound of modern Sinhalese, old Sinhalese, and a few Tamil words?and those of the South-Indian hill tribes.

There is nothing to indicate that the Vaeddas were ever the cannibals that the Jataka story represents them to be; the tale of their eating shipwrecked persons is an embellishment regarding the truth of which the later legends of the supposed habits of the true Yakshas would leave no doubt in an Indian mind. It may be taken to have no better basis than the fact that like many other aboriginal tribes they may have robbed and perhaps killed some of the traders wrecked on their shores, and seized the cargoes of their ships. On the other hand, the statement that the Pulayars of Travancore THE VAEDDAS 21

who are believed to be the aborigines of the plains in Southwest India, habitually file their teeth x must be admitted to afford some evidence that cannibalism was formerly a practice of that race, the habit of sharpening the teeth being almost always associated with anthropophagy. Had man-eating been also a custom of the aborigines of Ceylon, however, some distinct reference to it, |n addition to the very doubtful story of the habits of the Sirisavatthu residents, would almost certainly be found in the Sinhalese historical works, and the teeth of the Vaeddas would probably be filed to the present day, like those of the Pulayars.

On the whole, it may be concluded that the advance of the Dravidians to the south of India, which may have occurred before the entry of the Aryans into the north-western regions, may have eventually led to an exodus of an aboriginal and probably pre-Dravidian hunting and fishing tribe across the shallow strait that separates Ceylon from India.2

That this tribe in early times obtained food by fishing as well as hunting, may be gathered from the facts (i) that some Vaeddas live entirely by fishing at the present day; (2) that they are stated in the Valahassa Jataka to have wandered along the -shores round the southern and eastern part of the island; and (3) especially that in the eastern part of Ceylon, where the people who retain the name of Vaeddas are still found, the shark is a forbidden food to the Kapuwas (or demon-priests) of the jungles of the interior who conduct the worship in honour of their supreme deity. This prohibition must have arisen from an acquaintance with the man-eating proclivities of the shark, regarding which the natives of the interior could have no direct knowledge. Such a prohibition would never be thought of by any but residents on the sea coast who were accustomed to catch and eat the shark, and it would be quite useless among others who lived far from the sea. The shark is not a forbidden food to the Kapuwas in other parts of the

1 Rev. S, Mateer, Native Life in Travancore, p. 41.

% Dr. R. Virchow has already stated that *vre cannot avoid the conviction that they stand in a close affinity to the Aborigines of India.* (The Veddas of Ceylon, Translation, p. 131.)22 ANCIENT CEYLON

island. The custom is an evident survival from a time when a considerable part of the race gained a living by sea-fishing, and were aware of its necessity in order to preserve from defilement the officiators at the services in honour of their deity. I may add that it appears to completely negative the Indian story of the cannibalism of the aborigines. If they were eaters of human flesh they could have no reason for declaring the shark an impure fish because it ate the same food as themselves.

Many centuries must have elapsed before these wanderers could penetrate and spread through all the dense forests of the interior, and in considerable numbers occupy all the southern coast districts, as they are represented to have done by the fifth century B.C. It may thus be accepted as certain that their advent dated, at the latest, from the second millennium before Christ, if the primitive state of the wilder members among their descendants, and the advanced state of the more civilised portion of the race in early historical times, do not indicate an even more distant arrival in the island.II

THE ANCIENT VAEDDAS

Sinhalese histories contain several references to JL the aborigines of Ceylon, whom they usually denominate, in the Pali language, Yakkhas. The narrative of the Buddha's supposed visit to them has been given already. They are next mentioned in the tale of the arrival of Wijaya, the first Sinhalese king ; and the story, even if partly or chiefly fictitious, is valuable as an illustration of some of the notions which the invaders or new settlers held regarding them. On this occasion only two Yakkhinis (female Yakkhas) showed themselves and endeavoured to entrap the travellers, who were only saved because Vishnu had taken the precaution to tie charmed threads on their arms.

One of the Yakkhinis proved to be a princess named Kuweni, whom Wijaya married. She provided the adventurers with a good meal of rice and other articles taken from ships that had been wrecked on the coast of Ceylon. She is then represented as proceeding to recommend Wijaya to attack the Yakkhas of the neighbouring town, in the following terms (Mah. i, p. 33) :?" In the city Sirivattha [the Sirisavatthu of the Jataka story], in this island, there is a Yakkha sovereign Kalasena, and in the Yakkha city Lankapura there is another sovereign. Having conducted his daughter Pusamitta thither, her mother KondanamikS. is now bestowing that daughter at a marriage festival on the sovereign there at Sirivattha. From that circumstance there is a grand festival in an assembly of Yakkhas. That great assemblage will keep up that' revel without intermission for seven days/' The prince acted as advised by her, and * having put Kalasena, the chief of the Yakkhas, to death, assumed his court dress. The rest of his retinue dressed themselves in the vestments [or oma-

23

24 ANCIENT CEYLON

ments] of the other Yakkhas. After the lapse of some days, departing from the capital of the Yakkhas, and founding the city called Tambapanni Wijaya settled there.'

According to the narrative, Wijaya subsequently married a daughter of the Pandiyan king of Madura, and discarded the Yakkha princess, who went to Lankapura, where she left her two children outside the town(Mah.i, p. 35). 'The Yakkhas on seeing her enter the city, quickly surrounded her, crying out "It is for the purpose of spying on us that she has come back/' When the Yakkhas were thus excited one of them whose anger was greatly kindled put an end to the life of the Yakkhini by a blow of his hand. Her uncle, a Yakkha named Kumara, happening to proceed out of the Yakkha city, seeing these children outside the town, " Whose children are ye/' said he. Being informed " Kuwenfs/' he said,f Your mother is murdered; if. ye should be seen here they would murder you also; fly quickly/' Instantly departing thence, they repaired to the neighbourhood of Sumanakuta (Adam's Peak). The elder having grown up married his sister and settled there. Becoming numerous by their sons and daughters, under the protection of the king they resided in that Malaya [mountain] district. This is the origin of the Pulindas/ Thus it is plain that at the early date when the first annals consulted by the compiler of the Mahavansa were written it was known that the so-called Yakkhas were in reality the aborigines, the Pulindas.

In the time of the fourth king of Ceylon, Tissa, the chronicler returns to the old idea of the Yakkhas as a form of demon, and narrates (Mah. i, p. 41) that * A certain Yakkhini named Cetiya l {the widow of Jutindhara, a Yakkha who was killed in a battle at Sirivatthapura 2) who dwelt at the Dhurnarakkha mountain [which the context shows was close to the Kasa ford on the Mahawaeli-ganga, near Polannaruwa], was wont to walk about the marsh of Tumbariyangana in the shape of a mare/ which was of a white colour, with red legs. Prince

1 In this and all other transliterations the letter c represents the

sound ch, as in church. *'*?, : rf' .", ? ,

2 The words in brackets are only given, in Tumour's Mahawanso.THE ANCIENT VAEDDAS 25

Pandukabhaya, the nephew of the king, who had taken the field In an attempt to seize the throne, and now held all the eastern and southern districts, to the south of the river Maha-waeli-ganga, succeeded in catching this mare, and by her supernatural advice and help, that is, with the assistance of the Yakkhas or Vaeddas, defeated and killed the king his uncle, and the latter's brothers, with the exception of two, and thus secured the sovereignty.

He reigned at Anuradhapura, which he enlarged and rearranged, so that during his reign it became an important city. The chronicler relates that ' He established the Yakkha Kala-vela in the eastern quarter of the city; and the chief of the Yakkhas, Citta, he established on the lower side of the Abhaya tank [that is, on the south-western side of the town]. He who knew how to accord his protection with discrimination established the slave [Kumbokata], born of the Yakkha tribe, who had previously rendered him great service,1 at the southern gate of the city/ Thus he arranged that his Vaedda allies should be established on three sides of the city, doubtless as its ^defenders.

The cemetery was fixed on the western side of the town; and to the northward of it, and apparently near the main road which led to Mahatittha, the port from which travellers sailed for Southern India, * a range of buildings * was also constructed for the s Vyadas/ the Vaedda populace in general.

The Mahavansa also informs us that * he established within the garden of the royal palace the mare-faced Yakkhini.' It will be noted that this Vaedda chieftainess is no longer called a mare, but only mare-faced, just as nicknames such as * moonfaced/ ' crooked-nosed/ * large-toothed/ etc., were applied to the Sinhalese kings.

Thus it is clear that a large proportion of the population of Anuradhapura or its outskirts at that time consisted of the

1 She had saved his life when an infant. According to the history,

the so-called Yakkhas protected him from the time when he was born,

' his uncles having endeavoured to kill him on account of a prediction

that he would destroy them. If there is any truth in this, his father's

mother may have "teen a native princess.26 ANCIENT CEYLON

Vaedda supporters of the king. It has been already mentioned that he provided a site for the Vyada Deva, ' the Vaedda God/ also. The chronicler proceeds to indicate in unmistakable language the commanding position of the Vaedda rulers of this period : * In the days of public festivity, this monarch, seated on a throne of equal eminence with the Yakkha chief, Citta, caused joyous spectacles, representing the actions of devas [gods] as well as mortals, to be exhibited/

This important sentence proves that the supreme Vaedda chief of that day occupied a position little, if at all, inferior to that of the Sinhalese king.

The chronicler continues, ' This monarch befriending the interests of the Yakkhas, with the co-operation of Kalavela and Citta, who had the power of rendering themselves visible,1 conjointly with them enjoyed his prosperity/

It is easy to see that it was by means of a close alliance with the Vaeddas that this astute king, the greatest organiser the country has ever had?who is recorded to have made the first land settlement by defining the boundaries of the villages throughout the country?succeeded in deposing his uncle and gaining the throne. The natives were evidently far too numerous and powerful and well-organised to be put aside afterwards like the unfortunate Kuweni; and the politic king found it advisable to recognise the authority and influence of their leaders as nearly equal to his own. His political sagacity in this respect doubtless saved the country from many years of bloodshed and insecurity, and converted the Vaeddas into peaceable inhabitants devoted to his interests. In religious matters he was equally liberal and impartial; he made special provision for all religious bodies at Ms capital. It was he, also, who gave the first stimulus to reservoir construction in the northern districts, and probably also irrigation. The historian rightly referred to him as??' this wise ruler/ and stated that at his death the country was * in a state of perfect peace * (Hah. i, p. 44), This great monarch was born in about 345

1 We may recognise the hand of the reverend historian of the fifth

century in this little parenthesis.THE ANCIENT VAEDDAS 27

B.C., and reigned from 308 to about 275 B.C., or possibly a little later.

In the middle of the third century B.C. the account of the arrival of Mahinda, the son of the Indian emperor Asoka, on a mission to convert the Sinhalese and their king Devanam-piya Tissa to Buddhism, possibly indicates a certain retention of power by the Vaeddas, and the brusque manner in which they ventured to address the king. When Mahinda first met.the king in the jungle, * the thera [superior monk] said to him, " Come hither, Tissa." From his calling him simply " Tissa " the monarch thought he must be a Yakkha' (Mah. i, p. 50). Whether the story is true or false, it proves that the writer believed that the Yakkhas, who must have been either supernatural beings or the Vaeddas of that time, did not exhibit much deference towards the Sinhalese sovereign.

In the time of Duttha-Gamini (161-137 B-C0 there is a reference to a temple of a deity termed * Pura-Deva/ which is stated to have been on the northern side of the cemetery, where we have seen that the Vaeddas were settled. This god seems to be the Vyada Deva of the time of Pandukabhaya, the word apparently meaning * the Ancient God * of the country.

When the great Ruwanwaeli dagaba1 was constructed by this king at Anuradhapura, among the paintings depicted on the wall of the relic-room inside it the list runs : ' The four great kings of the Catumaharajika heavens stood there with drawn swords; and thirty-three supernaturally-gifted devas [inferior gods] bearing baskets of flowers and making offerings of paricchatta flowers [Erythrina indica, now used only for demon-offerings]. There stood thirty-two princesses bearing lighted torches, and twenty-eight Yakkha chiefs ranged themselves as a guard of protection [for the relics in the chamber], driving away the fierce Yakkhas' (Mah. i, p. 121).

In the Hatthi-pala Jataka (No. 509} a tree-deity is repre-

1 A dagaba is a solid mound built to contain relics of Buddha, or important personages, especially monks, or sometimes only to commemorate an event which occurred at the site. It is usually a semi-globe or a bell in shape, with a terminal spire ; but there are other forms, of which an account is given in a subsequent chapter. Dagaba =db§tii-garbha, * relic-chamber,128 ANCIENT CEYLON

sented as applying to the f eight and twenty war-lords of the goblins' to grant a son to a king. The beings mentioned in the Mahavansa are thus probably the same Yakkhas of the Indian authors.. At the dagaba at Bharhut, in India, these beings were carved in relief at the gateways of the ' Buddhist railing' in the third century B.C., as guards, together with Naga chiefs.

On the other hand, in Southern India it is the Rakshasas who always act as guards at the Hindu temples, in accordance with the derivation of the word from the root rdksh, to guard. When deities are represented on the gopuras or ornamental gateways at the entrances of the great temples, figures of the Rakshasas are invariably present as their guards, and the Yakshas are never found in such positions of trust.

In the later wall-paintings of the Buddhist wiharasin Ceylon, the Yakshas always form the army of Mara, the god of Death, which attacked the Buddha ; but this has been shown to be a conception of later date than the canonical works, and it may not have found acceptance in the country in the time of Duttha-Gamini It is, however, somewhat strange to find Mahanama inserting the description of these figures in such a position in the dagaba without some explanatory remark. He may have understood them to be representations of aboriginal chiefs.

I belfeve the Vaeddas only make their appearance twice more in the early Sinhalese histories. The" Rajavaliya relates that King Maha-Sena (277-304 A.D.) employed Yakkhas as well as [Sinhalese] men in the construction of a large number of reservoirs that were formed in order to store water for the Irrigation of rice fields. Some confirmation of this story may be seen in his deification at some subsequent period, with the title of Sat-Rajjuruwo, that is, * King of (all) living creatures/ ?both the men and the supposed demons whom he forced to work for him. Worship is still paid extensively to him in this capacity in the northern Kandian districts.

The Vaeddas still formed a great part of the population in the twelfth century. The Mahavansa (ii, p. 151) recounts how King Parakrama-Bahu I (1164-1197 A.D.), while his cousinTHE ANCIENT VAEDDAS 29

Gaja-Bahu ruled at Polannaruwa, made preparations for a campaign for the conquest of the latter's dominions, and enlisted for it large numbers of his subjects. Among these we are told that ' He trained many thousands ' of Vyadas, that is, Vaeddas, * and made them skilled in the use of their weapons, and gave them suitable swords, black clothes, and the like things/ Thus in the twelfth century we see the Vaeddas in a state of comparative civilisation, taking their place in the army with the other levies.

It is extremely probable that contingents of Vaeddas formed part of the Sinhalese army not only then but in every war. We find them still serving with the other troops under Raja Sinha in the early part of the seventeenth century. Captain Robert Knox, in his Historical Relation of the Island Ceylon, p. 62, states of those living near Hurulla, in the North-central Province, * The King once having occasion of an hasty Expedition against the Dutch, the Governour summoned them all in to go with him, which they did. And with their Bows and Arrows did as good service as any of the rest; but afterwards when they returned home again, they removed farther in the Woods, and would be seen no more, for fear of being afterwards prest again to serve the King/

As the immigration, such as it was, from the Ganges Valley appears to have practically ceased from the time of Panduka-bhayajs birth, his policy of admitting the natives to an equality with the Indian settlers must have caused a rapid fusion of the two races. This was the birth of the Sinhalese nation.1 We must believe that such a broad-minded ruler would not

1 The tradition of the origin of the name is given as. follows in the

Mahavansa i, pp. 33, 34. ' By reason of the King Sihabahu [the father of Wijaya] having slain the lion (Sl"ha), his sons and descendants

are called " Sihala " (the lion-slayers). This Lanka [Ceylon] having been conquered by a Sihala, from the circumstance of its having been colonized by a Sihala, it obtained the name of Sihala.* At a much later date it became the fashion to adopt Sanskrit forms of words in writing, and instead of the Pali word Siha the Sanskrit expression Sinha was used. The word meaning the country and people thus became * Sinhala* (pronounced with a nasal n, but no g sound). The Vaeddas have retained the old name of the country.;,;u; ri - ~~*w-*^*CT«'^^^^«*^

30 ANCIENT CEYLON

refuse equal rights to the northern Dravidians of Nagadipa, and thus the whole population must have gradually coalesced, with a great preponderance of the Vaedda blood. In the same manner as in England in Norman times or after the Roman domination, the natives in the lapse of years totally absorbed the newcomers, and a later very slight admixture of Tamil blood at last produced the race which we now find in the Kandian provinces. It differs from that of the western and southern coast tracts in all respects but colour, religion, and language.

In a note on the subject of Polyandry, the late Mr. E. Goone-tilleke, the learned Sinhalese editor of the Orientalist, said in Vol. iv, p. 93 of that publication, regarding the two races of Sinhalese, ' They are as distinct from each other in their dress, habits, manners, and customs, and in their very ideas and manner of thinking, as if they formed two different races, rather than two sections of one nation/ The Kandian villagers certainly look upon the people of the western coast tracts as a separate race, and do not term them Sinhalese, but always speak of them as Pata rate minissu, 'Men of the Low-country/

The difference is not altogether due to a preponderance of Vaedda blood in the interior. The dwellers near the western coast have always been exposed to foreign influences. The various races who have either settled among them in considerable numbers or held the western coasts as conquerors include Dravidian and Arab traders and settlers; and as conquerors, Malays, Chinese, Portuguese, Dutch, and lastly English. It would be strange if the resultant people did not vary greatly from those of the interior.

\ That the Kandian Sinhalese are thus the modern repre-,

; sentatives of the great bulk of the ancient Vaeddas is, I venture

. to think, beyond doubt. The people who were so numerous

throughout the country in the twelfth century, that in half

the island ' many thousands * could be enlisted as soldiers,

have certainly not been exterminated. They, like the Vaeddas

of preceding centuries, have simply settled down as Kandian

* villagers. An insignificant number still retain their ancientTHE ANCIENT VAEDDAS 31

designation, but even these, with the exception of a few families, have become ordinary villagers, and in outward appearance are indistinguishable from many other Kandians.

This abandonment of the wild forest life of their ancestors apparently began at a very early date. After the time of Pandukabhaya the next proof of the fact is found immediately after the introduction of Buddhism into the country. The evidence derivable from the caves or rock-shelters, thousands in number, under the sides of the boulders lying on the slopes of all the hills of the Low-country, whether in the eastern and southern part of the Northern Province, or the North-western, the North-central, the Eastern, or the Southern Provinces, all points to the settling down of the Vaedda populace in early times as peaceable villagers.

The researches of the Drs. Sarasin and Dr. C. G. Seligmann have shown that the first inhabitants of the caves were aborigines who made use of stone implements. Then, at a later date, which we know from the dedicatory inscriptions to be in nearly all cases pre-Christian, the caves throughout the whole of the above-mentioned Provinces (I have no knowledge of those of other districts) were turned into shelters for ascetic Buddhist monks. There is hardly a hill possessing such cave shelters, some of which, at least, were not so converted. Even where no inscription records the fact, the cutting of the katdr® or drip-ledge to prevent rain-water from trickling down the face of the rock into the cave is indubitable proof that this was the case.

Had the aborigines been forcibly ousted from these caves in order to permit the monks to occupy them, we cannot suppose that they would not have felt resentment, which would have led to reprisals of a violent character. It is clear that in many instances little establishments of only two or three monks must have occupied the caves on some of the most secluded of these hills, buried in the depths of the dense forests of the wildest parts of the island. In such sites the aborigines could have regained possession of their caves with ease and impunity, and with practically no fear of punishment by the Sinhalese authorities. In the histories, also, there is

32 ANCIENT CEYLON

no hint of any quarrels with the natives after the time when

Pandukabhaya became king.

If the monks who occupied the caves had been in danger of attacks by the aborigines, it is extremely improbable that they would have utilised the caves on practically all the hills during the short period between the middle of the third century and the early part of the first century B.C., as the form of the letters of the inscriptions cut on so many of them? ' hundreds and hundreds/ according to Dr. E. Muller?proves was the case. A few caves, but only an insignificant number, have inscriptions cut in letters of a later date than this. Thus there seems good reason to believe that when the monks came to occupy the caves their original residents had already voluntarily abandoned them, and, like the Vaeddas of Anuradhapura, had established themselves in villages.

Even the people who still call themselves Vaeddas are to some extent of mixed blood. This applies almost equally to the wildest members of the race, and is proved conclusively by the wide variation in the colour of the skin, and in the amount of hair on the face, even if the general outline of the features does not indicate it.

It was probably due to the union of the races on nearly equal terms that the Vaeddas accepted the language of the Gangetic settlers in preference to their own, which they have totally lost. Had they kept more aloof from the newcomers, they might have maintained their own tongue nearly intact down to the present time. The new language spread through Nagadlpa also; there is not a single very early Dravidian inscription in the whole of Northern Ceylon. The adoption of the Buddhist religion throughout the entire country? induding Nagadipa, as the numerous remains of ancient wiharas prove?must have accelerated this change of language ; at every monastery the monks would teach the dialect of Pali which had become the Sinhalese speech, in the same manner as at present.

Notwithstanding the alteration of language and ideas and the spread of the new religion, the population of whole districts must have remained more or less pure Vaeddas for manyTHE ANCIENT VAEDDAS 33

centuries, with some gradual slight intermixture of foreign blood as the intercourse with Nagadipa and Southern India led to an intermittent influx of Dravidians, culminating in occasional invasions of the island by South Indian armies. In some cases, in what are now thought to be pure Sinhalese districts, many of the people were still distinguished from the other inhabitants by the name of Vaeddas down to the seventeenth century, after which they appear to have abandoned this title to the wilder residents of the eastern districts.

Although declaring themselves Buddhists and attending the services at the temples, many of these Sinhalese-Vaeddas still adhered to the worship of the ancient Hill-God of their ancestors, the Vyada Deva of the old annalists. The philosophical reasoning of the new faith might appeal to their minds, but it did not afford the practical protection which they received from their old religion. They still felt the need of the kindly supreme deity to whom they could appeal in time of trouble, for which the new faith provided no remedy, but only taught resignation to the inevitable. The ancient god could still, it was thought, assist them out of their physical difficulties, without interfering with their general belief in the truth of the Buddhist doctrines. In some parts of the Kandian districts the two religions have therefore settled down side by side to the present day.

Dr. R. Virchow, as the result of an examination of a series of Vaedda and Sinhalese skulls, expressed the following opinion regarding the affinity of the Vaeddas and Sinhalese: * The Vaeddas would appear rather as representatives of the aboriginal race; the Sinhalese, on the other hand, as hybrids produced by a union of immigrant Indians with Vaeddas, and therefore varying according to the measure of their participation of either of these elements. This indeed strikes me as being the solution of the anthropological problem before us, so far, at least, as the material at present reaches. The linguistic difficulty, that also the unmixed * natives adopted the Aryan language of the conqueror, without, so far as we can

1 It is extremely doubtful if there are any groups of Vaeddas of unmixed blood in these days.

D

34 ANCIENT CEYLON

mdee having been forced to do so, appears to me no longer Lurmountable, since from personal experience I have established the fact that in the Baltic provinces of Russia one part of the Finnish population after the other, through impercepti-ble but steady progress, has become letticized to such an extent that the Courland language has wholly, the Livoruan almost wholly, disappeared, and only the Esthonian still offers any

nal conclusions on the subject are : ' (i) That manifold resemblances exist between the Vaeddas and the Sinhalese and that the origin of the Sinhalese race from a mixture of Vaeddas and immigrants from India possesses great probability, as well upon historical as also upon anthropological grounds. .

' (2) That the Vaeddas as well as the Sinhalese in the main features are distinguished from the Ceylon Tamils, and equally from those of Tanjore (Sola). ?

? (3) That, on the other hand, among the remnants of the old Dravidian or perhaps pre-Dravidian tribes of Hindustan we find even to-day evidence of analogies with- the Vaeddas ' (p. 136).

i Monograph on the Vaeddas, published in the Memoirs of the Royal Academy of Science of Berlin, in i88z, and translated for the Ceylon Branch of the Royal Asiatic Society m 1888, p. no.Ill

THE MODERN VAEDDAS AND

WANNIYAS

following account of these races depends on original JL observations of them made by myself during official visits to their districts, largely supplemented by quotations from articles on the Vaeddas by the late Mr. Hugh Nevill of the Ceylon Civil Service, which he published in 1886, in his magazine. The Taprobanian.1 I have endeavoured to credit him with all information taken from his papers. He had the advantage of being stationed in the Eastern Province for a considerable time, first as Assistant Government Agent of the Trincomalee district, and afterwards as District Judge of Batticaloa ; and being an indefatigable student and an accurate observer, and well acquainted with the native languages, he was able, owing to his official position, to collect a large amount of valuable information regarding the Vaeddas, as well as other subjects, which would not be readily available to others. It is greatly to be regretted that the part of it relating to the ceremonies used in their demon worship was never made public by him. I have also quoted some remarks on the Vaeddas by Professor R. Virchow, together with the sizes of their skulls as noted in the valuable monograph on them already referred to. Throughout this account of them I have instituted comparisons between them and the present Kandian Sinhalese. I am well aware of the defective nature of this account ; but as it contains some information which is not elsewhere available, I have thought it advisable to publish it.

1 1 am indebted to the courtesy of Ms brother, Mr. Ralph Nevill, for permission to utilise them.

3336 ANCIENT CEYLON

The Vaeddas of the present day, or those known as such, are found only in the eastern half of the island. They are usually divided into three classes, which I shall distinguish as follows :?

(1) The wild Forest Vaeddas, few in number, who live entirely by hunting, and dwell in the depths of the forests near the eastern base of the Kandian mountains. At Nilgala, where I expected to find them well known, I was surprised to learn that they are rarely seen; all of whom I could hear in that neighbourhood consisted of one small party who sometimes visited or resided on a hill about five miles away in the forest. There are more of them on the western side of the valley of the Madura-oya.

(2) The Village Vaeddas of the eastern interior and the south-eastern coast districts, who in many cases, but probably not in all, have some intermixture, recent or ancient, of Sinhalese blood, though practically forming the same race as the Forest Vaeddas. There are two villages of these Vaeddas in the North-central Province, near Hurulla tank, and several others on the eastern side of the lower part of the Mahawaeli-ganga, but the great majority live in the Eastern Province.

(3) The Tamil-speaking Vaeddas, who live in scattered villages on or near the central coast tract, from the north of Trincomalee to about ten miles north of Batticaloa. These have intermarried with the Tamil residents of that part of the country, and have adopted their language, and some of their customs, while still retaining some of their own.

Distributed among some eighteen small hamlets along the northern border of the North-central Province, which is the boundary between the Tamil districts of the north, the ancient Nagadipa, and the Kandian Sinhalese, there is also a race of hunters, probably less than 500 in number, who, like the others, are termed by the Tamils Ve Jan (in English pronunciation Verdan), plural Vedar. They themselves repudiate this appellation, except in. its ordinary meaning of e hunter/ and they deny that they are in any way connected with the Vaeddas, of whom they speak in very contemptuous terms. Their own name for themselves is Warujiya, ' person of the Wanni/THE MODERN VAEDDAS

37

as the forest and jungle of northern Ceylon to the south of Elephant Pass is called. They all speak Sinhalese, with the exception of the inhabitants of one or two hamlets lying to the west, but all the men also know a certain amount of Tamil. As their habits when engaged in hunting do not differ from those of the Vaeddas, it will be useful to include them in dealing with the latter, especially as some consider them to be true Vaeddas, with whom, in fact, it is not unlikely that they are connected, although they have lost all tradition of it, and neither know the Vaedi1 dialect nor, so far as I am aware, worship quite the same deities.

Like the Vaeddas, they all claim to be of good caste (in their case the Goyiwansa, or cultivating caste), although, like them also, many have names such as elsewhere now belong only to persons of the low castes like the Tom-tom beaters ; among these may be mentioned Kanda, Velan, Kata, Kona, etc. Others have what are considered to be good caste names. , On examining the inscriptions and histories, however, we leam that two thousand years ago, or more, the short names that are now confined to the lower castes were borne by the chiefs, and even by the members of the royal family. In Ceylon, in early times there seem to have been no names that were specially distinctive of the high and low castes ; where a distinction was made it was provided by the addition of a separate ending, of which instances occur in the names found both in the cave inscriptions and the histories, such as the

1 Vaedi Is the adjectival form. In Sinhalese, the masculine noun. is Vaedda, plural Vaeddo, and the feminine noun is Vaeddl. I believe these nouns are only employed by Sinhalese. I have not heard the Vaeddas term themselves otherwise than as * Vaddl men' (Vaedi minissu). In their own dialect this would be Vaedi minu, but a Vaedda has been represented as calling one of his race Wanniknde mina9 and the word Mai occurs for Vaeddas in the invocations collected by Dr. Seligmann. The Tamil-speaking Vaeddas call them* selves VSdan. Reasons have been given for doubting if the word Vaedda could be derived from the Pali word Vyadha. In any case, that VyMha, however,, signified Vaedda is, I think, clear from the use of this term in the Mah&vansa to describe the * many thousands J enlisted by Parakrama B3.hu I. In a footnote at the end of the chapter on the Primitive Deity of Ceylon I have given an intermediate form found in one old work.38 ANCIENT CEYLON

terminations Gutta (Gupta), Sena, Deva, Mitta (Mitra), and Naga. As regards their characteristic names, therefore, the Wanniyas and also the Vaeddas have simply retained the custom of pre-Christian times.

At the Census of 1901 the total number of all classes of Vaeddas, including, I presume, the Wanniyas, was found to be 3,971. The numbers obtained at the two preceding decennial. enumerations were so defective that no conclusions can be based upon them regarding the increase or decrease of the race.

Little is known of the Vaeddas of the first of the three classes, who are almost inaccessible in their wild forests.1 Formerly they were accustomed to lead a more or less wandering life, which in the case of each little family party was confined to a definite tract of forest, sleeping in caves at the foot of the hills, or under trees. They still make use of the caves, but their village neighbours informed me a few years ago that all now build huts in the forests and inhabit them at times when they are distant from their cave shelters. Those whom I have seen were indistinguishable from the Village Vaeddas ; they appeared to be healthy and well nourished. According to Mr. Nevill, they change their quarters from time to time when the game and "Iguanas' (large terrestrial lizards) of their neighbourhood are killed or driven away. So far as my own limited observation extends, I quite agree with Mr. Nevill that the Forest Vaeddas and the wilder Village Vaeddas are the same people. It is a mistake to suppose that aU Village Vaeddas are of more mixed descent than the Forest Vaeddas ; many are simply Forest Vaeddas who have settled down, in recent times in more or less permanent hamlets.2

Clothing.?They are a wild-looking race, wearing a minimum of clothing, which consists, in the case of the men, of a small rag or strip of calico suspended in front from a bark string tied round the waist, and when hunting a larger strip of discoloured

1 Dr. C. G, Seligmaim, accompanied by Mrs. Seligmann, has succeeded in finding some families of these Forest Vaeddas, and Is about

to publish an exhaustive account of them and their customs and beliefs.

s See the footnote at the end of this chapter.THE MODERN VAEDDAS

39

cloth which is passed round the abdomen in three or four folds, forming a narrow flat band about four inches wide. It is recorded that in. the early part of last century some Vaeddas wore a short skirt made of the liber or fibrous inner bark of the Riti tree (Antiaris innoxia), like the material of the bark bags which they still prepare for household purposes. It may be considered certain that where these trees were found this must have formed the general costume of the wilder individuals at a time when cotton cloth was unobtainable; and I was told that a very few of the poorer people still employ it for the same purpose.

Some have also been reported to wear green leafy twigs suspended from a bark string tied round the waist; but this may have been merely a hunting device to avoid notice of their cloth by wild animals. I have seen the Wanniyas using this primitive costume on such occasions, but only as a temporary expedient. Mr. Nevill mentioned that he was informed that in ancient times leaves were so worn as clothing in districts where there were no Riti trees. Only the poorest among them wore this dress, and that not from choice but necessity. He considered that there is no reason, to suppose that they ever went about in a state of nudity, I never heard that any of them have worn skins.1 The account of the natives at the time of Wljaya's arrival would lead one to suppose that some at least wore clothing which the newcomers did not consider primitive.

When in the forests, the Village Vaeddas of the interior, as well as the Wanniyas, dress in the same manner as the ordinary Forest Vaeddas, and roll up their cloth and fasten it round the abdomen like them. The females of both classes have similar clothing, a short skirt of cotton fastened round the waist and reaching to the knees or below them. When visiting other villages the men wear a similar cloth from the waist to the knees or below them.

1 Ribeyro, whose work was written in 1685, stated that those -who lived In the forests north-of Trixicoxnalee (? Wanniyas) wore the sMns of animals, but lie does not say that he ever saw them,. ICnox would not be likely to omit mentioning the custom, if it had be?n practised in his time.40 ANCIENT CEYLON

The Tamil-speaking or Coast Vaeddas dress like Tamil villagers, with a cloth reaching from the waist nearly to the ankles; the women wear a long calico robe which is passed round the body under the arm-pits and hangs straight down nearly to the feet. It is the ordinary costume of the village Tamil women of northern Ceylon, and is singularly ungraceful.

General Description.?I may premise that as regards Anthropology, so far as it relates to the scientific description of the human body, I possess neither qualifications nor knowledge, and I have therefore collected no information beyond that of a casual observer who is well acquainted with the other races of Ceylon.

The skin of the first two classes of Vaeddas is commonly of a dull dirty-looking dark reddish-brown colour, which may be termed a dark walnut hue. There is nearly always a distinct reddish tint in it. The difference between it and the colour of some low-caste Kandian Sinhalese is so slight that I am unable to define it; I should say that it consists chiefly in the duller appearance of the Vaedda skin. Many of the Coast Vaeddas and a few of the Village Vaeddas and Forest Vaeddas are much darker than this, and of a brownish-black colour, this shade evidently indicating a mixture with Dravidian blood.

Mr. NeviH considered that the Vaeddas belong to a light brown race, and the Sinhalese to a light yellow race, and he even thought that both the Sinhalese and Vaeddas * are of one original colour, yellow, with an olive tint.1 This does not account for the reddish hue of the Vaeddas, which can almost always be seen in a full light, and sometimes very conspicuously. It has reddish-brown or reddish-purple shadows. It is often present in the skins of Kandian Sinhalese, some few of whom are even of a clear dull copper-red colour. This tinge is never seen in the skins of Tamils, and is hardly observable among Telugus, at any rate those of low castes ; but I have noticed it very plainly in several Kanarese from Maisur, some of whom are of a clear copper-red colour. The pale brownish-yellow tint of Sinhalese is only found in the members of families of what is now thought to be the purest descent, such as those of many of the leading chiefs ; it is the colourTHE MODERN VAEDDAS 41

of those who most closely represent the original settlers from the valley of the Ganges, and is far from being the average colour of the race who comprise the Kandian or Low-country Sinhalese of the present day, which is much nearer a dark walnut tint. In the ordinary Kandian villager all shades are found from clear copper-red through varieties of reddish-browns to the deepest blacklead black, but the tints at the extremes of the scale are uncommon.

The height of the Village Vaeddas is less than that of the ordinary northern Kandian villagers, and in the case of the men averages probably five feet, or an inch more, the Sinhalese being two inches or three inches taller. Recorded measurements of Forest Vaeddas show that they, or many of them, are much shorter than this, and vary between four and five feet, but always above the lower figure.

Although their figure is always very slight, with narrow hips and weak-looking calves and thighs, the Vaeddas are active and lithe in the forests, and can thread their way for many hours among the trees and jungle without apparent fatigue. When alone one morning in thick forest remote from any villages, I met a party of Vaeddas who were in search of honey. In reply to my inquiry regarding their hamlet, they informed me that it was 'quite near * a tank (reservoir) which was four miles away, but I afterwards learnt that the place was several miles beyond it. They had made the journey that morning, and probably would return also, through'a forest full of undergrowth.

Nearly three hours later, as I was returning along the path after visiting the reservoir, I sat down at the side of a tiny streamlet of clear water, fresh from a neighbouring spring, in order to get a drink, and enjoy a quiet pipe under the cool shade of the tall forest trees,, when suddenly one of the party, an intelligent young fellow with a pleasant countenance, stepped out of the thick bushes and joined me. He had left the others some distance away, and had come on for a drink. I gave him the contents of my tobacco pouch, and found him quite communicative and acquainted with Sinhalese, which he spoke intelligently, although he addressed me as Umba,*;} f^r*

42 ANCIENT CEYLON

you, an expression which is usually applied only to inferiors. He stated that they only knew and visited the people of one small settlement several miles away. No others lived within some hours' journey from their huts. He laughed at the fears which some Tamils had expressed to me regarding the demons who were supposed to infest that part of the forest, though he admitted that it was full of them. These people were apparently true Vaeddas, but not now the Forest Vaeddas, who are, I believe, unacquainted; or only slightly acquainted, with ordinary Sinhalese. In physical appearance and colour they resembled Kandian Sinhalese of some low castes. Their ancestors were Forest Vaeddas in the first half of last century. Vaeddas have not the slightest negroid appearance. Their jaws are not prognathous, the facial angle is good, like that of the Kandian Sinhalese, and according to my observation their noses are usually straight and rather well-formed, though somewhat wide at the nostrils. They have not very large orifices. Mr. Nevill said that they are ' squat, with no bridge to them * ;.?': evidently they are of two types. Mr. F, Lewis, of the Forest Department, has informed me that the Village Vaeddas whom he has seen had commonly straight noses and somewhat thick lips. In the case of those whom I have observed the lips were perhaps thinner than those of the Sinhalese. The cheek bones are always somewhat prominent, but this may be partly due to the absence of superfluous flesh on the face. The eyes are rather deep set, but otherwise resemble those of Kandians. Some faces are practically hairless below the eyes, and there is rarely more than a very sparing growth of hair on the face, a very thin short moustache and a little short hair on the chin being all that is usually present. In this respect, also, they resemble many Kandian Sinhalese, but not Low-country Sinhalese, who are a distinctly hairy race, and often have thick beards, hairy chests, and a central line of hair down to the navel, which is said to be thought a mark of beauty. This is quite uncommon among Kandian Sinhalese, and apparently totally absent among the Vaeddas. A few Vaeddas have more beard than others, but it is always thin ; such a feature may indicate some mixture in theirTHE MODERN VAEDDAS 43

blood; I have seen it with a very dark skin. The forehead is narrow and not high ; it does not recede much from the line of the face.

Dr. Virchow gave the following proportions of their skulls, together with those of Sinhalese x and Tamils :?

Number of Capacity in Height Breadth

Skulls Cubic Centimetres Length Length

Vaeddas . . . 20 1261 74-9 71'6

Sinhalese . . .10 1438-8 75-4 72*4 *

Tamils . . .4 1247 77-5 75-3 * In sixteen skulls this was 72*2.

He remarked that the average index of the ratio between the length and breadth proves that the skull is * decidedly dolichocephalous/ only four out of the twenty being mesocephal-ous, with an index of seventy-five, while the index of seven was under seventy. He also stated that c no elaborate proof is needed that neither Sinhalese nor Vaeddas, at least in the form of their skulls, present the slightest indication of any relationship to the Mongols. Such a remarkably dolichocephalous tribe has never yet been found among the Mongols/ I may add that neither do they resemble the Australians in any respect, to judge by the illustrations of them in the elaborate works of Dr. Howitt and Messrs. Spencer and Gillen. On this subject Dr. Virchow said : * One glance at the skull, and still more at the skeleton, of the Australian convinces us that here a great and unmistakable contrast exists/ 2 Some have endeavoured to connect the Vaeddas with the Andamanese. This is at once disposed of by Dr. Virchow, who remarked : * The Andamanese, as well as the Negritos generally, are in reality brachycephalic, and this one circumstance distinguishes them definitely from all the Ceylon races. If we add to this that their hair grows in spiral coils, and is to be classed with the woolly hair of the genuine negro, then every possibility disappears of a union with the Vaeddas unless we assume that climatic influences have specially affected the hair/ 2

The hair of the Vaeddas is black with a slight brownish

1 It is uncertain how many of these were the skulls of Kaudians. 8 Op. a!, p. .131,44 ANCIENT CEYLON '

tinge, and, if attended to, is not more frizzly than that of ordinary Kandian Sinhalese. It is never cut, and is tied in a knot at the back of the head (as stated by Knox, p. 62), exactly like that of all Sinhalese. Photographs of some Village Vaeddas who have been brought to Kandy and elsewhere to be exhibited represent men with wild unkempt frizzly locks ; but I have never seen anything of the kind in their own districts, and it is probable that the heads of those who have been so pourtrayed have been ' made up ' specially, in order to increase their wild appearance?as, in fact, I was informed by their Sinhalese neighbours has been done on similar occasions. The wildest Vaeddas whom I ever met, in the middle of dense forest, had their hair tied up in a knot at the back of their heads in the usual way of the villagers; these were the true Forest Vaeddas who could speak only the Vaedi dialect. It may occasionally be a practice of the Vaeddas when hunting, as it is of other hunters in Ceylon, to wander in the forest with unfastened hair ; but from my own experience of them, and from that of Sinhalese who live in their district and are well acquainted with them, I am able to state that it is not otherwise done habitually by any but an extremely limited number. In answer to special inquiries, I was informed that some few individuals do neglect to attend to their hair, and allow it to stand out in this wild-looking manner. Instead of their hair being naturally frizzly, I have never seen a Vaedda with hair more wavy than that of the Low-country Sinhalese of the western coast districts. I may repeat that so far as superficial appearances go, there is nothing in the figure (except the smaller height), the features, or the ordinary coiffure, and very little in the average colour of the skin, to distinguish the Vaedda from many low-caste Kandians found in the northern and north-western Sinhalese districts.

There is only one race in Ceylon with curly hair ; they are the Kinnaras or Karmantayo, the mat-weavers, the lowest caste in the island. In the case of some of the men the whole hair of the crown consists of a mass of very short thick curls, while the lips of those I have seen were invariably rather thick, although the jaws were not prognathous. Their faces resembleTHE MODERN VAEDDAS 45

in other respects those of Kandians, and are not of the Mongolian type. The hair of the women is tied up in a knot like that of the ordinary Sinhalese. The men never allow their hair to hang down beyond the "upper part of the neck, even in the case of those whose locks are not so curly as others ; it is always cut off when it reaches this length. The colour of these people is the same dark brown as that of the average Kandian villager; I have seen none who were much darker than this.

Their mode of life does not indicate any connexion with the Vaeddas, none of them being either hunters or fishers ; all gain their living by weaving mats in frames, and by cultivating millet and rice. They have village tanks and rice fields, and keep cattle; their villages and houses are clean and neat, being exactly like those of the Kandian Sinhalese. They have no tradition regarding ttheir origin, and no dialect of their own, knowing not one word except Sinhalese ; and nearly all their folk-stories are the same as those of the Kandians. Those which vary from the latter are chiefly Buddhistic, the race being all Buddhists, though not permitted by the Kandians to enter the wiharas, or the houses of other villagers. Their rank is so low that, as some of them admitted to me, they address even the Rodiyas, whom many wrongly believe to be the lowest race in the island, as Hamaduruwo, * my Lord/ and do not pass them on a path without first asking permission to do so. I was informed that the Rodiyas at once interfere if any of the men attempt to allow their hair to grow beyond the tipper part of the neck, and order them to cut it shorter.

I believe that they are now found only in the district immediately to the north and north-west of Kandy and near Kuru-naegala ; but a Sinhalese folk'tale places some on the western coast. This may indicate that we have in them the remnant of another tribe who came from the Malayalam 'country. It is interesting to note that, like the Vaeddas, they have completely abandoned their original language.

On the other hand, there is another race, of which only a few villages exist in the North-western and perhaps in the North-central Provinces, called ' Waga * or * Waga men / who46 ANCIENT CEYLON

are traditionally supposed to be the descendants of some of the Tamil captives brought from Southern India by Gaja-Bahu I, in the second century A.D. These people, though nearly as much isolated among the Sinhalese as the Vaeddas, hut not so much as the Kinnaras, still retain and speak their original Tamil tongue, in addition to Sinhalese. They closely resemble Sinhalese of some low castes, and are rather darker in colour than the average Sinhalese villagers. Why some races should have abandoned their mother tongue and others have retained it is a fact for which I am unable to offer any satisfactory explanation.

The Waga people, although they are supposed to have been originally only charcoal burners, are now cultivators exactly like their neighbours. They term themselves of good caste, and the men have the usual names which denote that position, such as Maenikrala, Kapurala, etc.; but the women have names that belong to persons of low caste, such as Bokkl, Bandi, Bad!, Kombi, Gaembl, Tikiri, LattL One might expect the name of the race to mean Vanga, that is, Bengal, but that the people both speak Tamil and claim to be Tamils.

The figure of most of the Tamil-speaking Vaeddas naturally approximates to. that of the Tamils with whom they are intermarried?so much so that there is little in it to distinguish them, and especially the women, from many village Tamils of a rather low caste. In the greater width of the hips and the amount of posterior tissue, the difference between the females and the Village Vaedda women is marked. Their colour is also commonly darker than that of the Vaeddas of the interior, and is sometimes black, with brownish shadows.1 The character of the features of the men approaches that of the Village

1 The only races I have seen with jet-black skins, which always have distinctly blue or purple shadows, are many of the Tamils of Southern India (not Ceylon), and all the Wolofs of the Senegal and Gambia coast districts, who have no resemblance to the true negroes, Some of the Andarnanese are also described as having skins of this black-lead colour. The same peculiar colour is to be seen in some few northern Kandians, but such cases are quite exceptional,, and are doubtless due to a strain of Dravidian blood. It does not occur among the Vaeddas.THE MODERN VAEDDAS 47

Vaeddas; there are the same scanty hair or absence of hair on the upper lip and chin, and the somewhat prominent cheek bones, and, according to my observation, straight noses. The hair is always tied in a knot at the back of the head.

The description of the Village Vaeddas is generally applicable to the Wanniyas, who, however, are perhaps an inch or two taller, on an average, and I think have slightly less prominent cheek bones. Their eyebrows are low and fairly straight, their eyes deep set, their noses generally straight, and their lips not thicker than those of the average Kandian villager. There may also be a slight difference in the shade of the skin, which is perhaps not quite of the same dull dirty tint as that of the Vaeddas ; but otherwise, like theirs, is nearly always a dark brown with a reddish tinge, though darker shades are also seen. There are variations in the colour, some having distinctly reddish skins, and others skins of a deep walnut hue. The hair is nearly straight, and excepting sometimes when they are hunting is always tied in a knot at the back of the head. The face is commonly nearly hairless below the eyes. The women differ in appearance from Tamils ; they have oval faces, pleasant comely features, and not ungraceful figures. Among all Vaeddas and Wanniyas the superciliary ridge is rather prominent; it is never absent in Kandian Sinhalese, but is often unnoticeable in Tamils and the so-called ' Moormen.*

Ornaments.?The Tamil-speaking male Vaeddas and those of the south-eastern coast tract, who are brought into communication with the Tamils, or Sinhalese who have adopted some of the habits of Tamils, carry a ring or stud in the lobe of each ear after marriage, and some of the former also wear silver bangles. The Vaeddas of the interior and the Wanniyas of ten'have silver rings in their ears, and I have observed the Forest Vaeddas with similar ornaments, which some of the most northern Kandian villagers, as well as the Rocjiyas, also commonly wear, but not other Sinhalese men, nor the Kinnaras.

Mr. Nevill remarked that the females put on necklaces of coloured glass beads when they can get them, and shell, ivory, glass, or brass bangles. The Village Vaedda women are said

48 -: ANCIENT CEYLON

by him to have worn in former times a considerable amount of costly jewellery made of gold and gems, in the form of necklaces and bangles, but not anklets or nose ornaments (which Sinhalese also never wear) ; there cannot be much of this left among them now. They also had ear-jewels, set like those of the Kandian Sinhalese, in a large hole which is bored through the lobe of the ear and expanded to receive them, to a diameter of about three-quarters of an inch; some of them were made of ivory, horn, or bone, and were carved and etched. Brass ones are now worn. Sinhalese women have a cylindrical tube of silver, closed at the outer end and having a projecting rim at it; in this end are inserted pieces of red glass or garnets, round a central stone or boss.

Mr. Nevill also observed that when properly dressed in their villages both men and women adorn their hair with bright or fragrant flowers and leaves, and occasionally add garlands of flowers for their necks, red and orange being their favourite colours. I have noticed that Kandian girls do the same. He added that the Vaeddas also crush fragrant leaves and rub them on their hair, neck, arms, and breast. He learnt that the marrow of the Sambar deer (Rusa aristotelis] is applied about once a week to the hair, if procurable ; or the fat of the Talagoya or Monitor Lizard (Varanus dracaena), commonly called in Ceylon the ' Iguana/ is used for this purpose. He was of opinion that the number of split bones left by prehistoric people may be due to a similar custom.1

Dwellings.?Mr. Nevill states regarding the Forest Vaeddas : ' If possible, a cave is chosen for the home, and improved by a slight roof in front, if too exposed, and around this the food-winner ranges* during the rainy season, when the Sambar deer frequent the neighbourhood of the hills. ' A good cave becomes an hereditary possession. . . . Where an overhanging rock can be found, it is of course sufficient. Otherwise any rock is chosen, and some sticks being laid sloping from in front of it, it is roughly thatched with twigs, rushes, and large pieces of bark. A few elk [Sambar] hides, if ^ot bought up

1 The Taprobanian, Vol. i, p. 189.THE MODERN VAEDDAS

49'

i

m

by pedlars, will form a screen at one end. If it is only to exclude dew, a very few branches or bits of bark suffice.1

' In the dry hot months when brooks and ponds dry up,, the game collects in the low forests around the half-dried riverbeds. He then takes wife and children, aged parents, or crippled relatives, and settles them in a hut close to where water can be got. From this he makes his hunting forays,, and returns to it with his game.

' Besides his high-ground [cave] residence, and his low-ground residence, if a tract of forest burst suddenly into flower that attracts vast swarms of bees, or into useful fruit, the family will make a little picnic party, and go there for a week or a month, if it be too far from the home for daily visits. He cannot, however, be called nomadic/ 2

The houses of the Forest Vaeddas are flimsy, easily erected,, low rectangular huts or shelters under shady trees, built of thin sticks, and usually in a reversed wide V shape, without walls, though some have them. They have a covering of grass on the roof, or in default of it the skins of Sambar deer, or broad pieces of bark. The temporary huts of the Village Vaeddas are quite similar ; and their more permanent houses are also rectangular,3 with a low roof raised on walls which are covered with broad strips of bark, or have the spaces between the sticks filled with leafy twigs. A few fill up the walls with mud. Nearer the eastern coast, where suitable trees for barking are scarce or absent, they have only grass roofs, and leafy twigs are almost always employed for closing the spaces in the walk. Mh Nevill remarked that there is little cpfference between the homes of the Village and Forest Vaeddas except that the former makes his house sufficiently substantial to keep out rain as well as dew; and that he leaves his family at it, and does not usually take them to his temporary hunting quarters. The Wanniyas erect similar huts roofed with grass ;

1 Dr. Seligmann is giving a full account of the cave dwellings of the Forest Vaeddas.

2 The Taprobmnian, Vol. i,,p. 180".

3 With the exception of a few Tamil villages in the northern Province there are no circular dwelling-houses in Ceylon.

Ei'.;P ~ZiEZ

50 ANCIENT CEYLON

nearly all those I have seen had only walls of sticks, filled up with leafy twigs, but a few possessed mud walls?or rather, mud was used in them instead of the twigs.

Any bushes growing at the front of the huts are cleared away, so as to leave an open space under the trees, in which the occupants can sit, or lie, or cook, and peg out deer-skins for drying, or dry their surplus meat on a rectangular stick frame over a slow fire, this being a common custom of all hunters in Ceylon. They all abandon the site for very slight reasons, and establish themselves a mile or more away, often, in the case of those who cultivate millet, in order to be near the piece of ground which they are clearing for millet-growing, and at which, in any case, the men generally reside for some months in huts like those of the Forest Vaeddas, to protect the crop from Elephants, Deer, and Buffaloes.

Sometimes they form a new hamlet because they find themselves too near a road used by the public, or on account of an outbreak of sickness. In the latter instance it is thought that the old site was haunted by local devils who caused the disease. I have known the northern and north-western Kandian Sinhalese abandon villages for the two latter reasons, even when their huts had mud walls and raised earthen floors, which require much more labour to reconstruct.

Food.?The food of the Forest Vaeddas consists of fruits, roots of wild yams, and especially honey and the flesh of any animals they can kill, which are chiefly * Iguanas/ Pigs, and Deer. All the Village Vaeddas, and the Tamil-speaking Vaeddas (with the exception of a very few who are solely fishermen), and the Wanniyas eat the same food, and have in addition the small millet above-mentioned, called Kurahan by the Sinhalese, the Indian? Ragi (Eleusine coracana). This is grown in temporary clearings (termed hena in Sinhalese), made in the forest, all bushes and grass being cut and burnt off, but not the larger trees. After one crop, or sometimes two, have been taken off the ground, the clearing is abandoned, and allowed to be overgrown once more with jungle, and is not recultivated until from five to seven years have elapsed. In these clearings, which are exactly like those of the Sinhalese,THE MODERN VAEDDAS 51

are also grown a few red Chillies and Gourds, and sometimes a little Indian Corn, and a small Pulse called Mun (Phaseolus mungo). A very few Village Vaeddas and Wanniyas who live in suitable places for it grow and irrigate a little rice, which the Forest Vaeddas are now learning to cook and eat when they can procure it.

Mr. Nevill was informed that ' of all food the greatest delicacy is considered to be little bits of lean flesh, chopped up, and wrapped in fat of the Igilana, taken from the entrails apparently. This is broiled/ 1 The flesh of this lizard is white, and rather wanting in flavour, but not in any way unpalatable ; I have often eaten it when stationed in the jungle, and it is a favourite dish of the Kandian Sinhalese villagers.

Following the example of their Tamil neighbours, the Wanniyas and the Tamil-speaking Vaeddas do not eat Monkeys, which, however, form a regular item in the diet of all Vaeddas of the interior, and with the exception of the small brown Monkey (Thersites) are eaten by the majority of the northern Kandian villagers. The flesh is dark-coloured, and somewhat strongly flavoured; I have tried it more than once, feeling at the time that I was, as it were, the next-door neighbour of palaeolithic man, and practising something allied to cannibalism.

The Tamil-speaking Vaeddas informed me that they have no forbidden meats excepting the Monkey and some of the ' Vahanas * of their Hindu Gods, that is, the animals on which the Gods ride, such as the Peafowl and the Rat, the Vahanas of Skanda and Ganesa.

The Coast Vaeddas subsist on fish, in addition ; they alone catch them by netting or spearing them. Like the Sinhalese and Tamils of jungle villages, all are accustomed to capture fish in the dry seasons either by baling the water out of shallow pools, or by stupefying the fish by means of poisonous leaves or fruits thrown into the water. The crushed leaves of the Timbiri tree (Diospyros embryopteris), or the crushed fruit of the Kukuru-mahan bush (Randia dumetorum) t and also, accord-

1 The Tapfobanian, Vol. i, p. 191.52 ANCIENT CEYLON

ing to Mr. Nevill, the roots of a species of creeper called Kala-vael (Denis scandens) are especially used for this purpose.

Unlike the Low-country Sinhalese, they never fish with the hook, a peculiarity that they share with the Wanniyas and nearly all Kandian Sinhalese, who for some reason, unknown even to themselves, hold that it is quite improper to do so.1 Whether the Sinhalese name for fish-hook, bili-katuwa, the word Uli meaning also offerings made to devils, has had any influence, I cannot say; but the feeling may be connected with the fact that the north-western Kandians also think it a disgraceful act for a female, even though a child, to capture a fish in any way whatever. I have never been able to discover an explanation of this prohibition. Whatever the objection may be to the fish-hook, it is not applicable to the Tamils; I have seen Tamil women of jungle villages fishing with a line and hook, and proud to show the number of fish they had taken.

The millet is ground into flour on a flat stone, or in a quern by those who possess one, and is cooked by baking it inside a wood fire. The flour is first mixed with water on a deerskin or some broad leaves, into a stiff paste, which is made into a circular cake more than an inch thick and some nine inches in diameter. This is then covered on both sides with the large green leaves of the Halmilla tree (Berrya ammonilla). After the fire has burnt for some time, so as to contain a supply of redhot charcoal, it is raked away, and the cake is laid on the hot ashes, and covered up by more ashes and the burning charcoal, the heat of which in a few minutes is considered to have baked it sufficiently. The Wanniyas term this cake Alupota, * Ashes-slab' ; it is the Ginipuwa, or * Fire-cake/ of the Sinhalese hunters, who also make it. Mr. Nevill states that cakes are also made of the dried and ground-up seeds of the Tree-fern (Cycas circinalis) ; the ' cabbage/ or bud of unopened leaves at the crown of the wild Date (Phoenix zeytanica), is doubtless also eaten, as by Sinhalese villagers.

As in the case of all hunters, meat is cooked by broiling.

1 Plutarch mentions that the natives of Oxyrhynchus in Egypt did not eat fish that had been caught with a hook.THE MODERN VAEDDAS 53

The few who have rice boil it; being in the neighbourhood of Sinhalese or Tamil villages, where common pottery is obtainable, such persons are able to procure earthenware pots for the purpose. .

Including even the wildest Forest Vaeddas, all are accustomed to chew sliced Areka-nuts with Betel-vine leaves, when they can get them from other villagers. In default of them they (like the inhabitants of remote Kandian villages who are without them) use the leaves of aromatic herbs, especially a Basil, Tala (Anisochilus suffruticosus), and the bark of the Kaeppitiya (Croton lacciferum)?one of the bushes on which stick-lac is found?and other trees, among which Mr. Nevill includes the Demata (Gmelina asiatica) and Dawata (Carallia integerrima}, and the seeds of a Lac-bush (Gardenia carinata). He states that lime is sometimes burnt from shells of Cyclo-phorus involvulus, and taken with the barks as a luxury. Some Forest Vaeddas looked with suspicion on some cut tobacco which I offered them for chewing, and refused it, as they had not previously seen any like it; but they readily took the uncut leaf.

According to Mr. Nevill, ' they will drink the clear water in a natural [rock] cistern, but win not drink the clear water of pools in the bed of a river or in forest hollows. If water is wanted at a stream, they scoop a little hollow in the sand, where it looks clean and sharp, and wait until the water filters through into it. They particularly like water lightly tinged yellow with mud, called Borct-diya, and it is considered better flavoured and more wholesome than plain water. They will drink river water, unless it be clear and stagnant; and the clear water of streams, running, they also drink if there be no sand in their bed in which to* scoop a hollow. Stagnant clear water is considered very bad, in fact,, poisonous/ * Kandian villagers also prefer * bora-diya/ and the water of pools which are covered with a green vegetable growth. I have found this water always good and sweet.

Utensils*?At their dwellings the simple wants of these people are easily supplied. In some parts of the interior the 1 The Tapfobanian, Vol. i, p, 187.54 ANCIENT CEYLON

wilder Vaeddas have a few large hollow black shells of the hard fruit of a high tree which grows in the eastern forests, the name of which I omitted to note, slung by some bark strings for carrying. More commonly they use the shells of small Pumpkins, with a section cut off at the stem, similarly strung, and termed Panliya. These are about seven and a half inches in diameter, and are used for carrying water or honey (Fig. 18).

The only other household article that they really require is a bag, or perhaps two, made of the inner bark of a short slightly tapering length of the Riti tree, which is stripped off or drawn off in one piece, after being well beaten, and is sewn together at the larger end. This makes a strong and very durable bag, called a Riti-matta, which lasts for some years, and has almost the appearance of having been woven. One in my possession, blackened with age, is thirty inches long, ten and a half inches wide near the mouth when laid flat, and fourteen inches wide at the other end (Fig. 19). The bag is used for carrying or storing millet, or any other food. Some also make small baskets of the same inner bark. The Wanniyas and those who live near the sea have, like the Kandians, whole gourds (labba) for holding water, and also use common earthenware pots, obtained from Sinhalese potters, for cooking and for containing water. Mr. Nevill learnt that in ancient times the [Village] Vaeddas had household vessels made of copper and even gold, for holding water and for cooking, and he saw copper ones still in use. There is no probability that the wilder Vaeddas ever possessed such articles. Neither Vaeddas nor Wanniyas are acquainted with the art of making pottery, and certainly the former, and I believe also the latter, do not understand any form of mat or other weaving. Deer-skins supply the place of mats for sleeping on, or when preparing food.

The blades of axes and especially those of arrows answer all the purposes for which knives are usually thought to be indispensable. Those who cultivate millet or rice purchase for the purpose, by exchange of honey, meat, deer-skins, or horns, or beeswax made into thick circular cakes, the digging56 ANCIENT CEYLON

hoes termed by us ' Mamoty '?(more correctly, the Tamil word mm-vettei, earth-digging implement)?and by the Sinhalese Udaetta. For excavating purposes, such as taking up wild yams, or digging out of their burrows the Pangolin or Scaly Ant-eater (Manis pentadaotyla) and the ' Iguana, 'they, like the Kandian hunters, merely use a sharpened stick. All who make clearings for millet-growing buy the Bill-hooks (kaetta) which are used by their Sinhalese or Tamil neighbours.

Fke-maMng.?Fire, is commonly got by striking a spark with the aid of the axe, the word for it being gini-gahanawd,

* to strike fire/ A piece of flint and a little tinder are generally carried, or the latter is soon made from a bit of rag. But all Vaeddas and Wanniyas are also able and accustomed to obtain it by means of friction with two dry sticks. There are two ways of doing this. In one they use the twirling-stick, both races invariably turning it between the hands while the point rests in a hollow in a lower stick which is held on the ground by the feet. The expression used for this by the Vaeddas is gini-gdhen ginna gannawa, * to take fire from the fire-tree'; it is one of the very few alliterative sayings used by them or the Sinhalese, with the exception of simple duplicated words and the refrains of songs. The Vaeddas and Wanniyas use various woods for getting fire by this method, but Velan (Pterospermum suberifolium) is a general favourite.

The other method, which when practised with wood picked up in the forest is much more laborious, is by simply rubbing one stick across another ; the Wanniyas and Sinhalese express it by the verb mandinawd. Only extremely dry Velan wood is used for obtaining fire by this process, which, as the wood is probably even then not thoroughly dry, I was told sometimes occupies nearly two ' paeyas/ or forty minutes.1

This is the mode of fire-making employed by some tribes of Central Australia, but not other Australians, the edge of a

1 Dr. Schweinlurth, in. The Heart of Africa, 3rd Ed., Vol. i, p. 254, describes this method of obtaaniEg fire in the Higher Nile districts,

* the whole proceeding being a marvel which might well nigh eclipse the magic of my lucifer matches.' It is also practised in Senegal.

(CailMe, Travels through Central Africa, Vol. i, p. 123.)THE MODERN VAEDDAS 57

piece of wood used as a spear-thrower being rubbed ' backwards and forwards upon the shield ; in a short time the light wood is charred, then it glows, and with judicious blowing the glow is fanned into a flame/ *

This method of getting fire is found in Malayalam and Tra-vancore, the very district from which it is probable that the earliest 'settlers came to Ceylon. In Mr. Thurston's Ethnographic Notes of Southern India, pp. 468, 469, it is stated that fire is made by cross-friction by the Pulayans of Travancore and the Paniyans who live at the base of the Western Ghats of Malabar. He gives an illustration of two members of the latter race engaged on this work, which he describes as follows : * A portion of a bamboo stem, about one foot in length, in which two nodes are included, is split longitudinally into two equal parts. On one half a sharp edge is cut with a knife. In the other a longitudinal slit is made through about two-thirds of its length, which is stuffed with a piece of cotton cloth. The latter is held firmly on the ground with its convex surface upwards, and the cutting edge drawn, with a gradually quickening sawing motion, rapidly to and fro across it by two men2 until the cloth is ignited by the incandescent particles of wood in the groove cut by the sharp edge. The cloth is then blown by the lips into a blaze/

When no flint or chert is available, the Kandian Sinhalese also employ both processes, but naturally they prefer the twirling stick, which they always turn by means of a bow and slack string, using in the north either Velan wood for both sticks, or often the wood of the Lolu tree (Cordia myxa) for the lower stick, and Mayila wood (Bauhinia racemosa), which Is very hard, for the upper one, or twirling-stick. The use of

1 Messrs. Spencer and Gillen's The Native Tribes of Central Australia,

p, 586 ; The Northern Tribes of Central Australia, p. 619*

2 In the illustration only one man is doing the sawing work, while the other holds the lower stick. Captain Lewin described a nearly

similar method employed by the ' Chittagong Hill Tribes. A semi« circular groove was cut round a split bamboo, and a flexible strip of bamboo worked in it until the dust became incandescent.?Wild Maces of S. E. India, p. 207.

58 ANCIENT CEYLON

the bow for this purpose is one of the very few practices which

differentiate the Sinhalese from the Vaeddas. There is a third method of making fire by means of two

sticks ; in it the pointed end of one stick is rubbed in a long

groove made in the other. I believe it is unknown in Ceylon.

Dr. Guppy, who calls it' the Polynesian method/ saw it used in the Solomon Islands, and stated that * the friction in some three or four minutes produces smoke; and finally a fine powder, which has been collecting in a small heap at the end of the groove, begins to smoulder. After being carefully nursed by the breath of the operator, the tiny flame is transferred to a piece of touch-wood, and, the object is attained/ * Darwin also observed this mode of fire-making in Tahiti, and wrote of it, ' The fire was produced in a few seconds' ; he himself tried it, and found that it required * the greatest

exertion.'2 In the Eastern Archipelago A. R; Wallace noticed that

cross-friction was employed.3

Thus we find that fire may be obtained from two pieces of dry wood by three different methods : (i) by drill-friction of a point in a hollow, the mode most generally used, which is again subdivided into hand-drill friction, bow-drill friction, and cord-drill friction (as used for the sacred fire of Hinduism) ;

(2) by transverse friction of a knife-edge in a groove ; and

(3) ty longitudinal friction of a point in a groove.

The fact that even those Vaeddas who have seen fire obtained by turning the twirling-stick with a bow never copy this method, although they understand the action of the bow and the ease with which fire can be obtained by using it, shows how extremely conservative in their ideas such people are. When they had said of it to me, " The Sinhalese do it, but it is not our custom," there was an end of the matter, so far as they were concerned. This is exactly the way of the northern and north-western Kandian Sinhalese. When I asked one of the latter whom I knew well why he did not try the effect of

1 The Solomon Islands, p. 65. *

2 A Naturalises Voyage, ed. 1882, p. 409. 8 The Malay Archipelago, 5th ed. p. 325,THE MODERN VAEDDAS 59

manure on his rice field, which he complained 'was not very productive, he made the usual reply, " We are not accustomed to do it." After I had explained the matter further, and suggested an experiment on one small patch, he ended the discussion by remarking, " My father did not do it. Am I a better man than my father ? " l When this is the mental position of primitive races, it is clear that immense periods of time must be allowed for the development of the slightest and simplest advances towards civilisation.

Weapons and Tpols.?The weapons of all the Vaeddas and Wanniyas consist only of a diminutive axe (Fig. 10) and a bow and arrows, generally two in number according to Mr. Nevill, and rarely three among the former race ; but usually three among the latter people. Mr, Nevill had an axe that was two and a half inches wide and five and a half inches long in the blade; but some are much smaller than this. These axes have handles from eighteen inches to two feet in length, which are passed through a socket-hole in the head. Nearly similar tools are in general use by the Kandians, and are illustrated in a later chapter. Neither Vaeddas nor most Wanniyas carry knives, which Kandian Sinhalese find indispensable. The steel heads of these tools are obtained from Sinhalese or Tamil smiths in exchange for skins, honey, or meat.

The correct length for a Vaedda or Wanniya bow is considered to be a little more than the owner's own height, but there is no fixed standard, the length partly depending on the strength of the person who is to use the bow. Some considerably exceed their owner's height; but short ones are often preferred for use in thick forest, as being more convenient to carry than long ones. One Wanniya bow that I got (Fig. 16) is only four feet ten inches long,2 and the old man from whom I obtained it stated that he always used similar short ones;

* It is laid down in the Ordinances of Maim .(iv, 178) that a good man should always follow the path of his father and grandparents, so the attitude of the villager was quite correct.

2 Some bows of British archers in mediaeval times were only five feet long.

60 ANCIENT CEYLON

other men informed me that they preferred longer ones. The longest are perhaps six feet in length.

The Vaeddas of the interior make them of Kolon (Adina ?cordifoUa) and Kaekala wood (Cyathocalyx zeylanicus), split and thinned down to the required size, and also of Kobba or Kobbae-wael (Allophylm cobbe). The Tamil-speaking Vaeddas to the south of Trincomalee employ the wood of the Ulkenda tree for them. All are rough, round in section, and not always straight, and are without notches. They are not decorated in any way. The Wanniyas informed me that they only cut their bows during the south-west monsoon, as they have an idea, possibly well founded, that the constant bending and relaxing of the fibres caused by the strong winds of that season render the wood more elastic and tougher than at other times. The roughest sort of Sinhalese bow does not differ from that of the Vaeddas ; but others vary in the material used, the length and thickness, and in having elaborate decoration in coloured lac. The length is usually greater than that of the Vaedda weapon.

Mr. Nevill -states that pellet-bows like those of the Sinhalese, with two strings at the middle of which a piece of skin is fixed, are used by Vaedda boys for killing small birds;

When shooting, the bow is commonly held by the left hand, but occasionally by the right. Some Vaeddas and Wanniyas are also accustomed to shoot while sitting on the ground, holding the bow by the foot, between the big toe and the next one. This is chiefly, if not entirely, done in shooting animals at night when they come to drink at a water-hole.

The twisted inner bark of two or three different trees is used for bowstrings, or where they are available the exceedingly tough fibres found in the long narrow leaves of a rock plant called Niyanda (Sansievera zeylanica). Many Vaeddas of the interior employ for this purpose the fibres of the thin aerial roots of the Banyan tree (Ficus indica). The Tamil-speaking Vaeddas make use of the inner bark of a creeper called Gaera-vaela in Sinhalese or Tevalan-kodi in Tamil. The Wanniyas employ the inner bark of Velan trees. The string is some-times rubbed with the split fruit of the Timbiri tree, which isTHE MODERN VAEDDAS 61

said to strengthen it. It is permanently fastened to one end of the bow, and tied round the other when about to be used (Fig. ii*).

The shafts of the arrows (Figs. 12-15)l of all alike are made of small Velan saplings, thinned down to about half an inch in thickness, and the whole length is often three feet, but varies from two to three. A wide notch is cut at the butt end. Whether used by Vaeddas, Wanniyas, or Sinhalese, they invariably have flat, narrow, and elongated steel heads rounded at the points, without barbs. The Vaedda arrowheads are wider near the butt than those of the modem Sinhalese, and very slightly concave on the sides, but some ancient Sinhalese arrows nearly resembled those of the Vaeddas in shape. They vary from two and a half inches to nearly eighteen inches in length, the latter size being of course rarely used, and only for large game such as Elephants ; the smallest are required for Hares and birds. The usual length of the blade is four or five inches. A set of three which a Wanniya used varied from four to eight inches in length. The arrows have nearly always either three or four feathers, which in every case, even among the Sinhalese, are the primaries (or long feathers) from the wing of the Peahen.2 These are rarely fixed in slight grooves cut in the stem. Occasionally five ?feathers are employed, and Mr. Nevill stated that in some instances they are placed in a slightly spiral direction. The fine strings of bark which tie the feathers to the shaft or bind the shaft at the head are sometimes protected from wear by being covered with a hard gum. A Wanniya arrow in my possession (Fig. 17) is wrapped at the head with a thin strip of deer-skin in a spiral In former days, according to Mr. Nevill, pieces of the shells of River Mussels (Unio lamellatus

1 The arrows numbered 13 and 14 are in the possession of Mr. H. B. Christie, recently Provincial Engineer in the Public Works Department, and were obtained by him from Village Vaeddas, They differ from the usual Vaedda type shown in the other arrow heads, and Fig. 14 resembles those now employed by Sinhalese.

2 Used in India also. Kalidasa describes an arrow which Raghit used against Indra, as being * fledged with peacock's plume."' (Raghu-van fa, Johnstone's translation, p. 26.)

62 ANCIENT CEYLON

and U. marginatus) were used by some Vaeddas of the interior as arrow-blades ; and he observed that the Sinhalese who live in their district in the Eastern Province still term these bivalves ' arrow-head mussels/-1

Until recent times no ancient stone weapons or implements had been discovered in Ceylon, and it was therefore assumed that the aborigines were unacquainted with the art of their fabrication. It was thus with great interest that I learnt from my friend Mr. F. Lewis, of the Forest Department, that for a considerable number of years several types of primitive stone implements have been found in the Kandian hill-tract in Maskeliya, by Mr. John Pole and Mr. G. B. Gardner. Through the courtesy of these gentlemen I am able to append the following particulars of their discoveries.

Mr. Pole, who has recently published a short account of his collection in the Journal of the Ceylon Branch of the Royal Asiatic Society, writes that the first examples of these weapons and tools were discovered twenty-five years ago by himself and Mr. E. E. Green, who is now the Government Entomologist in Ceylon, on some hillocks at Imbulpitiya, near Nawala-pitiya. He states (in epist.): *I have collected within the last twenty-four years over a thousand of these stones, in all their fantastic shapes and material, and my conclusion is this : The men of this age arrived at no type of implement. They split the stone and made the implements they immediately required, from the shards as they split them off, according to their adaptability. A serviceable shard or flake was helped to an ''edge/' and when they found a " point" amongst their shards they chipped the sides to make the point more serviceable. There was no attempt at copying any known design; the material was too obstinate to allow this/ He states that he considers that the agreement of a few specimens with some primitive types is merely an accidental coincidence. Of course in these remarks Mr. Pole is referring only to the stone implements in his own collection.

In his paper on them he mentioned that similar flakes have been found in |he districts of Puttalam (that is, in the early 1 The T&probanian, Vol. i, p. 33.THE MODERN VAEDDAS 63

Nagadipa), and Hambantota in the Low-Country ; and Matale, Nawalapitiya, Dimbula, Dikoya, and Maskeliya in the mountain region. He added that the Drs. Sarasin of Basle also discovered some in Uva, and in caves at Nilgala; and from Mr, Pole and Mr. Gardner I learnt that Mr. J. Still, formerly Assistant to the Archaeological Commissioner, Mr. H. C. P. Bell, met with some made of quartz and chert in the North-central Province.1

In a letter to the Ceylon Observer, dated August 8, 1907, Mr. Pole remarked of the makers of those found by him on the mountain ridges in Maskeliya, ' These people must have lived in cordons of single families, for they must have entirely occupied the vantage points of every spur of our mountains. Not many flakes are found in the flats.'

He expressed the opinion that * There was never more than a single family in one spot; there was no village artizan, around whose domicile a larger number of shards and debris might be found. ... Each man was his own armourer, found his own quartz stones, smashed off his own pieces from the native rock, just as he was able ; made use of the most serviceable by coaxing off from them some extra thick edges. . . . There was no getting rid of an obstinate angle in the stone fractured. There was no subsequent rounding off of the edges and corners.

f As to shape and size, they took the chance chippings of the stones, satisfied with the natural fracture, and worked the blunter edges. The material was quartz of various formations and compactness ; some as clear as glass, some clouded and milky, and others of a granulated structure/ It appeared to him that the stones * had been brought from great distances ?for although there is scarcely a ridge up-country on which no " flakes" occur, it will be found that the material of which they are formed exists nowhere in the vicinity.' He added that there could hardly be any distinct classification of the stones excepting into those with points and those with edges.

1 Dr. C. G. Seligmann informs me tliat he also found worked quartz chips under the earth in some caves now or formerly occupied by

Vaeddas, and he has since published an account of them in * Man.'64

ANCIENT CEYLON

Mr. G. B. Gardner, of Belihuloya, had, however, a different experience, and all the worked stones which he discovered seem to have been arrow-heads. They were lying on the ground on the summit of a hill. He states (in epist.}: ' They generally consist of a natural stone of the right size with one natural edge, and the other edge seems to have been chipped to make it double-edged/ These stones were not of the type of any of those found by Mr. Pole.1 They were roughly triangular, and were notched on both sides near the butt end for tying to the shaft. He noticed the resemblance between them and others in his collection, from Arizona and New Mexico, made by the Red Indians within the last forty years.

Specimens of all the types were sent from Ceylon to Mr. Bruce Foote of the Indian Geological Survey, and were reported by him to be identical with those which he has found throughout Southern India; he considered them to be of Neolithic age. According to Mr. Pole, the Drs. Sarasin (who, I believe, had not inspected those found by Mr. Gardner) thought them to be palaeolithic, and of the ' Madeleine ' or Magdalenien period?the time when the Mammoth and Aurochs and Reindeer were hunted in France and England.

Any doubt as to their date which these conflicting opinions might leave has now been definitely removed by the high authority of Mr. C. H. Read, of the British Museum?to whom I submitted Mr. Pole's tracings of typical examples of the stones and Mr. Gardner's drawing of the type found by him? and who has been courteous enough to furnish me with the following expression of his opinion of their age : * I should think there can be no question that the age of the stone implements is either neolithic or relatively modern. These stones seem to me to have much the same relation to the Vaeddas as the stone implements of North America have to the existing Red Indians/

1 I am indebted to Dr. Seligmann for a cutting from the Ceylon Observer (weekly edition) for March 5, 1909, in which Mr. Pole gave an account of the discovery of numerous flint implements and cores in a cave on Scarborough Estate, in Maskeliya. Among them was one 'beautiful example3 of an arrow-head, but of what type is not stated.Knives and Scrapers

Quart zite \ implements' di*covered bjr M* John Pole.

Arrowfwodl Gardner

FIGS* 20-34. Stone Implements. (Half Scale.)66 ANCIENT CEYLON

Through the kindness of Mr. Pole and Mr. Gardner I am able to supply illustrations (Figs. 20-34) of & typical series of these articles which will indicate their shapes and character better than attempts at description. Mr. Pole was good enough to send me tracings of many of his * finds/ and Mr. Gardner gave me a sketch of an arrow-head. Their extremely rude nature Is quite evident.

Notched arrow-heads have been found in England (rarely),1 and in Neolithic Lake Dwellings in Europe ? Switzerland, France, Italy ? with a slight broad stem or' tang ' at the butt,2 and also in Egypt 3 and Japan 4 ; but chiefly in North America, where many types with a straight or very slightly curved base or butt end like those of Ceylon have been obtained. These last are all illustrated by Mr. Gerard Fowke in the. Thirteenth Annual Report of the Bureau of Ethnology, under the heading 4 Stemmed Flints.*

After the later settlers, whether Nagas or Magadhese, introduced the arts of smelting and working iron,5 the Vaeddas would find little difficulty, in the accessible districts, in obtaining steel axes and steel arrow-heads, which they still continue to procure by barter from the Sinhalese or Tamil smiths.

Since no stone axes have been discovered in Ceylon, it is not certain that the prehistoric Vaeddas made them. Doubtless serviceable articles of this nature would be much more difficult

1 Jowitt. Half-Hours among some English Antiquities, p. 45,

Fig. 61.

* Dr. Munro. The Lake Dwellings of Europe, pp. 65, 103, 268.

* Seton-Karr. Report of U.S. National Museum, 1904, Plates

2» 4, 9*

* Sir John Evans. Ancient Stone Implements of Great Britain, p.

2f. , ;

5 I have met with, low mounds and scattered fragments of refuse from very ancient smelting furnaces in three or four places in the Northern Province, but I think not elsewhere. All trace of the furnaces had disappeared. To the best of my recollection all the heaps were in uninhabited forest in places where Mdney iron abounded on the surface of the ground. There were soiree small fragments of a very rough type of pottery mingled with the refuse in at least one of the heaps, but nothing else to assist in. determining the age. The pieces of refuse resembled black slag from English smelting furnaces, and not the scoriae rejected from forges.THE MODERN VAEDDAS 67

to fabricate than the simple tools shown in the illustration. The axe seems to be more indispensable in Ceylon than the arrow; both Vaeddas and Wanniyas (and I may add Sinhalese hunters also) are accustomed to procure a supply of food by its aid, without employing the bow and arrow. It is difficult to comprehend how the aborigines could exist in the wild forests of Ceylon without it.

If they did not make stone axes, it is just possible that in some way or other the primitive inhabitants may have been able to procure metal ones. If so, they must have got them from India, as it cannot be assumed that the Nagas, who may have made them at a later date, arrived in the country until several centuries had elapsed after the coming of the Vaeddas, otherwise they would have occupied a greater portion of the island. There may have been a trade in such articles at an extremely early date. Iron or steel weapons and tools of various kinds were in common use by the Aryans in the early Vedic times, and it is possible that their manufacture may have been understood in some part of Southern India also, in the second or third millennium B.C. The Vaedda word for an axe, gal-raekki, in which the first half of the compound means ' stone/ appears to refer to the sharpening of the weapon on a stone, according to Mr. NeviH's information,1 and until some examples have been discovered in Ceylon it cannot be accepted as affording any proof of the employment of stone axes by the first comers, no tradition of their use having survived.

The earliest Sinhalese iron or steel axes that have been found in Ceylon, apparently belonging to the second or third century B.C., are mere socketless * celts/ They are described and illustrated in a subsequent chapter. They are of a shape which was found elsewhere in the later Neolithic period, the polished-stone age. Although such tools must have been in general use by the Sinhalese from the time of the arrival of the first Gangetic settlers, the fact that only two examples of this form of axe hav4 been discovered, and that by the mere accident of the excavation of a deep channel at Tissa through 1 The Taprobanicrn, Vol. i, p. 189.

68

ANCIENT CEYLON

a hollow in which the refuse of an artificers' settlement was deposited, shows how unsafe it would be to assume that tools which have not "been found yet have never existed. All later types of axes found in Ceylon are removed by centuries from this primitive form. If, therefore, of the immense number of iron or steel axes that were used in clearing away the forests throughout all the civilised districts of Ceylon for probably more than five hundred years, only two examples have been met with, it may easily have occurred that the dwellers in the forests had axes, either of stone or iron or some other metal, of which no specimen has yet been seen by us.

A few Wanniyas and Village Vaeddas who can afford to buy guns now use them in the dry season, when the rustling of the crisp leaves that cover the ground at this time renders it difficult to approach game without being observed.

Some of the Coast Vaeddas, but no others, have an iron-headed spear or harpoon for catching fish, but I have not examined one, though I have seen them using it while wading |n the brackish or salt water of the lagoons near the sea. As &o other form of spear is employed by Vaeddas, they may have learnt its use from their Tamil neighbours in comparatively recent times.

Bathing. ? Many of the Forest and Village Vaeddas do not bathe. One man stated that he caught cold after the only bath he ever took, and therefore he had abandoned the practice as too dangerous. It will easily be understood that many of them are not very cleanly in their persons. A gentleman in the Survey [Department who had occasion to make use of some of them as guides in the forest informed me that they appeared to spend most of their spare time in pursuit of their insect cdmrate ; tbe& appear to have been unfavourable specimens of their race. It is also a common recreation of Sinhalese villagers, especially females, and is looked upon as an exhibition of disinterested 'friendship, to institute a searching examination of 'the hea^fe of their friends for this purpose.

is- ? especially as hunters in thick forest that

the Forest -saaA- Village Vaeddas and Wanniyas are distinguished, and in ibis respect they are exceedingly skilful, if notTHE MODERN VAEDDAS

69

altogether unrivalled. Lazy and inexpert as they seem when idling about their houses, the rapidity with which they can pass like shadows through thick jungle, without making the least sound, is astonishing. They have assured me that when the leaves lying on the ground are not too dry they can steal up to any animal in the forests without rousing it, and kill it while asleep, or at the least give it a mortal wound, with the sole exception of the Peafowl, which is too wakeful to be caught in this manner. Living in woods frequented by Elephants, Bears, Buffaloes, and Leopards, they state that they have no fear of any beast that the forest contains ; and judging by my own experiences when in the forest with some of them, I should suppose that in any ordinary circumstances they could escape from any of the three first-mentioned animals with ease; the Leopard does not attack them. Occasionally, however, a savage Sloth-bear (Ursus labiatus) mauls them when met face to face at a sudden turn in a narrow jungle track.1

A Vaedda once related to me a story of an incident of this kind, which cost him the loss of half a finger. On rounding a corner in such a path he found himself dose to a Bear ^hich immediately attacked him, knocking him down and endeavouring to seize his face. He described vividly how he felt its hot breath on his face as he caught its open jaw with both hands while he lay on his back, with the Bear standing over him. He succeeded in holding it thus for some minutes, in the meantime getting half his finger bitten off; and at last by a great effort he threw it backward and sprang to his feet. Luckily for him, the Bear thought the adventure not worth pursuing, and did not renew the attack, T ut disap| eared in the jungle.

On another occasion a Village Vaedda was assaulted in the same manner by a Bear, and came out of the encounter much more seriously injured, being badly bitten on the arms and head. He told those who found him lying on the path and carried him home, how he heard a loud report whilfc the Bear

1 I have seen Kandian villagers who have been frightfully injured by these Bears. In one case the whole side of the face was bitten away.TfjT

'"1

7o ANCIENT CEYLON

was worrying his head; this was caused by the fracture of his own skull by the animal's teeth. He was seriously ill when the account was given to me, and I did not learn whether he succumbed to his injuries or not. The way in which these jungle-dwellers recuperate after extremely severe injuries is sometimes surprising. I have known a Kandian recover under home treatment by a village practitioner or ' Vedarala' when his thigh was half cut through in the middle and the bone exposed, by his falling backwards across a razor-edged piece of newly-blasted granite.

While engaged on a hunting expedition, these hunters?

and Kandians likewise?glide along in single file, avoiding

every leafy twig the rustling of which might betray their

presence, or if game be near holding it until the next man can

take charge of it, and hand it over in the same manner to the

man behind him. At such times all tread in the footprints

of the first man, who when putting his foot on the ground

Ajft glides his toes along it in order to push aside any twigs

W1^^ if crushed. Their eyes and

jears are MJy alert to catch the slightest sound or movement

amoijg the thick jungle around them. With a lifetime's ex-

pepen ^ and hereditary perceptive faculties to assist them,

the secrets of the deepest forest appear to them as an open

book which they read as they pass. They hear sounds and

see object that to a person whose perception is dulled by

^ well be altogether absent, so far as his

-ppfer o| observation is concerned Their trained ears detect

iiii^^;|^lis»3I.,o£ the jpild forest animals walking through the

!;j|||^^ .and 'can distinguish even

ll^ which,is..quite inaudible

Stci^^^ any ..uncertainty' exists

,;refi|^^ or. taeel down .with one. ear on

.. l^^^&^^^^^^i^^^^^: up /their,, doubts.,,. When they are

in seiir^p^ with keen sight, they hide

tb^r-diC^ round their waist-string,

TMs?? certain ;;^^:;.|liap,,,a.very wild appearance^ but there

is ao triistwcKflaf show that it was the primitive

.'dress of the aborigines of Ceylon*!THE MODERN VAEDDAS 71

Wild honey being one of their favourite foods, their vision and hearing are trained to an astonishing quickness in detecting every Bee that flies across their path, and noting its species, and whether it is flying laden or is only in1 quest of food. When it is carrying a load of honey and flying straight through the trees, they at once move off in the same direction, if it be the season in which the hives contain honey, that is, August and September, knowing of course that the laden insect makes a direct flight to its hive?the proverbial bee-line. As the nest is approached other Bees are seen converging towards it, and in a few minutes it is certain to be discovered.

Four species of Bees are found in the forests of Ceylon. The greatest one, a giant among Honey-bees and as large as a Hornet, is called the Bambara, its hive being a Bambaraya. It hangs an immense white comb longitudinally under a substantial branch of a tall tree, or high up in the face of a cliff, sheltered by overhanging rock. The Wanniyas have a belief that the next species of Bee does not permit the Bambara to make any part of its comb on the upper side of the branch. If it did so, the Dan Iuwaella would carry off the honey in that portion, the right to place any above the branch belonging to it alone.

The largest of these combs is about five feet deep, but somewhat less in length. The comb is without any cover excepting that provided by the bodies of the Bees, which usually cluster thickly over it, and completely hide it, thus protecting it from both sun and rain. The honey is chiefly used medicinally by the Sinhalese, but for the Vaeddas it is am important addition to the dietary. An old Wanniya once told me, as a good joke, that when moral pressure was put upon him by a Ratemahat-maya, or principal district chief, in order to make him supply some honey, he took care that it should be of this kind, and after receiving the thanks of the chief, who anticipated some pleasant eating, decamped before it was tasted. " I was never ordered to bring honey again/* he said, with a chuckle.

Mr. Nevill noted that to get this honey when the hives are attached to rocks, the Vaeddas sometimes descend from above by long frail ladders made of cane. These swing about in anANCIENT CEYLON

alarming manner, rendering the task a very dangerous one, I

especially at night. In order to appease the Spirit of the rock, |;

called Kande Yaka, 'the Demon of the Rock/ and induce f

him not to cause or permit the climber to fall, they sing songs loudly [I presume in his honour] while engaged in the work. Before undertaking the task a song is also sung, and a little honey sprinkled, to propitiate the Spirit.

The next Bee in size, called Darwjuwaella, its hive being the DanduwaeUaya, also hangs a single uncovered small white ;

comb vertically under a branch, but never under a very high ?

one; it is commonly found in a low bush. A very small j

portion of the comb is always constructed on and round the \

uPPer side of the branch also. No larvae are placed in this part, which is reserved for storing honey. This Bee protects the comb in the same manner as the large species. The honey is clear, rather pale-coloured, and sweet; and is eaten by all who find it. As in the Bambara's comb, the cells are on both sides of the comb, the more advanced larvae being in the outermost cells; these are often separated from the rest by one or two rows of short empty cells. In the middle portion of the comb the largest larvae are found round the centre. The largest combs I have seen were from twelve to fourteen inches wide and deep.

The third kind of Bee makes its hive in hollows in trees. It is termed the Mi-maessa, the ' Bee-fly/ the hive being the Miya, and it bears a dose resemblance to the common Honey- t

bee of Europe. The honey is darker coloured but perhaps sweeter than that of the last species. This is the kind that is specially searched for by the hunters; as there are many combs in the hive, of course much more honey is obtained froM it tiaan from the single comb of the Dan JuwaelIa, and eight or evm ten quarts of honey are taken from a very good Mve. The Forest Vaeddas are said to still occasionally pre-serve surplus meat in this honey, placing it in the hollows of trees, which they fill up with honey, and afterwards closing the orifice ^Rth 'day or wax.

The last Bee is an interesting one called the Ku Ja-Ml-niaessa, the * Small Bee-fly/ no bigger than a small House-fly,THE MODERN VAEDDAS 73

»

which at a first glance it somewhat resembles. It is a fat-bodied little insect, less than a quarter of an inch long, and is extremely tame; when one perspires with the heat in the jungle several of them often alight on one's hand to drink the moisture.1 It is black in colour, as are also its comb and honey. There is very little of the latter in a hive, but it is the sweetest of all. The nest is often found in a small dead branch or stump; and the entrance is built up with wax so as to leave an orifice sometimes not wider than the lead of a common pencil, barely permitting the insect to enter. The Wanniyas consider this Bee to hold Mgher rank than the others, notwithstanding its diminutive size; it is the Himi, ' Lord/ of the Bees, because, they say, its hive is sometimes established at a higher level than those of the other kinds.

The honey of the two last-mentioned Bees is procured by enlarging the entrance to the hive, or cutting a new one, with the little axe which these hunters always carry by passing the handle downward through their cloth belt. The work is easily done, as the stings of the Bees are ineffective and rarely cause injury; in fact, they are not often inflicted upon the hive-robbers ; but the Bambara is a dangerous insect when the community is aroused?there being often several combs in proximity on the same rock?and its hive can be cut down only at night, after stupefying the Bees with a smoking torch on which resin has been sprinkled.

Unless the wax be required for household use or barter, the finders divide and eat it, and everything taken out of the hive, excepting only the full-grown young which crawl out of their cells in time to escape this fate; all the rest of the larvae, however much developed, being thought to be little, if at all, inferior to the honey, and having, as a Kandian assured me, ** a pleasant flavour like milk/'

In districts where there is suitable forest, the Kandian Sinhalese make exactly similar, but temporary, excursions in search of honey, and are fairly expert in observing the Bees, without which they could not expect to meet with any success.

1 I met with, a similar Bee, which was equally tame, in the Gambia district in West Africa.V ' .l

ANCIENT CEYLON

Among some Sinhalese it is a custom for the man who discovers a hive which he intends to take afterwards, to make a cut with Ms axe on the stem of the tree; the honey will not then be removed by others. It is believed that if more than five cuts were made the Bees would abandon the nest. While on such expeditions in one northern district, it is a significant fact that they still address each other as " Vaedda."

All the forests and jungle where the hunting races live are apportioned among them for the purposes of hunting, getting honey, taking fish, and collecting shed deer-horns ; and they informed me that they respect each other's rights over them. When I was out in the forest with some Wanniyas on one occasion, one of them observed a half-broken twig hanging at the end of a small branch?a common hunter's mark in the jungle?and remarked at once that somebody had been passing through their forest, which was a wild tract far from villages. It was evidently a matter which caused them considerable misgiving, and they discussed it long and eagerly, and eventually agreed that it was done by a certain person of another hamlet, who was known to them as an unscrupulous character. " It must have been Tikka," they said, " he is a bad man; no one else would do such a thing. He has been collecting some of our horns."

It is well known that Deer shed their horns annually. At the season when they are dropped the hunters wander about in the forest in all directions in search of them, knowing that they are useful for barter at the little roadside ' boutiques/ or shops, which they visit in order to procure cloth, salt, etc. It is somewhat strange that many horns are found badly gnawed, sometimes more than half through ; this is said to be done by Porcupines. The work of collecting the horns is laborious, and to our minds would not appear to be worth the little that their finders obtain for them. A Wanniya who was with me carried for three days, at considerable inconvenience, a small gnawed horn for which he only expected to receive a penny. After reaching his home he would still have a journey of eight miles in order to dispose of it, but probably he would carry some honey or other horns with him. At any rate thisTHE MODERN VAEDDAS 75

work would appear to be performed without much danger ; but I have known a man when so engaged to be attacked by a Sam-bar deer, which knocked him down and broke his collar-bone.

The Vaeddas, and also Kandian hunters, usually go on hunting or honey-collecting trips for a few days at a time; but the Wanniyas are absent in the forests for about two months together, returning home at intervals in order to fetch a little millet-flour, or to leave horns, skins, or honey* They take with them as food merely a small bag of millet-flour. When other food fails they cook the large cakes that have been described above, one of them sufficing for a day's eating. Of course the wilder Vaeddas who do not cultivate milet are without this resource, and live entirely on the forest products and animals at these times.

The Vaeddas are sometimes reduced to starvation if continuous rain fall while they are distant from their home on these trips. At such times, they informed me that they seek a large Riti tree (the bark of which is easily detached in large pieces), and immediately make a long cut across it with an axe, near the foot, and from each end of this a vertical cut of about their own height, or a little more. The piece of bark within the cuts is then lifted off the tree at the lower end, and supported at the loose corners on two sticks set in the ground for the purpose. This makes a tiny watertight shed under which a man can sit and .sleep while the rain lasts. I was assured that sometimes they have been obliged "by bad weather, when the forest streams were impassable, to remain in such a shelter for three, and in extreme cases even four days, without food. They axe so well inur&| to privation of this kind that they seem little the worse for it, in the opinion of the Sinhalese who know them best. They remarked that they had never heard of a Vaedda's dying of starvation.

When I was in the forests for several days with a party of Wanniyas, a heavy rain-storm came on in the evening, and lasted all night. Using my breakfast-cup as a gauge, I found that the fall amounted to more than three inches. It was an awkward predicament,, as we were quite without shelter, and were merely camping under trees. There were no RitiANCIENT CEYLON

trees in that part of the country, but the hunters were equal to the emergency, the threatening appearance of the sky having given us warning of the approaching storm, which many earnest supplications addressed to one of their special Forest-Deities, the Sat-Rajjuruwo, the deified King Maha-Sena; had failed to avert, though accompanied by abundant offerings of leafy twigs hung over the horizontal stems of suitable creepers.

They scoured the forest all around until they found a tree with a large hollow up the trunk, at its foot. Dried wood was collected, and hastily crammed inside this shelter. Then a fire was made round another large dead tree, which soon became-ablaze, and the whole night's rain failed to extinguish it. It will be observed that the fire was taken to the wood, and not the wood to the fire; this is a hunter's custom; a hunter always makes his fire close to a supply of dry wood. Round this tree we all camped, the men lying on improvised beds of small leafy twigs which kept their bodies off the wet ground, while I was in a hammock, between two blankets, out of which the water was wrung in the morning. When the rain at last ceased at 7 a.m., the dry wood was brought out of its hiding-place, and a roaring fire was made at the burning tree. This soon warmed us, and thoroughly evaporated all the moisture in my clothes?no one else was much overburdened with such articles?and the drenching had no injurious effect on any one.

When animals have been wounded by their arrows, the hunters track them through the jungle until they find them exhausted or dead. Elephants-are killed by means of heavy anxyws' With the eighteen-inch blades. These are driven into them behind the shoulder at very close range?a distance of two or three yards?and as both edges of the blade are sharpened, every branch touched by the shaft of the arrow as the animal rashes through the jungle causes it to enlarge the wound, until the loss of blood is so great that the Elephant is exhausted. An old Wanniya, Kona by name, told me that he had killed mnt' Tusk-elephants in this way. Sambar deer are taken in a similar manner. Deer and Pigs are often killedTHE MODERN VAEDDAS 77

on moonlight nights, while drinking at small pools in the forest, the hunter sitting behind a low bush or a small shelter of leafy branches made on the leeward side of it. It is on such occasions that the bow, if a very strong one, is sometimes held by the foot.

By the Wanniyas, at least, if not also the Vaeddas, the flesh of the Pig is never removed until the epidermis has been scorched off by fire. On one occasion when one was shot they refused to cut it up until this necessary preliminary work had been done. " Whoever heard of cutting up a Pig before the skin was burnt off," they said ; and I was obliged to wait and watch the proceeding. As the Pig is considered to be an 'unclean' animal by the Kapuwas, or demon-priests, in the Vaedda districts, there may be some idea of first purifying the meat by the application of the great purifier, fire, before taking it away. A fire is made against one side of the animal until it is charred, after which the body is turned over and the other side, and, in fact, all parts are equally burnt, firebrands being applied to the legs. The skin is then easily removed by scraping it with sticks.

Kona was quite an original character. I never saw him sleep in the ordinary way; he merely sat with his back in a comfortable position against a large tree, and he seemed to obtain a good night's rest in this manner. He hobbled about with a bent back, and supported by a long stick, and appeared to be quite incapable of any useful work; but as soon as we began to make our way through the bushes he took the lead and kept it, at a pace that was almost too rapid for me. He knew, he said, every rock and game track in the forests in which he and his friends were accustomed to hunt, and his opinion was always listened to with respect, and his advice followed.

The idea of locality of these hunters is perfectly developed. On one trip I was taken by some Wanniyas through a piece of wild pathless forest ten or eleven miles across, near Pada-wiya tank, at the north-eastern boundary of the North-central Province. The jungle was dense, and the journey therefore occupied all day. Of course we were unable to proceed in aI -

ANCIENT CEYLON

straight line, and more than once we deviated into a right-angle from our proper direction in order to avoid thorny jungle that was said to be in front of us. At about one o'clock we came to a high rock, as they had promised, on the top of which good rain-water is always retained in a hollow. There we cooked and ate some food, after which we resumed our tramp. In the middle of the forest, as we were proceeding along a deer-track, one of the men drew my attention to a half-broken twig hanging at the side of the path. " I broke that two years ago/' he said ; he was then proceeding at a right-angle from the line we were taking.

When I asked him if he never lost his way in such thick forest, full of undergrowth, he at first could not understand my meaning. After I had explained it?feeling while doing so that I was making an interesting exhibition of my ignorance? he laughed consumedly, and thought it a capital joke. " How can one lose it ? " he said. He had never heard of such a thing before; to him it appeared to be quite impossible,1 apparently as much so as getting lost in an open field would be to us. " When we look at the sun we always know which way to go," he remarked. The men justified my confidence in their powers by emerging, just before dusk, at the very spot where I wished to arrive, many miles from the homes of any of the party. Those who had acted as guides lived some twelve miles or more away, by the nearest footpath; and the house of the man who lived nearest was five miles from the point where we left the forest. I have always thought it a very clever feat.

There can be no doubt that something more than the mere sight of the sun is necessary when one is in the midst of such thick leafy jungle as that of Ceylon. Accompanied by two Kandian trackers, I once followed the tracks of a ' Rogue-elephant * that I had alarmed, for more than half a day, in thick forest, ending, nearly at dusk, seven miles from my quarters by the shortest path; and nothing would convince

1 A hunter near Benin, in West Africa, stated that l it was quite impossible for him to be really lost in the forest.* (Roth, Great Benin, p. 144.)THE MODERN VAEDDAS

me that we were not returning in a diametrically wrong direction out of the jungle, until we got into a path which I recognised. I was then no longer inexperienced; I had lived for several years in jungle stations, and had been accustomed to jungle shooting and elephant tracking. The men who were with me could not possibly be acquainted with the part of the forest where we ended, as it was eight or ten miles from their village, and was totally uninhabited; yet they understood their position perfectly, and rightly decided that if we adhered to a game-track it must lead us to a village tank which they knew.

Progression in the right direction in open forest is a simple matter; it is different when one is in the midst of thick leafy jungle. Some in Ceylon is so dense and full of leaves that it is no exaggeration to say that an Elephant would be invisible at a distance of six feet; and in one case I was charged by a Rogue-elephant which I could hear approaching but of which I could not get a glimpse until his head was ten feet from me. I can recommend such an experience as a good test for the nerves. In this instance, a Kandian young man of about twenty years of age, who at his earnest request had been allowed to accompany my two trackers, was so overcome by fright that he stood perfectly still, paralysed and speechless, with wide-open mouth and staring eyes, and shaking all over more violently than the proverbial aspen. I have also seen a "Moorman/ perhaps thirty years old, in exactly the same state under similar circumstances. Some minutes elapsed before they recovered the power of speech. Of course all the forests frequented by these hunters are not so dense as this ; some of the high forest is comparatively open in parts, and they avoid the thicker jungle.

As illustrating the observant nature of the Yaedda, I may mention that I once showed some Village Vaeddas who lived far from others in the forest the illustrations, the first they had ever seen, in a copy of the Graphic, among which was one representing the landing of some troops from, boats. They understood the scene immediately, one of them having once seen some boats at the coast, he said; and to my surprise a

8o ANCIENT CEYLON

Vaedda remarked that the persons in the background who appeared to be smaller than the rest must be at a greater distance than the others. He explained that they had noticed that the more distant objects always seemed smaller than those near at hand.

On the other hand, when I exhibited a drawing in the same paper to a learned Buddhist Abbot or Anunayaka, who lived at a remote temple, and was deservedly respected by all, and well acquainted with the Pali and Elu (old Sinhalese) languages, he said, regarding the more distant persons, " I suppose those men are a smaller race." The Buddhist scholar, deeply versed in the classical languages of his country and intimately acquainted with the abstruse philosophy of. his religious works, who, in fact, was then about to found a small college for training Buddhist monks, was surpassed in intelligence by the Vaedda, who had never looked inside a book in his life, perhaps had never seen even the outside of one before.

A road was opened near the hamlet of these Vaeddas, and when I passed that way again and wished to renew my acquaintance with them, I found that they had withdrawn some miles further into the forest to avoid the publicity thrust upon them. The Wanniyas believe that when the Grey Mungus 1 (Her-griseus), which they term the Nay Mugatiya, or ' Cobra Mungus/ meets with a Cobra that it is afraid to attack or which has attacked it, it goes off in search of a White Mungus or Eli Mugatiya, which is said to be a very small and rare and fetches it to the scene of combat, where it pays to it, bowing down before it. Fortified by the presence and authority of this superior animal, the Cobra Mungus at the Cobra and kills it, after which it and any proceed to eat the snake, the White Mungus, however, no in the feast.

The Wanniyas closely resemble the Kandian

as their intelligence. The instances I have

are of the amount of mental quickness shown

1 It is to the name * Mongoose * or * Mungoose *;

tike is Mungma,THE MODERN VAEDDAS

81

by the Village Vaeddas with reference to subjects with which they are acquainted.

The Vaeddas and Wanniyas bear the character of being thoroughly honest, and they are said to be faithful in their marriage relations. Unlike the Kandian Sinhalese, they are strict monogamists, and do not practise polyandry, according to my information ; and the former, at any rate, are reported to be good to their wives according to their ideas. I have no reason to doubt that the same can be said of the Wanniyas.

They are quite as lively and ready to enjoy a small joke as the Kandian villagers, but there is not much to amuse them in their forest life. While the Vaeddas often dance and sing on suitable occasions this does not appear to be a trait of the Wanniyas, who thus resemble the Sinhalese villagers as regards the former amusement. Fortunately for them, they are not exposed to the temptation of drinking alcoholic liquor, and probably not one of them knows the taste of it. Crime is practically non-existent among them all.

With respect to their truthfulness, of which Mr, Nevill had a very high opinion, my own experience is that although they are generally truthful, many individuals are prepared to deny a knowledge of facts of which they are fully aware, when to do so suits their convenience for the moment. In this respect they are like the Sinhalese villagers, so far as concerns their dealings with strangers. They will not work for hire except under the compulsion of hunger, and they might thus be thought lazy by those who see them idling about their huts at times when they are not engaged in hunting. But their active life at other times, when they are out in the forests, entirely disproves it.

I found them all converse readily with me, without any appearance of the fear, or hesitation, or shyness that one often notices in Kandian villagers. Many Forest Vaeddas have loud harsh grating voices. I was told by those who knew them well, and I observed the same peculiarity in those I met, that under ordinary circumstances, as well as inside their dwellings, the conversation of some of them is carried on in an extremely loud tone, the people almost shouting at each

Gg2

ANCIENT CEYLON

that they appear to strangers to be in a towering el otte when in reality they are having a

a are said to be fairly healthy ; but owing

to want of good drinking water, in very dry years outbreaks of dysentery sometimes occur which carry off many of them as well as the adults, who also suffer considerably from malarial fever and the peculiar disease called ' Parangi Leda/ allied to the West Indian ' Yaws.' .

Every race has its own etiquette. When visiting an ancient abandoned reservoir in the forest with some headmen who knew the Vaeddas well and could speak their dialect, I once offered the usual ' chew' of Betel-leaf and Areka-nut to two wild-looking Forest Vaeddas whom we met there. The elder man said immediately, "What is there here for me to take to my wife ? " and refused it ; but he accepted the offer of a whole roll of the leaves and an adequate accompaniment of the nut. It was explained to me that everything they receive is invariably shared with their wives. They expect, therefore, never to be given less than a handful of anything, and to present a smaller quantity to them is considered to be a breach of ordinary courtesy. As an example of this feeling, I was told a story of a gentleman who offered a Vaedda a rupee in turn for information supplied by him. It was scornfully declined, but was readily taken when changed into copper cents, one hundred to the rupee.

The wildest Vaeddas now understand the use of money ; one of the men above mentioned suggested to me that I should give him some.

I cannot do better than quote some of Mr. Nevill's remarks respecting their character : ' The true Vaedda varies between a taciturn and almost morose state when hungry, and a laughing reckless mood when not hungry. Their temper changes rapidly, and hence, if offended, in former times they were often guilty of sudden murdef . They would carry on a feud until they considered justice done,1 and then their minds would

1 This must always be necessary among people who have no chiefs or court to which they can carry their grievances for redress. TheTHE'MODERN VAEDDAS 83

cherish no future malicious rancour. The Vaedda is proud in the extreme, and considers himself no man's inferior. Hence he is keenly sensitive to ridicule, contempt, and even patronage,

' He is thoroughly truthful and straightforward; a little kindly sympathy makes him an attached friend, and for his friend, as the Sinhalese nobles over and over again proved, he will readily give his life. The women are chaste and industrious, and have seldom a wish to attract the envy of other women, or the admiration of men.

* They are a merry people, delighting in riddles, songs, and jests. Those I have seen, of all clans, laugh often and merrily. They burst into a verse of song now and again, apparently from sheer exuberance of spirits, and any ludicrous incident amuses them as much as it would a Malay.

' A Vaedda is exceedingly jealous, and this jealousy, coupled with a quick temper and a reckless craving for revenge, probably developed the chastity and monogamy of the race. In any case, its honesty, truthfulness, and obedience to family or clan discipline, stand out in bright pre-eminence.

' As a rule, among the purer Vaeddas the younger women are rigorously excluded, or rather protected, from contact with strangers. They occupy, however, an honourable and free position in the society of their relations/

A ' Mission' established a few years ago to f rescue ' and civilise these people was, like previous attempts, a failure. Nearly all the persons who joined it had Sinhalese names, and probably most of them were not true Vaeddas, though leading nearly the same life as the Village Vaeddas. I learnt that they only remained at it for the sake of the free food which they received. The true Vaedda is not a person who could be induced to settle permanently at such a station. When the hunting season came round it would be impossible to prevent these hunters from feeling an irresistible desire to return to their forest life, which some of them informed me they greatly prefer to any other. A small grant of funds to enable a supply of millet to be given to them in years when unfavourable seasons

same custom is, or was, in vogue among the Jolas of West Africa, among

whom the conditions which affect this practice were similar.84 ANCIENT CEYLON

damage their crops, and, if possible, the provision of some kind of inexpensive wells at their hamlets, such as those made by the ancient Sinhalese and lined with rings of common earthenware, would be of more practical and immediate benefit to them.

The late Mr. Frank Fisher, who was formerly in charge of the Eastern Province as Government Agent, and who understood the natives of Ceylon better than most Europeans, was of opinion that the best method of dealing with the Vaeddas would be to restore one of the larger ancient reservoirs in the middle of their district, and to induce them, by a little pressure if necessary, to settle on the irrigable land below it. As such a scheme would be of benefit to the other inhabitants of the district it might eventually prove successful, but not for some years, and possibly never as a commercial undertaking. In any case it would be a costly experiment. Probably it was through the introduction of irrigation and rice cultivation that the ancient Vaeddas were converted into the Sinhalese of the present day. It was certainly not by means of well-meant but ineffective c Missions/

As one village tank after another was constructed?until every valley, however shallow, had a chain of them, one below the other, each supplying a separate rice-field with water? and the benefits due to these works became appreciated, the Vaeddas who lived near them would be gradually led to adopt rice cultivation as a chief means of gaining a livelihood, while still, like the Wanniyas and many Kandians of jungle villages, devoting a large part of their time to hunting. The example of agricultural settlers from Southern India, and occasional intermarriages with them, would doubtless give a further impetus to this transformation of the race into a nation of cultivators. We can see the very same advance in civilisation taking place among the Vaeddas of the present day. Some who live near the recently constructed irrigation works hat^e already voluntarily adopted rice cultivation, and of their o*ttn accord have planted Coconuts and other fruit trees about their houses, Time Reckoning.?Neither the Forest nor Village VaeddasTHE MODERN VAEDDAS 85

keep any account of time. They have no words for the days of the week, and do not recognise such periods as the hours and their subdivisions, nor even weeks, months, or years.

Counting.?I now come to the question of the Vaedda's ability to count, which has been denied by some. I did not specially investigate the extent of the knowledge of the Wanni-yas in this direction. All those whom I met appeared to resemble the ordinary Sinhalese villagers in this respect, and their common reference to numbers up to a thousand showed that they are well acquainted with them.

Regarding the Vaeddas, I may state that my inquiries were made without interpreters, in Sinhalese or Tamil. I was definitely assured by the Village Vaeddas, and this was confirmed by Sinhalese headmen who speak their dialect, that in the dialect which they call their own they have no words to express either numbers or periods of time. A Village Vaedda who came from the wild tract in the Madura-oya valley in which the Forest Vaeddas are chiefly found, informed me, in Sinhalese, that Vaeddas never make use of any numbers when conversing, and are unable to count. He remarked that he himself could not count; but on making further inquiry I learnt that this only referred to the Vaedi dialect. He could count quite correctly in Sinhalese, and seemed rather proud to do it for me until I stopped him. As apparently all Village Vaeddas are more or less acquainted with Sinhalese, it is safe to assume that they are all able to count in that language.

Regarding the knowledge and use of numbers possessed by the Forest Vaeddas, I have no positive information. If their dialect does not, as I was told, contain words for them, it is just possible that they are unacquainted with them; but before believing this I should require convincing evidence which at present is not forthcoming. That they have a considerable acquaintance with Sinhalese is certain, and if so why should they omit to remember the words for numbers ? The parents of many persons who are now ordinary Village Vaeddas were true Forest Vaeddas sixty years ago,1 yet all the former class understand and speak Sinhalese.

1 See the footnote at the end of tlie chapter.

86 ANCIENT CEYLON

There is a Vaedi measure of length, the Pilluma, which represents the Sinhalese -S'aetaepma or Hatakma, the distance marched by a man carrying a load while on a journey, between two resting-places, called Ruppe in the Vaedi dialect. Its use appears to postulate the employment of some method of stating a distance of several Pillum; and the Village Vaeddas readily mention (using Sinhalese words) the number of Pillum on a well-known path, for instance one which leads to their own village.

It is also quite likely that the Forest Vaeddas, even if they are unacquainted with any words for expressing numbers, may indicate them by means of marks made on the ground, or pieces of stick, or stones, or by their fingers, a common method used by Sinhalese villagers. In the course of conversation the wild Village Vaedda above mentioned indicated a number to me by his fingers of both hands, and a half by crossing his right forefinger over his left one ; and they may do the same.

Some have remarked that the Vaeddas can count only up to five ; and the same reply has been made to me by Tamil-speaking Vafeddas. On inquiry, however, I ascertained that it merely meant that they, who spoke Tamil and could count easily in that language, were only acquainted with the Sinhalese words for numbers up to five; they thought them Vaedi words.

The Village Vaedda above referred to, who was much nearer the state of a Forest Vaedda than the ordinary villager, declared that he and his acquaintances never employed numbers when conversing among themselves. In reply to my special questions he assured me that they would fiever use such expressions as ' three trees' or ' three buffaloes'; he insisted that they would only say the words * trees * or * buffaloes/ without specifying the number. He seemed to think that the actual number would be of little importance ; it would be enough to know that there were more than one. I have no doubt that this is correct, as others confirmed it; but it is far from proving their inability to count when they desire to do it.

Mr. Nevill remarked on this subject; * The earlier observers-. THE MODERN VAEDDAS 87

are right in saying that they do not count. Practically one, two, several, many, very many, make up their use of numbers/ I am not satisfied that this can be accepted as final, even in the case of the Forest Vaeddas, if it was meant to indicate not only their use but also their knowledge of numbers, since it is quite certain that the Village Vaeddas, at any rate, both can and do count without difficulty by employing Sinhalese words or their fingers, although they, too, have been supposed to be unable to do it.

Whatever the final result of the investigation of the knowledge of numbers possessed by the Forest Vaeddas may be, the absence of special Vaedi words for them is of little value as evidence of the state of Vaedda civilisation, either now or in past times. If the vocabulary which I append be examined it will be found that there is in it only a single pronoun, and that is practically a Sinhalese, that is, an Aryan word. If they adopted the Sinhalese pronouns in the place of those which they possessed originally, they could equally employ the Sinhalese words for numbers instead of their own. Their long and intimate connexion with the Sinhalese is evident in their vocabulary.

The only indication of their use of numbers in early times is the statement in the Mahavansa that the wedding festivities at the marriage of one of the local chiefs were to last seven days. Even if this was an invention of the early Sinhalese annalist, it proves that he, who must have had some acquaintance with the ways of the aborigines, believed not only that they were able to count, but that they kept a time record.

I am strongly of opinion that if any Vaeddas do not habitually make use of numbers, it is merely because they do not find it necessary to employ them, and not from any incapacity to understand them.

As an illustration of this, I give a practically literal translation of a few lines from a folk-story told in Sinhalese by a Village Vaedda of the interior, called Yapa, a typical Vaedda name, and written down verbatim in that language.

It is evidently a story originally learnt from the Kandian Sinhalese, and there is nothing in it to indicate any connexion

88 ANCIENT CEYLON

with Vaeddas ; but the mere recollection and repetition of it prove that the Vaedda and his hearers had a good acquaintance with numbers, including simple addition and subtraction, otherwise they could not understand the point in it. It is too long to be given in full here. It relates the adventures of a Gamarala, or minor village headman (who is often the butt of folk-story jokes), who being anxious to eat some cakes, after a quarrel with his wife over them, took some rice and asked the women at another house to prepare them for him.

It then continues as follows: ' Having given them a little rice, he said, " Make and give me five cakes (Kaewun) out of this, please/' The people of the house replied, " Very well/' and taking a little of the rice fried some cakes. The woman who fried them then looked into the account. "For the trouble of pounding the rice and grinding it into flour, I want ten cakes/' she said. " Also for the oil and Coconuts * I want ten cakes; and for going for firewood, and for the trouble of frying the cakes, I want ten cakes." So that on the whole account for cooking the cakes it was made out that the Gamarala must give five cakes.

'Next day the Gamarala, having eaten nothing at home, came to eat the cakes. Having sat down, " Where are the cakes?" he asked. Then the woman who fried the cakes said, { Gamarala, from the whole of the rice I fried twenty-five cakes. For pounding the rice and grinding it into flour I took ten cakes. For the oil and Coconuts I took ten cakes. For going for firewood, and for the trouble of frying the cakes ten more having gone, still the Gamarala must bring and give me five cakes/*

'Then the Gamarala thought, "Ada! What sort of a cake-eating has happened to me ! " '

It is evident that the persons who can enjoy such a story as this have as good an acquaintance with numbers as the ordinary Kandian villager.

There is also among the Sinhalese and Tamils of the villages in the jungle a laconic mode of expression which closely ap-

1 Scraped Coconut was put inside them, and they were fried, in Coconut oil.THE MODERN VAEDDAS 89

proximates to the Vaeddas' talk without using numbers. For instance, if a drove of wild pigs or a herd of deer cross the path, these villagers would rarely draw attention to them by uttering more than the single word " Pigs/' or " Deer " (pi.). If a man fire at some birds and miss them, his comrade does not say, " You missed them; all three have flown away " ; but merely Giya, " gone/' It is nearly the same in some parts of India. In the folk-tales, the European man-eating giant exclaims when he scents the youth hidden in his house, " Fee, fi, fo, fum ! I smell the blood of an Englishman " ; but the laconic Eastern ogre only says, Manmh-gandah, " Man-smell"!

In the same way the Vaedda finds the simple plural a sufficient numerical expression. He is not conversational, but very laconic ; he says no more than is necessary. The intelligence of his hearers easily supplies the blanks in Ms speech, which consists of concise statements of simple facts, or opinions upon matters with which all who are present are conversant. Thus he needs few nouns, and still fewer verbal forms, and practically no abstract expressions, except such adjectives as good, lad, and the like. The ordinary conversation of the Wanniyas aid of the Kandians of the more secluded jungle villages is ntarly similar.

Stesttires.?When a gesture will convey his meaning, the wilder Vaedda uses it in preference to spoken language. If the thumbs be placed side by side, and the forefingers be raisMl and curled forward until their tips approach each other, he tdnks that any one will understand that this must mean a bufalo's horns, and therefore a buffalo.

head is often utilised for this purpose. Instead of to, * there/ or Oba, * on that side/ a slight inclination of the lead sideways fully expresses his meaning. A similar movemtat implies * yes/ or ' very well/ when an affirmative answer | expected, just as the least shake of the head suffices for a negative, as with us. Anything in front is indicated by the chin/yhe head being tilted slightly backward.

A doul% nasal grunt conveys various ideas according to its tone ; ismeans ' yes/ * is it so/ * no * or * do not/ in which

9o ANCIENT CEYLON

senses it is used by the Sinhalese. A near approach to it is the affirmative aspirate Hd, which is also in constant use by the Kandian Sinhalese, but not those of the Low-country; it is noteworthy that I found it (as well as the Sinhalese affirmative ehe) nearly equally employed by fhe inhabitants of the Gambia valley, in West Africa.

When out in the forests, the Wanniyas and Vaeddas call to each other by an exact imitation of the bleating cry of the small Hornbill (Tockus gingalensis). This does not disturb any animals, of course. The former race, and probably the Vaeddas, are also on such occasions accustomed sometimes to utter the grunt of the Buffalo. I think this is done if a slight rustle be made when game is near, so as to allay any suspicion which it might arouse. It may have other meanings, and it is also a call to each other when near at hand. I do not remember hearing them imitate any other animal.

Domestic Animals.?The Vaeddas are said by Mr. Nevill to keep only Dogs as their domestic animals, but I was informed by the Sinhalese headmen of their districts that many of the Village Vaeddas also possess Buffaloes. A few Wanniyas have some fowls, as well as Buffaloes, Black-cattle, aid Dogs.

The dogs are trained for hunting, and will track any wouncfed animals, or follow up unwounded ones, through the thickest jungle; they are specially taught to catch the snail Mouse-deer, or' Miminna ' (Meminna indica), and the ' Igusna/ and Mr, Nevill says also Porcupines and Hares. He f und that from three to five are generally maintained by each Wage Vaedda household.

Well-trained dogs of this kind, of no particular bfeed,sharp-snouted, pointed-eared, little bigger than an Airedale terrier, in colour commonly yellow-brown or black, the ordinary nondescript dogs that are seen in every village, are waderfully intelligent in the forest. I have myself seen a si?all pack, the general set of curs that are found about cooly Ws in the jungle, perform a feat that astonished me.

I was then engaged in the restoration of an anc$Rt tank or reservoir, which had an embankment a mile inuength, andTHE MODERN VAEDDAS 91

covered 170 acres. A party of earthwork labourers were excavating soil in the jungle at the low side of the embankment, at about half-way from each end. On going to work one morning the men startled three Axis deer that were grazing close to their working-place, and the dogs belonging to these coolies at once set off by themselves in pursuit of them. They followed them for some hours, gradually bringing them close round the upper side of the reservoir, as we could hear by an occasional faint yelping which reached our ears across the water; and after a chase of several miles through the thickest thorny jungle, they finally drove the exhausted animals completely round the reservoir, and into the very spot from which they had commenced the hunt; and their masters killed all three there.

The same or similar dogs were greatly interested in a tame ?Leopard which I had at that time, and parties of three or four of them, or on rare occasions single individuals, made periodical visits to my quarters, a mile from their homes, to inspect it. On their arrival they sat on their hams at a very safe distance, and watched the Leopard for some considerable time, finally trotting back after, as a rule, behaving in the manner customary when dogs meet with odorous corners or objects.

A trained dog of this description will lie flat on the ground, with his ears, if they be not cropped close to avoid injury by thorns, laid close to his head; and in this attitude and on his own Initiative draw himself forward by his forelegs until he has passed completely under heaps of thorny bushes that have been piled up for burning, and seemed to have no passage through which such an animal could crawl. These are favourite hiding-places for the Mouse-deer and Forest Hares. One hunter with a gun assured me that with a single trained dog in a leash, to prevent its too rapid progress, he was certain to loll any wounded Sambar deer that he followed up.

Such dogs as these are invaluable assistants to the hunters in the dense forests of Ceylon, and an old Wanniya informed me that four which he kept had run down and captured many Sambar deer for him. As Mr. Nevill remarked, the dogs act as guards of the huts as well as the camp, and when they are92 ANCIENT CEYLON

present their masters know that they will have ample notification of the approach of strangers, whether bipeds or quadrupeds.

We are told by Mr. Nevill that in former times the [Village] Vaeddas kept Buffaloes which were trained for use in hunting ; they are still employed for the purpose by some few Sinhalese and Tamils. The animal obeys orders communicated to it by means of a string which passes through the septum of the nose, * and the archer stalked his game behind it, shooting either over or under it, as occasion required. They are now trained to allow use of firearms.1 x A gentleman who had been out shooting with one informed me that he experienced no difficulty in approaching various kinds of game in this manner, round the sides of open grass plains. The time selected for the purpose is a bright moonlight night, when the animals can be seen at a considerable distance.

Mr. Nevill also learnt that when they had them the Vaeddas used milk taken from the Buffalo cows; and he remarked that ( well informed old Sinhalese have told me that the Una-pana Vaeddas, and allied clans, used to ride Buffaloes, the wife sitting beside her husband. This is mentioned in one widely known song also/

A very few Vaeddas who grow rice must make use of either their own or borrowed Buffaloes in its cultivation, for converting the surface of their rice fields into mud prior to sowing, by trampling it continuously while wet. Some Wanniyas also use them for the same purpose; at a hut in one of their hamlets the mud hole in which the animals wallowed was so close to the door that the occupants could hardly avoid passing through part of it on entering or leaving the house. When I asked one of the occupants if they did not get malarial fever in such a site his reply was characteristic. " Why not ? " he said ; " we do get it." He added that they were considering the advisability of moving their quarters, and abandoning the site to the Buffaloes.

Games.?I made no inquiry regarding the games played by either the Wanniyas or Vaeddas. Mr. F. Lewis has informed 1 The Tapr $anian, Vol. i, p. 191.THE MODERN VAEDDAS , 93

me that the only one of which he heard was played with the small red seeds of the Olinda Creeper (Abrus -precatorius), which one person tossed to another. I feel no doubt that .other amusements are known to them.

Folk-Stories.?The most secluded Village Vaeddas1 of the interior told me that they are acquainted with many folk-stories. The names of several that they mentioned, as well as others that I have collected from people of different villages, show that they are the same as the tales related by the Kandian Sinhalese of the North-western and North-central Provinces. They appear to have been learnt, like the one already given, from Kandian Sinhalese visitors or settlers, or perhaps have been passed down from the earlier Vaeddas of the North-central Province, who must have acquired them from their neighbours in that case. There is not one among them, so far as my information extends, which describes the primitive life, or ideas, or customs of the Forest Vaeddas. This almost makes one doubt if the Forest Vaedda is an altogether primitive being.

I reserve the stories for publication with a collection of other Sinhalese tales ; but I append a translation of one, evidently of early date, about a Vaedda, that was written in Sinhalese for me in the North-central Province. Its conclusion is interesting. It will be observed that notwithstanding his poverty, the Vaedda is represented as being appointed the local king of the district in which he lived. I have adhered to the words of the story as they were written, and have inserted in brackets a few others that are required to explain the meaning in some places.

* A STORY OF A VAEDDA '

Once upon a time in a city a dana [or feast for Buddhist monks] was given at the royal palace. On the next day the surplus rice was deposited for animals to eat, and dogs, cats,

1 Excepting one small Vaedl hamlet, there is not a village within ten or twelve miles of theirs.94 . ANCIENT CEYLON

pigs,1 fowls, and crows came and began to devour it. Then a Vaedi youth, who had gone to kill some game and was hungry, came and saw the fowls and pigs eating some cold cooked rice, whereupon he went to the heap of rice, and pushing aside the upper part of it took a little from the bottom and ate it.

At that time the royal Princess was at the open upper story of the palace. She saw this action of the Vaedda, and remarked to her mother, " Ane ! Amme I However poor a man may be he does not do that disgusting work." The Queen admonished the Princess, and said to her, " Appa ! My daughter, do not say so of any man whatever; you do not know what may happen to you. [It might be your fate to be married to such a person.] " Then the Princess, speaking in ridicule of the Vaedda's want of good looks, replied, " If so, why should I wear this costume ? [I may as well begin to dress like my future husband's people.] ?" The Vaedda, after stopping and overhearing this conversation, went away.

As a lion used to come to that city [and carry off the inhabitants] the King subsequently caused the following proclamation to be made by beat of tom-toms : " I will give my daughter to any person whatever who kills the lion which comes to this city [and devours the people.]" On hearing this, the Vaedi youth dug a hole in the path by which the lion came, and having got hid in it, when the animal approached shot it with his bow and arrow and killed it.

When the King learnt that somebody had killed the lion, he gave public notice that its destroyer should be sought for. The Vaedi youth then caine forward, and after he had [proved that he was the person who killed it] the King gave that royal Princess to him in marriage [and he went away with her].

While she was living with him another good-looking Vaedi youth accompanied him one day. On seeing him, the Princess tricldshly drove away the Vaedda who was her husband, and married that handsome Vaedi youth.

It was not long before this Vaedda one night killed a Buffalo, and [taking some of the flesh] said to the Princess, " Cook this

1 No pigs are now kept by Kandian Sinhalese of the North-central or North-western Provinces.THE MODERN VAEDDAS 95

and give It to rue." The Princess replied, " It would be disgusting work for me to do ; it is no business of mine " ; [and she added] " What does it matter if my first husband is not good looking ? he was good to me/' Saying this, she drove this Vaedda away, and seeking the place where the first Vaedda whom she had married was stopping, went up to him, and said, " Let us go [home together]/' But the Vaedda refused.

After that she put on her Princess's robes as before, and came away.

In a little while afterwards that very Vaedda was appointed to the kingship, and everybody subsequently lived prosperously and in health.1

When they can repeat, as.they have done for me, page after page of these stories, varying in almost no detail from those of the Kandians, it does seem rather absurd that some who have described these people should have remarked that their memories are defective. What better test of their retentive powers could be desired ?

Are the Forest Vaeddas Primitive ??I have ventured to utter a doubt as to the position of the Forest Vaedda of these days. Is he, at least in part, the degenerate descendant of more civilised ancestors, and not altogether primitive ? There are one or two facts which to a certain extent tell in favour of such an hypothesis.

It is made clear by Captain Robert Knox 2 that in the middle of the seventeenth century the majority of those who retained the name of Vaeddas were such as we should now term Forest Vaeddas. They were then found throughout a large tract of country in addition to the present Vaedi-rata or ' Vaedda Country/ on the east of the Kandian mountains.3 He mentioned that about Hurulla, in the North-central Province, ' there are many of them that are pretty tame and

1 A proof, according to Eastern notions, of the excellence of the ruler.

2 An Historical Relation of the Island Ceylon, 1681, p. 61.

3 At the beginning of last century Percival mentions Vaeddas as being found in considerable numbers in the Northern Province. Probably these were Wanniyas. (An Account of the Island of Ceylon, p. 273.)96 ANCIENT CEYLON

come and buy and sell among the people/ and that he saw manyof their camping-grounds in the forests between. Anurad-hapura and Arippu. Even if' the tamer sort' could be found ' it must be with a great search in the woods/ as ' they have no Towns nor Houses/ All lived solely by hunting; ' they never Till any ground for Corn, their Food being only Flesh.' I shall assume, therefore, that a few centuries ago the ancestors of all the present Village Vaeddas were in reality Forest Vaeddas?as we know was actually the case with many of them during the last century?and that at that period they acknowledged the same deities as their descendants.

The evidence, chiefly found in succeeding pages, which tends to indicate either the lapse of the Forest Vaeddas from a more civilised state, or their close connexion in former times with civilised people, is as follows:?

i. They claim to belong to the highest castes of Vaeddas. Some of the wildest of them are members of the Bandara Warige,' the Chiefs Clan/ from which alone the Vaedda chiefs and kings were taken in ancient times. If these chiefs were civilised, many of the other members of the same leading clan were probably equally civilised.

2. Their knowledge of the Sinhalese language, which they spoke even in the time of Knox. Had they always been isolated from civilisation, as at present, it is difficult to comprehend how they could acquire this language. The fact that they understand and use in invocations such classical expressions as Nirindu^ ""Chief of men/ a poetical title meaning a king, proves a more or less intimate acquaintance with the tongue in ancient times. Such a word is never employed in modern colloquial Sinhalese.

3. Their adoption of the worship of the Goddess Mohia!, which must have been acquired through Sinhalese who had taken it over from Tamils, if not directly from Tamils. In either case it postulates an intimate and lengthened acquaintance with civilised people.

1 This word, the Sinhalese form of Nara -h indra, occurs in an invocation of the Vaeddas which Dr. Seligmann was good enough to

show me.THE MODERN VAEDDAS

97

4. Their cult of Panikki [the] Vaedda, a distinguished Vaedi chief who lived in the North-western Province, and was created a Bandara Mudiyanse or Mudaliyar (the title of a superior chieftain), in the latter half of the fifteenth century.

5. Their adoption of a whole series of the demons of the Sinhalese, which were acquired by the latter from the Dravi-dians of Southern India. Nothing but a very close connexion with the Sinhalese or Tamils can account for their taking over these evil deities and learning their attributes.

6. The mixed blood of the Forest Vaeddas, as well as that of the Village Vaeddas. While the majority are brown, some have black skins, which cannot have descended from Sinhalese, among whom a really black colour is quite exceptional; it must be derived from a strain of Dravidian blood. To acquire it they must have been on terms of intimacy with Sola or Pandiyan Tamils,1

In the face of these facts it is difficult to resist the conclusion either that nearly all were once partly, if only slightly, civilised, or that at the least they must have been joined in their forest life by considerable numbers of Sinhalese and a few Tamils, that is, by civilised people. Knox even stated that this was the case. He remarked, * They are reported to be courteous. Some of the Chingulays [Sinhalese] in discontent will leave their houses and friends and go and live among them, where they are civilly entertained' (p. 63). This adoption of the hunting life by occasional civilised villagers most probably continued for many centuries, and the cumulative effect of its influence on the Vaeddas is evident in their language and beliefs.2

I have already drawn attention to the incontrovertible fact that there was a considerable Vaedda population at Anuradhapura in the time of Pandukabhaya; and I may remark that the evidence of the caves is conclusive as to the abandonment of the cave life by nearly all the Vaeddas in pre-

1 Dr. Seligmann has met with some Tamil expressions in the invocations of the Vaeddas.

2 Mr. Bell says of the Vaedda villages in the North-central Province, * Low-Country Sinhalese squatters have settled in every hamlet.* (Archaeological Survey. Annual Report for 1897, p. ro, footnote.)

H

g8 . ANCIENT CEYLON

Christian times. There is good reason to believe that the caves were not re-occupied by them until several centuries had elapsed after the time of Christ. The people who had lived in them must have become villagers. It is possible that heavy taxation,.or misgovernment, or Tamil invasions induced a certain number of these villagers, who had always lived parfly by hunting, to revert to the forest life of their ancestors. Parties of Kandian hunters often occupy some of the caves for a: ^considerable time at the present day.

High Bank 0$ Vaeddas.?The Vaeddas claim to be of high caste, and their leading clans hold that they are not inferior in this respect to* any Sinhalese, whom they consider to be interlopers. One of them remarked to me, " The whole country was ours before the Sinhalese came.0 It is significant that their rank does, not depend on their present state of civilisation ; some of the wildest Forest Vaeddas belong to the highest clan, from which their chiefs were selected.

The ending of the story which has just been given was considered by, Sinhalese, villagers of the North-western Province to be quite appik priate, and they stated that it was in accordance with their traditions. They saw nothing incongruous in the appointment of a Vaedda over people of their own race. There are other examples which confirm the Sinhalese and Vaedi traditions of the high rank held by their chiefs. One of them occurs, in an inscription.

At the side of a flight of steps cut in the rock at Dambulla to facilitate the descent from the celebrated cave-temple, the largest in Ceylon*, to the quarters occupied by the Buddhist monks, near which many other monastic buildings stood in former times, several short inscriptions in colloquial Sinhalese of about the third or fourth century A.D. were left as records of the commencement of the chiselling work for cutting out the steps. They record the names of pious personages who perhaps bore part of the expense of the work. Such records as Amatayft Wahabaha tani patagati, ' The place begun by the Minister Wasabha'; Naka lakhi kahi patagaii, * Naga having made a mark began (the work here) *; HumanayaJm patagati, * Begun by Sumana ' ; Mitaha buja patagati, * Begun by theTHE MODERN VAEDDAS 99

landed proprietor Mitta/?leave no doubt as to their general import.

Another x of these notices runs, Sidha. Raja Pulida Abaya nakare Sidahata kapa gala,' Hail I the stone cut by Siddhattha, King Abhaya, the Pulinda, having caused it to be done/

The appellation Pulinda shows that this king was a Vaedda. When the expression occurs in the Mahavansa it certainly refers to the Vaeddas, and there is nothing to indicate that in the present instance the word has a different meaning.

So far as it is of value, the Sinhalese story also supports this interpretation, which at once sets aside all doubts as to the high caste-rank of the ancient Vaeddas, and the commanding position of the superior Vaedda chiefs even seven or eight centuries after the accession of the first Sinhalese king.

Coming down to much later times, there is conclusive evidence of their power in a manuscript (the Wanni Ka Ja~in Pota, * the Book of the Wanni Boundaries') of the time of King Bhuvaneka Bahu VI of Kotta (1464-1471 A.D.), which contains an account of the appointment of a chieftain called Panikki Vaedda, of Eriyawa, a village near Galgamuwa in the Kurunaegala district, to define the boundaries of the Four Wanni Pattus or divisions of what is now the North-western Province. He was granted the title of ' Bandara Mudiyanse/ an expression which could only be applied to a chief of very high caste. After stating the limits of the district, the account concludes as follows in one manuscript: 2. * Having received the orders from the Lord, the Sinhalese 3 King, Bhuvanaika Bahu, Panikki Vaedda fixed and gave the boundaries/

Panikki Vaedda was evidently one of the most important chiefs in Ceylon at that time. He was not merely the supreme chief of the Four Wanni Pattus (Puttalam Pattuwa, Munis-saram Pattuwa, Demala Pattuwa, and the Wanni Hat Pattu) ; these districts were granted to him and his heirs for

1 A facsimile will be found in Fig, No. 153.

2 There are variations in the wording, but not many in the matter, of different manuscripts.

3 As this expression also shows, there is some reason to believe that the book was written by a Vaedda, reference being made in it to ' our servitude' (ape daskama), which Bhuvaneka Bahu abolished.loo ANCIENT CEYLON

ever. This record is so important that I give the words in full, with a translation.

Sitawaka waeda un Bhuvaneka Balm devi maha raj jura-wannen yedi Eriyawe Panikki Vaeddata me hatara pattuwa kacja-in kota irahanda pawatina tek laebunaya.

* Having fixed the boundaries, these four Pattus were granted to Panikki Vaedda of Eriyawa as long as the sun and moon last, by the Great King His Majesty Bhuvaneka Bahu who dwelt at Sitawaka/

He is elsewhere termed Wanni hatara pattu Eriyawe Panikki-rala, * the Elephant-catcher Chief of Eriyawa over the Four Wanni Pattus'; and the leaders under him, called Panikki-ralas or merely Panikkiyas, are mentioned as me hatara pattuwe Vaeddan, ' these Vaeddas of the four Pattus' or districts.

He was an Elephant catcher (Panikkiya) ; and as stated in Upham's Buddhist Tracts, p. 236, he and another chief named Dippitigama Liyana Vaedda, or in another manuscript, Lekan Polpitiye Liyana Vaedda, a Secretary or Registrar, were ordered by the king to capture a Tusk-elephant and take it direct tof Sitawaka, where they showed the king the manner of tying up a wild elephant, the newly captured animal having been freed for the purpose inside a circle of tame female elephants. The men who tied up the elephant received presents and high-sounding titles ; one of them became Eriyawa Wanni-nayaka Si^happuJMudiyanse, and another was called Raja-paksa Kumara Si^ha Wanniya.

The villages of these men, or the chiefs who assisted in the capture of the animal, are mentioned as Eriyawa, Gala-waewa, Dunupota-gama, Kaekuna-waewa, Wilawa, Wara-gammana, Hulugalla, Hata-gammana, Wenda-kaduwa, Mahagalla, Udu-weriya, and Polpiti-gama; they are nearly all still occupied by Kandian Sinhalese who must be the descendants of these Vaeddas of the fifteenth century. Large tracts of rice fields were cultivated at these villages, the sowing-extents being stated in the manuscript.

These are not the only records of the deeds of Panikki Vaedda. When some princes with armed followers arrived from India at Ponparappu, his * Archer Vaeddas * (Malalu Vaeddan}THE MODERN VAEDDAS 101

at once notified the matter to this chief, and Panikki, who is also termed Panikki Maetiyd, * the Minister Panikki/ proceeded to the spot with a large force of Vaeddas to inquire into the cause of their coming. He translated into Tamil the words of the Vaeddas, for the benefit of the visitors, made them show him the presents which they had "brought for the king, and sent his royal master a full report, stating that they carried swords slung from their right shoulders and shields in their left hands, but that they stated that they came as friends, and were in want of food ; he awaited instructions. Eventually he was ordered to feed them, and to allow them to proceed to Sitawaka for an audience with the king, A large guard of Vaeddas under Panikki accompanied them, apparently to see that they caused no damage on the way. The visitors stopped at Munessaram to pay their devotions at a temple of Vishnu, who granted them permission to proceed to the king.

In the first half of the seventeenth century we find Vaeddas still holding important positions in the country. A short manuscript in my possession which apparently dates from about 1640, contains some particulars of the efforts made by Prince Wijapala to retain the control of the Matale district-As we learn from the Mahavansa (ii;'p. 330), the Prince's father was King Wimala Dharma Suriya I (1592-1620); and his uncle Seneratna (1620-1627) having succeeded this king placed him in charge of the Matale district.

The account commences by stating that * Wijapala Maha Rajayano, of the Godapola Maha Wasala/ or palace, having failed to conquer his enemies?that is, his cousin, Rajasinha, who had followed Seneratna on the throne, and with whom he had quarrelled?called out his adherents in the Matale district, and with their assistance dispossessed several chiefs of their territories. The representatives of * the three Matale Houses ' responded to his summons; they were Kulatunga Mudiyanse of UdupihiUa, Candrasekara Mudiyanse of Dubu-kala, and Waniseka Mudiyanse of Alu Wihara.

The following Vaedda chiefs are also mentioned: The Vaedda chief of Hulangomuwa, Yahamipat Vaedda, Kannila Vaedda of Pallakanan-gomuwa, Herat Vaedda of Nikakotuwa,102 ANCIENT CEYLON

Maha Tampala Vaedda of Palapatwala, Maha Domba Vaedda of Dombawala, Wall! Vaeddi of Wallivela (a female Chief), Maha Kawudella Vaedda of Kawudupalla, Nairari Vaedda of Naran-gomuwa, Herat Bandara Vaedda of Madawala, Imiya Vaedda of Kampalla, Makaraya Vaedda, Koduru Vaedda, Raeka Vaedda (evidently a title, as he was the Guardian of the district boundary), Maha Kanda Vaedda of Kandapalla, Hempiti of Galevela, Baju of Udugoda, Minimunu of Pallesiya Pattuwa, Devakriti of Melpitiya, and Kadukara of Bibile. All these are stated to be Vaeddas ; they were ' of the Vaedi wasagama/

As no other leaders are mentioned, it is certain that these Vaedda chiefs were included among the most important personages next to the three superior Kandian chiefs. The Matale district was evidently full of Vaeddas at that period.

The manuscript also contains a bare reference to the reason of the invasion of Ceylon by the S51ians of Madura in the reign of Wankanasika Tissa (110-113 A.D.). This is termed 'the War of the short-horned Buffalo (ankota miwuwage hatana) of the widowed Vaeddi, Simi of Dodandeniya/ Unfortunately, no expltoation of the phrase is furnished. Doubtless it commemorates some incident that was popularly supposed to have led to the war between Ceylon and Madura, regarding the cause of which the histories contain no information. We may conjecture that some traders from Madura killed or carried off the widow's buffalo, and that the reprisals made by the Vaeddas eventually induced the Solian king to avenge his subjects by invading the country. Whether the dispute originated in this manner or not, the traditional phrase may be taken to prove that the Vaeddas possessed buffaloes in the second century A.D.

Their high caste-rank is still admitted by most Sinhalese who are acquainted with them. I was informed thirty years ago by the brother of one of the Ratemahatmayas, the superior |,Kandian chiefs, that Ms family was intimately allied to the Vaeddas by marriage, and that such a connexion was considered to be by no means a mesalliance.

No one who knows the intense family pride of the KandianTHE MODERN VAEDDAS 103

chiefs could suppose that they woul$ allow their sons to marry wives selected from the Vaedda clans if these were thought to be of much lower social status than themselves; and still more rigorously would they be debarred from marriage with them if they had been, as Professor R. Virchow-said, mere primitive savages of a lower type than the Australians and Andamanese. It would be an insult to them to even suggest that they would ever, in a single instance under any circumstances, consent to such unions.

It is also impossible that a race of savages "would be selected as the special guardians of the important Hindu temple of Skanda, the War God, at Kataragama, in 'South-eastern Ceylon.

Evidence of Former Civilisation.?Professor R. Virchow has written of the Vaedda race that * One may call it among the smallest [in stature] of the living human tribes' ; and after stating that he thought it just conceivable that some remains indicating their former higher culture might yet be discovered, he proceeded to remark, ' But what will be gained even by this ? At best the possibility of placing the Vaeddas on a level with the Andamanese and the Australians, whilst, according to present facts, they must be placed decidedly lower. A people who do not even possess clay vessels, who have no knowledge of domestic animals beyond the dog, who are unacquainted with the simplest forms of gardening and agriculture, who lack almost every kind of social institution, who are not even counted among the outcasts by their * civilized neighbours, cannot possibly ever have had the means which make a higher culture of any kind possible. The hypothesis of a return to barbarism must hence be d?finite#y given up/ *

Had the learned Professor been in possession of the information which I have given in the last few pages, he might perhaps have modified this sweeping condemnation of the race to the lowest place among the lowest savages. But even the early Sinhalese annalists furnished particulars which, if they are to be credited, disprove the Professor's conclusions.

The references to the Vaeddas in the Sinhalese histories and 1 Monograph on the Vaeddas (Translation), p. 108.

104 ANCIENT CEYLON

the Valahassa Jataka story show clearly that in pre-Christian times, when it must necessarily be admitted that they were numerous and well known in the country, part of them at least were believed to have held a far higher position in the scale of civilisation than their direct representatives of the present day. It must not be forgotten that the accounts which we possess were compiled from annals that were almost certainly?as the accuracy of the details in other respects shows:?committed to writing by the second if not the third century B.C. The more our knowledge of the early history of the country progresses the more evident does the general truthfulness of the early accounts become. The careful Sinhalese chroniclers of that time would be most unlikely to attribute to the aborigines more advanced customs than those which they saw for themselves among them, or to place them in a higher social position than they occupied in their day or in the traditions of their forefathers.

In describing the uncivilised natives of a conquered or newly acquired territory, the general tendency among writers down to comparatively recent times, and not among the early authors only, has been in the opposite direction. They have represented people with a certain amount of culture as mere savages, and savages have been even described as no better than the wild beasts, and as using no human form of speech. On this accoult, any evidence of the civilisation of the ancestors of tfieVaeddas which occurs in the early histories may be accepted with much confidence.

What is this evidence ? 1 Assuming it to be trustworthy, let us see what deductions may be legitimately drawn from it. We are told that the country was politically organised, that is, that in the fifth century B.C. it was ruled over by chiefs who lived at settled towns or villages which had a considerable population. Eighty years after the first Sinhalese king began his reign, we find a supreme sovereign of the Vaeddas, whose name is given as Citta, residing at Anuradhapura almost on an equality with the Sinhalese king, and sitting on a similar throne to his when the royal party were present at public 1 It will be found in detail in the preceding pages.THE MODERN VAEDDAS 105

festivals and sports. It is specially added, in order to mark the position held by the Vaedda chief, that both the thrones were of the same height. According to eastern custom, and even western also, this proves that the Vaedda ruler took precedence of all persons in the country except the Sinhalese king himself, who thus publicly acknowledged their equality of rank. Had the annalist been a Vaedda, we might suspect that he had invented such a description of his sovereign's status at the court; we may feel sure that no Sinhalese chronicler would have deliberately perpetuated a story which placed the ruler of the aborigines in such a prominent position unless he and his compatriots believed that the Vaedda chief had actually occupied it.

In addition to the sovereign of the Vaeddas, another Vaedda chieftain, Kalavela, who held a post of almost equal importance in the country, is mentioned as residing at the Sinhalese capital. It is explicitly stated that it was with the assistance of these two chiefs that the Sinhalese king ruled over the country. It may be said, therefore, that this account completely supports the more doubtful one which is given of the social position of the local chiefs in the time of Wijaya. They were persons with whom the Sinhalese rulers could associate on terms of practical equality. I suggest also that it is difficult to account for the devotion of the Vaeddas to Pan Jukabhaya, before he became king, unless he was connected with their race through his grandmother.

The reference to the wedding festivities of even the rulers of the Vaeddas indicates that they were elaborate festivals which lasted some days, and that the etiquette of the country rendered it necessary for the princess who was to be married to be escorted by her mother to the town or at which the ruler dwelt to whom she was to be united.

The Vaeddas are described as being well dressed, The had a special ceremonial costume which even a prince the court of one of the sovereigns of the Ganges valley was not .ashamed to wear when he assumed the sovereignty over them. The costumes or ornaments of the royal .also.found suitable for the followers of the Indian Itio6 ANCIENT CEYLON

is clear that the dresses of such people were no mere waist-cloths of Riti bark, or girdles of leafy twigs. They must have consisted of imported cotton cloth of an ornamental pattern, brought into the country either by Magadhese or South-Indian traders.

These statements are supported by modern Sinhalese traditions, and the accounts of the Vaeddas which were collected by Mr. NevilL These name even the clan, the Bandara warige?the 'Chiefs Clan/ which still exists, and to which some of the wildest Forest Vaeddas belong1?from which the kings and chiefs were chosen in former times ; and they mention the coloured dresses and jewels, and the golden household utensils which their more settled representatives still possessed in the last century. Among the names of modern Village Vaeddas given below it will be seen that one is called Randunu Wanniya, .' the Wanniya of the Golden Bows/ 2

If the Vaeddas were in the state of civilisation which these facts indicate, it would be unjustifiable to suppose that they could be ignorant of all knowledge of numbers.

The Sinhalese annalists and the writer of the Valahassa Jataka agree that trading vessels were often wrecked on the shores of Ceylon before the advent of Wijaya, that is, in or before the fifth century B.C. The tradition of the Vaeddas is also quite definite as to the arrival of their supreme deity in a ship from Southern India, ' in the olden time/ which we know by the reference to him in the reign of Pandukabhaya must have been prior to the fourth century B.C.

These were not local ships ; it is practically certain that they were vessels which came from ports on the Indian coasts. In the Saixkha Jataka (No. 442) there is a reference to a ship built of planks, with three masts ; and voyages were certainly made at an early date from the Ganges valley to Suvanna-bhinm, * the Land of Gold/ that is, Burma. In the Indian Antiquary for 1876, vol. v, p. 340, Dr. J. Muir published translations of some maxims from the Maha-Bharata, one of which

1 I have stated that I met some wto belonged to it.

2 This is strong evidence that the Wanniyas axe really Vaeddas ; another Vaedda is also called * Wanniya.*THE MODERN VAEDDAS 107

runs, ' On seas, in forests wild, the bold will risk their precious lives for gold ' ; and even in Vedic times sea-voyages, some of which occupied several days, are often mentioned. It must have been such vessels as these which brought the first Gangetic travellers, and at a much later date Wijaya and his relatives, and their followers.

With what object did the first Magadhese traders venture upon the dangerous voyage to Ceylon from their distant country on the Ganges, a journey of more than 1,600 miles ? This long voyage cannot have been undertaken for any other purpose than to obtain the articles produced in the country, ivory, wax, incense, and probably also pearls and gems,1 being part of them. We know also that these were not paid for with money, which would have been useless to the natives; the traders must have brought with them cargoes of other goods, like those taken to Burma according to the Jataka stories (in which whole shiploads of merchandise are mentioned)?to be disposed of in exchange for the local commodities. We shall probably be correct in assuming that these cargoes consisted largely of cotton materials, beads and other ornaments, axes and arrow-heads of steel, and cooking and other vessels of earthenware, copper, or brass, all of which would be readily taken by the natives in exchange for the produce of the country.

This at once presupposes an internal trade in these articles, like that of prehistoric people in Europe. All would not be retained in the hands of the dwellers on the coast; a part of them would be distributed throughout the whole country by some form of barter,2 or possibly by local traders established at settlements far inland, in the ' forests wild * of the Maha-Bharata, where the produce of the district would be collected in exchange for them, exactly as at present.

1 The Mahavansa states that Wijaya sent to his father-in-law the King of Madura, gems, pearls, and chanks (i, pp. 34, 35).

2 So also it is stated of the natives of Central Australia, ' The trading propensities of the Australian natives have led long ago to the disposal far and wide over the continent of the iron tomahawk of the "white man. . . . One group barters what it makes for the products of another living, it may be, a hundred miles away.' (Dr. Howitt, The Native Tribes of Central Australia, p. 575.) There was a similar prehistoric trade among the American Indians, and in Europe in Neolithic times.io8 ANCIENT CEYLON

The mere fact that Indian traders came so frequently to Ceylon that vessels were known to be often wrecked on the coasts, proves, without any other evidence, that many natives were in a far more advanced state that the present wild dwellers in the eastern forests, with whom no regular trade could be possible. Permanent trading centres must have been established at fixed and well-known points on the coast, near native settlements, at which the vessels called, and to which the articles produced in the country would be conveyed for barter with the adventurous merchants who came for them with the monsoon winds. All this must necessarily follow if such traders came to Ceylon; and that they did visit the island is confirmed by the presence of the natives of the Ganges valley as settlers in the fifth century B.C. It is impossible that these settlers were the first persons to visit Ceylon from Magadha. The Magadhese were a nation of traders (as the Ordinances of Manu tell us), and probably knew all the coasts of the Bay of Bengal, If they were acquainted with the voyage to Burma they would experience no greater difficulty in finding their way down the Indian coast to Ceylon* In the Sussondi Jataka {No. 360) merchants are stated to have proceeded by water from Benares to Nagadipa, that is, northern Ceylon, in former times.

If some of the inhabitants were carrying on a trade with Indian merchants, and their rulers were considered by the Sinhalese sovereigns to be sufficiently civilised to associate with them, we may still surmise that a great number of the natives continued to gain a living wholly or partly as hunters, leading while in the forests the same wild life as their descendants of the present day.

The annalists evidently believed that no rice was grown in the country before Wijaya's time, since they specially explain that the rice which was cooked for travellers when they landed was procured from stores brought by ships that were wrecked on the coast. According to the custom or law regarding such matters in other countries, of which many examples might be quoted, these wrecked vessels would be looked upon as lawful prizes, either sent by their Gods for their special benefit, or rejected by the God Who ruled over the waters.THE MODERN VAEDDAS

109

I assume that as the newcomers from the Ganges valley, introducing various arts of their own country, settled down permanently, and exhibited a more stable form of civilisation than that of the aborigines, they must necessarily have intermarried with the more advanced natives. While they were being gradually absorbed by them?which would not occupy a long period after immigration, which was probably never extensive, from the Ganges valley ceased-?they imparted to them their own culture, and to a great extent their language.

But the physique, and colour, and hunting proclivities remained unchanged. Many of the villagers of the North-central and North-western Provinces merely require to be sent to live in the forests in order to become once more practically the same Forest Vaeddas who lived by hunting before the time of Wijaya. If these people were isolated in the forests for a very short period, I am sure that in most respects they would be indistinguishable from the Vaeddas, just as the Wanniyas resemble them. It would be an unavoidable result of the environment. They could make neither pottery, nor iron or stone implements *; and dogs would be the only domestic animals that they could retain in the forests. All Sinhalese and Wanniya hunters lead the life of the Forest Vaeddas after they leave their villages on their hunting expeditions, carrying only a small bag of millet-flour, gourds for water, an axe, a knife, and usually, but not always, either a gun or bows and arrows. They all anticipate such a life with pleasure ; they are still Vaeddas at heart. They dress almost like the Vaeddas, and get the same food in the very same manner.

This shows that the appliances of the Vaeddas are such as are best suited to their forest life, and that the absence of others is not a proof that they are the lowest savages. It only proves that they have practically all the implements that are necessary in these dense forests. I cannot imagine that any others but the knife would be of the least use to such hunters.

The omission to keep any record of time, whether days of the week, or months, or years, cannot be considered to be 1 If the potters and smiths were excluded.no ANCIENT CEYLON

conclusive evidence of a primitive state. I found it equally absent among the-Adjammateyi or Jolas (the Diolas of French authors) of West Africa, who are admitted by other natives to be the best agriculturists in the Gambia valley. They stated to me that they only recognised the season for preparing the ground for crops by observing the flowering of certain forest trees. Yet they have fully inflected verbs, with eight tenses and eight persons, and no less than eight regularly formed conjugations derived from each verbal root. So, also, to the early Greeks Hesiod said :?

' When Atlas-born, the Pleiad stars arise Before the sun above the dawning skies 'Tis time to reap ; and when at sunrise now They sink beneath the West, 'tis time to plough.1

The small cranial capacity of the Vaeddas is not a proof of their low intellectual status. Dr. Virchow has shown that the size of the brain in four Tamil skulls is practically identical with theirs, and he states that other South-Indian skulls are similar. No one, I presume, will venture to maintain that the Tamils, or rather the Dravidians, are not a highly intellectual race, to whom India possibly owes a part of Its present culture. The Rt. Rev. Dr. Caldwell, the greatest authority on the subject, said in the Preface to his Grammar of the Dravidian Languages, * It is impossible for any European who has acquired a competent knowledge of any of the Dravidian languages?say Tamil?to regard otherwise than with respect the intellectual capacity of a people amongst whom so wonderful an organ of thought has been developed' (2nd ed.,p. ix).

M. de Quatrefages also remarked that ' the development of the intellectual faculties of man is to a great extent independent of the capacity of the cranium, and the volume of the brain/ l

As to the opinion which is sometimes expressed regarding the intellectual effect of variously proportioned brains, there is nothing to show that the Vaedda cranium is inferior in mental power to that of other dolichocephalic people. As a matter of fact, it is open to doubt if the mere proportions of the cranium are more tha|i insignificant factors in the case. Bra-1 The Human Species, p. 384.THE MODERN VAEDDAS in

chycephalic races are not necessarily of greater mental power than dolichocephalic races. Thus the Lapps are at the limit of brachycephaly, with a cranial index of 85 ; and Mongols, Turks, Javanese, North Americans and even Andaman Islanders have a higher index than Parisians.1

In dealing with the position of the Vaeddas, we are faced with this difficulty?that a portion of the race was relatively civilised in ancient times, while certain members of it are foupd at the present day almost in the state occupied by some of the most primitive peoples. We must adopt a theory which will include all the facts of the case ; and not one which ignores some of the most important and significant and incontrovertible historical details and traditions. We cannot select the smallest and wildest group of Vaeddas, and because of their simple life as hunters place the whole race in the position which they continue to occupy, not because, like the aborigines of Australia and the Andamans, they are intellectually incapable of rising above it, which the example of the others has completely disproved, but partly by accident and partly of their own free choice.

My conclusion therefore is that whether there has been any retrogression of the present Forest Vaeddas from a certain low state of civilisation or not, in very early times a great part of the race had reached a much more advanced state of culture than the wilder members of it, whose more or less isolated life either as hunters, or as htmters-and-viUagers, did not in many cases induce them to feel any desire to participate in it. This more civilised portion has absorbed the Gangetic settlers, and acquired their status and language, and with some intermixture of Dravidian blood, or in many instances without it, has become the existing Kandian Sinhalese race.

The ancestors of the present few hunting Vaeddas?who now most probably number much less than one hundred?either abandoned, some centuries after Christ, a form of village life in which they were partly or chiefly hunters, and reverted to the forest life of their forefathers; or, like some of the wild hunting tribes of the South Indian Mils, remained, at least until very

1 Topinard, quoting Broca and Hamy, in Anthropology, pp. 241. 242.IK ANCIENT CEYLON

recent years, in nearly the original condition of the first comers to Ceylon, apparently simply because they preferred the free untrammelled life in the woods, and found their accustomed habits and household articles suited to all the requirements of a hunter's existence in the forests of Ceylon. The evidence afforded by the caves appears to me to be in favour of the former theory, which is also supported by the loss of their original language and their adoption of the Sinhalese tongue. The majority, however, of those who did not coalesce ^yith the Gangetic settlers and their descendants, or accept their mode of life and culture, have, in comparatively modern times,1 and in certain instances partly through compulsion?since portions of the forests in which they were accustomed to hunt have been cut down in order to permit rice and millet cultivation?to some extent adopted the more civilised existence of their neighbours.. Many keep buffaloes, and all but those few who live only by hunting and fishing, grow millet and other plants suited to their jungle clearings. An exceptional few in favourable sites for it even cultivate rice, and, as some of them informed me, in recent years have settled down permanently and have planted such fruit trees as Coconuts, Areka-nuts, and Plantains about their houses.

No arguments of the supporters of the hypothesis that the Vaeddas are, * at the best, on a level with the Andamanese and Australians/ (which must imply an incapacity for intellectual development), can lay aside the examples which have been given of their high status in former times. Historical facts such as these must necessarily supersede any theories that are not in accordance with them ; if the theories do not agree with the facts, so much the worse for the theories.

1 As a number

an example, I may note that according to Sir Emerson Tennent __ jer of Forest (or, as he terms them, Rock) Vaeddas settled down In hamlets between 1840 and 1850, at one of which there were twenty-five families. He adds, 'it may thus be said that the distinction of the Rock Vaeddas has ceased to exist in that part of the country ; all having more or less adopted the customs and habits of villagers/ ^Ceylon, 2nd ed., Vol. ii, p. 447.)SOCIAL DIVISIONS AND CUSTOMS

Vaeddas are socially divided into a series of tribes JL or clans, called by them Warige (or Waruge, according to Mr. Nevill), of which three hold much higher rank than the rest, with whom their members do not intermarry. These are (i) the Bandara warige, * the Chief s clan ' ; (2) the Morani warige; and* (3) the Unapana warige.

The members of at least these three clans, and I believe those of the other clans also, are admitted by the Kandian Sinhalese to belong to the Goyiwansa or Cultivating caste, the highest among the Sinhalese, though there are several different grades in it. Mr. Nevill was informed that in ancient times the Vaedda kings and chiefs were selected only from the Ban- Jara warige, as the name indicates. He stated that this clan is supposed by some to derive its origin from the children of the Vaedda princess Kuweni, whom Wijaya married, their names being thought to be Sabara and Sabari. Of course no dependence can be placed on any claim to such a descent, though the fact remains that the clan is acknowledged by all Vaeddas, as well as the Sinhalese who are acquainted with it, to be of higher rank than the others.

How it came about that part of the Forest Vaeddas are members of this clan is a matter deserving special investigation, It may possibly be an indication of their relapse from a more cultured state for the reasons suggested by me, or owing to some cause which cannot now be traced.

Below these come the following clans: (4) Urana warige, which Mr. Nevill called Uruwa, and put, In the sixth place; (5) Nabudena or Namada warige ; (6) Urawadiya warige; (7) Aembalana, or Aembala warige ; and Mr. Nevill added also (8) Kovila waname ; (9) Tail warige ; and three terri-

113ii4 ANCIENT CEYLON

torial groups, those of (10) Tambalagama, (n) Kattakulam, and (12) Anuradhapura (? Tamankachiwa) ; as well as (13) the Coast Vaeddas. The warige names of the last four have been lost. Possibly the Wanniyas should also be included as an additional clan.

He found that the Kovil waname has four territorial sections, those of Dambana, Miyangoda, Makanda, and Galkaeta ; but their representatives are now very few in number, and apparently they could give no account of their ancestry, beyond a tradition that it was some members of the Dambana section who discovered the Goddess Valliyamma as a child in the forest near Kataragama, and adopted and reared her until the War-God Skanda married her. He learnt that it was formerly the duty of this clan to act as guards of the Kataragama temple in south-eastern Ceylon, and that they resided in the district adjoining it.

This temple, dedicated to Skanda, is considered to be one of special sanctity, and is visited by pilgrims from all parts of India, including even the North-west Provinces. How it came to be established in such a site, and to acquire such importance is, I believe, unknown ; it must have been partly due to encouragement and support given by the kings of Southern Ceylon in the times when they resided at Tissa or Magama, which is not far distant.

Possibly Kataragama may have been an important site of the worship of one of the deities of the aborigines. Dr. C. G. Seligmann has informed me that the Forest Vaeddas highly reverence a deity said by them to be the spirit of a Vaedda known during life as Kande Wanniya, by which title he is frequently addressed in their invocations. If he was an ancient deity the new settlers may have identified him with Skanda, who is also a Mil-god, and to whom worship is paid on the hills by some of the wild tribes of Southern India, according to information derived from a respectable Tamil eyewitness of it. Skanda's t$$aJl liape in Ceylon, Kanda Kumara, may have assisted in this justification, which would account for the Vaeddas* becoming ilie guardians of his temple, with which, however, Kande Wanniya is not now" connected.FIG. 35. Skanda and Valliyamma (Taiijore Temple).

To /ace p. 115,SOCIAL DIVISIONS AND CUSTOMS 115

The princes (who may have been Vaedda chiefs) of Kajara-gama, as the place was then called, were included among the distinguished persons who were present when the celebrated Bo-tree was planted at Anuradhapura in 244 B.C. ; and that an important Buddhist monastery was established at the spot at that time is proved by its being selected by King Devanam-piya Tissa as one of the first places at which a shoot of the Bo-tree was planted. The only inscription that Dr. E. Miiller saw there was a defaced one of the fourth century A.D.1

Mr. Nevill referred to a local legend that it was at Katara-gaiha that Skanda and his forces defeated the Asuras; and that he also met Valliyamma and married her there, after she had been adopted in a Vaedda family 2 ; but I never heard of her being treated as a special goddess by either Sinhalese or Vaeddas. I give an illustration of these two deities (Fig, 35), It represents a panel at the great Saivite temple at Tanjore, and may date from the thirteenth century A.D.

When Dr. Davy visited Kataragama in 1819 he found two enclosures there, and said of them, * In the largest square are the Kataragama Dewala [temple], and the Dewala of his brother Gana [Ganesa]; a wihara dedicated to Buddha in a state of great neglect, and a fine Bo-gaha [Bo-tree]; and six very small kovils [temples], mere empty cells, which are dedicated to the Goddess Pattini3 and to five demons. In the smaller square are contained a little karan Juwa sacred to Isvara [Siva], the Kalyana Madama [shed], a kovil dedicated to the demon [God] Bhairava, a rest-house for pilgrims, and some offices/4

The tenth, eleventh, and twelfth clans are said by Mr. Nevill to be practically extinct, their members having died out or been absorbed by the surrounding people. The others* who with the exception of the Urana warige, the Uraw5 Jiya warige, and the Coast Vaeddas, are very few in number, appear to

1 Ancient Inscriptions in Ceylon, p. 46.

2 The Taprobanian, Vol. I, p. 180,

3 An account of her will be found in a later chapter, on the Ancient Games.

4 An Account of the Interior of Ceylon, 1821, p. 420, ! corrected the spelling of the names and native words.

n6

ANCIENT CEYLON

have lost all tradition of their origin; but it is possible that patient inquiry by a competent person might elicit some interesting facts, or legends at least, regarding their ancestors.

Pending the publication of Dr. Seligmann's researches, there is not much definite information concerning their social customs. The Village Vaeddas have a simple but formal ceremony similar to that of the Sinhalese, on the day when the child first tastes the food of adults. When the infant is about seven months old the parents fry some Indian corn, or cook some millet porridge, and give a little of this to it, together with a taste of any other food that they possess. It is probably at this ceremony that a name is bestowed on it.

As regards their marriages, Mr. Nevill has condemned the inaccuracy of those who stated that the Vaeddas are accustomed to marry their sisters. This was an Indian custom of early times, however, and the Mahavansa mentions two traditional cases at least?the parents of Wijaya, and the children of Kuweiu?in both of which the contracting parties were the children of the same parents. According to the Dasaratha Jataka (No. 461) Rama married his sister, Sita. It seems also probable that Maha-Naga, the brother of Devanam-piya Tissa, married his sister Anula; while in the Kunala Jataka (No. 536) we find the people of Koliya reviling the Sakyas of KapOavastu (from whom Maha-Naga was descended) as persons who ' cohabited with their own sisters/ Regarding the practice of the Vaeddas I have no direct information ; I should accept Mr. NevilTs statement on the subject in preference to those of others who had not the same opportunities of obtaining correct accounts of their ways. Such marriages with sisters are unknown among the Sinhalese of the present day.

Mr. Nevill learnt that among the first eight clans marriage into another territorial section was usual. His remarks are so interesting that I quote them rather fully : * Thus if A and B are adjacent territories, the bachelor in A goes to B, and marries a wife, if possible his mother's brother's daughter, or the nearest collateral relative similarly connected [that is, through the mother]. The difficulty of finding out who this might be was remedied by adopting the territorial title as aSOCIAL DIVISIONS AND CUSTOMS 117

quasi-gotram division ; and our bachelor in A, whose mother came from A and bore that name, would go to B, and marry the first eligible maiden of the B name, with perfect confidence that she belonged to the correct division of the family.

'This caused an apparent division into family clans [subdivisions] which, however, are better termed territorial clans [sections], to avoid confusion with the true hereditary clan. The necessity of taking a wife from a territorial clan [section] C, D, or E, if no available maiden could be found in B, often led to breaks in the propinquity of the descent, and even intermarriage with a new waruge, as distinguished from a territorial section of the same waruge. The waruge divisions [clans], once political or ethnic in origin (at times of mixed ethnic origin by intermarriage out of the Vaedda race), have now become quasi-g5tra or family clans, and are thus to be distinguished from what I term the territorial clans [sections, which are] all branches of the one true gotra or family clan.

'The rule for marriage was stringent. The daughter represents her mother's family, the son also represents his mother's family [? father's family, as among the Kandian Sinhalese]. In no case did a person marry one of the same family, even though the relationship was lost in remote antiquity. Such a marriage was incest. The penalty for incest was death. Thus the daughter must marry either her father's sister's son or her mother's brother's son, neither of whom would be of the same clan name. [This is the common Kandian custom also.] Failing these she may many any of their name, and should no such bridegroom be available, marriage into a third family becomes necessary.

' The Vaeddas marry young, and are strict monogamists. Consequently the husband and wife are watchfully jealous, each of the other, and love intrigues are few and far between. Nothing short of murder would content the injured party. This strict morality extends to unmarried girls, who are protected by their natural guardians with the keenest sense of honour. It does not extend, however, to widows, however young and pretty, and a widow who avoids exciting the jealousies of the wives may have love affairs with half the men around,(;_ n

118 ANCIENT CEYLON

without exciting any wish for revenge among her relations, who would have given their lives at once to avenge any impropriety of conduct while she was single. The women also are said not to show any excessive jealousy of a widow, if her allurements be not too openly talked of.

' As might be expected, when a wild race marries young and the husband and wife remain constant, any unusual festival is often the occasion for riotous sensuality between husband and wife, who then discard all decency in their private intercourse, and break out into licentious love-songs and gestures. * There are no special marriage ceremonies/ 1 I was informed by the Village Vaeddas that when a young man thinks of marrying, he selects a suitable girl himself, and speaks on the subject to either her father or mother. Having obtained the necessary consent, he takes up his residence at their house without further ceremony, and the girl becomes his virtual wife. After three or four months have elapsed, and he has cut, and sown, and reaped a temporary clearing in which millet is grown, or has otherwise assisted in providing a supply, of food for the family, he is considered to be formally united to the girl. Prior to this, I presume that the union is looked upon as a probationary one, according to the similar practice which is still occasionally followed in the more backward Kandian villages of the interior. It is also a common custom of Kandian villagers not to register their marriage until after the birth of the first child; this leaves the parties free to separate, if they desire to do so, without the trouble of applying for a formal divorce. In such cases the marriage prior to registration is practically a probationary one, like that of the Vaeddas.

The formal consent of a parent, or of the natural guardian if the parents be dead, is the only absolutely necessary part of the Kandian marriage ceremony, which is thus in agreement with the practice of the Vaeddas, and is doubtless derived from them. This consent having been obtained, the living together of the young couple, with or without any other ceremony, constitutes a valid marriage, by ancient Kandian custom. I have known several cases of this kind, in which the permanent union 1 The Tapwbanian, Vol. i, p. 177.SOCIAL DIVISIONS AND CUSTOMS 119

was unaccompanied by any ceremony. A recent law of the last decade renders registration compulsory in order to secure the legality of all marriages.

To show that this practice of the Vaeddas is not a mere primitive trait, it is only requisite to refer to the custom in China, where we are told that ' the only essential feature of a Chinese wedding is the delivery of the bride at her husband's home/ 1 Among the West African Mandinkd and J51as, too, who are certainly not primitive, the consent of the parents renders any marriage valid, and among the latter people there is no formal ceremony.

I am obliged again to borrow the following information regarding funerals from Mr. NevilTs account of them :2 ' Bodies were never buried until the English Government endeavoured to enforce burial. The Vaeddas have not the least objection to the corpse being buried, but object greatly to being forced to dig the grave, a waste of labour, over mere perishable matter, from which the spirit has gone free, they say.

' The Vaedda religion seems to have been such that the spirit alone was recognised as human, and the flesh, when the spirit has left it, receives neither veneration nor superstitious reverence. Where the life left the body, there the body was left, if safe from wild beasts, or if the family were in a hurry. It was usual to put it in a crevice between rocks, or to cover it with boughs ; if no rocks were near, boughs were laid over it. This was merely done in a sense of decency, to prevent wild beasts from feeding upon it. Spirits were not thought to haunt the spot, as among Sinhalese and Tamils, nor did superstition require any funeral rites. Two to five days after the death, however, the relatives were invited to the scene of funeral, and a feast was held. The original object of this seems not to have been religious, but civil. It was in fact a coroner's inquest, and was held to satisfy relations that there had been no foul play. I have Htherto had great difficulty in getting real Vaeddas to discuss the funeral, as they seem to think that I am secretly

1 Rev. I r. Smith. Village Life in China, p. 269. The italics are his,

2 The Taprobanian, Vol. i, p. 179.J-

120

ANCIENT CEYLON

laughing at their want of etiquette on such occasions, and there is nothing a Vaedda dislikes and dreads so much as being

despised as a savage.

' The Vaeddas of Bintenna, however, having assembled relations and neighbours, procure rice or other grain, and decorate the pot in which it is cooked with sprays of the Liniya tree (Helicteres isora), a shrub with leaves like our hazel, but with bright scarlet flowers. If no flowers can be got, bits of red cotton or other cloth should be used. The celebrant then dances round the pot of food with an arrow in his hand, singing any chant he knows, and making obeisance to the food by a wave of the arrow. The food is then distributed, and it is etiquette not to revisit the spot until the flesh has decayed away. There does not seem to be a dread of pollution ; but rather that feeling which makes us think it bad taste to be seen in a nightdress, etc., by our friends makes the Vaedda think it bad taste to go and stare at the decayed and abandoned body of his friend

and neighbour.

' It is evident that this custom cannot apply to those who formerly did not eat grain. These, however, were few. Roasted game would probably with such take the place of grain, and the latter seems only used as the best and most unusual food procurable, much as our poor try to provide cake, and not bread and cheese, etc., at weddings/

My own information regarding this ceremony is scanty. I was told by them that a few days after the burial they prepare food, in the same manner as the Sinhalese make ready a ' dana * or feast for the Buddhist monks who attend at their houses on such occasions, and proceed with it to the grave, upon which they place it. They then call the deceased loudly by his name, to come and eat it. After waiting a little time, during which the spirit is supposed to partake of the essence of the food, all the persons present at the grave themselves eat up the whole of the food. After this feast they return to their houses.

The summoning of the dead to share in the repast makes it clear that this ceremony is a farewell feast with the spirit of the deceased person, who, as the honoured guest, is first fed before the rest of the party take their shares of the food. ?SOCIAL DIVISIONS AND CUSTOMS 121

The Telugu-speaking Gypsies of Ceylon have a similar custom, but in their ease the cooking of the food is also done on the grave itself immediately after the burial. Those present at the grave then eat it at the spot. The * dana' of the Sinhalese appears to be in its origin the same, or a closely allied, ceremony.

Among the Tamil-speaking Vaeddas the ceremony becomes a religious one, approaching in character that of their Tamil neighbours in honour of the Manes, rather than a farewell feast with the dead, and I therefore include it with other religious ceremonies.

I have no information regarding the social customs of the Wanniyas.

I append a list of some names of Vaeddas, in addition to the few chiefs previously mentioned:?

Vaeddas of the Interior.

MALES.

Patabaenda Dematanan Siripala Rani Konaruwa Kira

Milalanan Raddtiwa Sella

Knppeya Bonda Wanniya Rodda Vira Tissahami

Kanda Ktinda Yapa Randunu Punca Mutuwa

Kumma Kayira Vaedibekki-rala "Wanniya Nila Maenika Ranga Punya

FEMALES.

Wall! Rongi ?HudI

Madalena Ukku Wayiri

Rang! Kandi Maeniki

Atti KM Garo

Siribaddanaya

Herald Haedaya

Milalana Panamaya Badapissa Kiritanda Vannaku Banda Hudu Maeniki

Werattl RamI Rangana Latta

Names given by Mr. Nevitt, in * The Taprobanmm*

MALES.

Janaya Kattandi Ahngada Kalingnrala Waeliy a Dahagonaya Siyatii Binakaya Raka

FEMALES.1

Handi Maedini

Kawenibami Rukkuli Ku,mi

1 Probably the final vowel is long in all these names ; but it Is not so marked by Mr, Nevill.122

ANCIENT CEYLON

Southern Vaeddas, chiefly Sinhalese names.

MALES. FEMALES.

Nanharni Kanda Kanatti Puncihaml

Maeddam-Appu Manawalaya Sudu-Appu Isanhami Sinno Hudu Bandara Raterala Ganaeti Umayattl Kalu Ranhami Maenik! Silinduhami - Garu Salangu Kiribandu

Undiyarala Potana 1 Wasani Kir5 j

Tamil-speaking Vaeddas.

MALES. FEMALES.

Kanakkan Palan Kali Naki

Siddena Periya Tutan Valli Karati

Pattar Sinna Tutan Tevi Vayiri

Sinna Kayilan Kandi Sinni

Karaval Yapan Sempala Patti

Kanavadi Velan

Sinnaya Karatan

Kandan

1 So-called because he was born at a village of that name. There are several place-names in the list of chiefs previously given.THE VAEDI AND KAELE-BASA VOCABULARIES

THE Vaedi dialect is to a great extent the colloquial Sinhalese tongue, but is slightly changed in form and accent. Yet closely as it resembles the latter, these differences and the manner in which it is pronounced render it quite an unknown language when it is spoken to one who has not a special acquaintance with it. Besides this, the Vaeddas use their own terms for the wild animals and some other things about which they often find it necessary to converse. Such words are usually a form of Sinhalese, or admit of Sinhalese or Tamil derivations; but a very few may possibly belong to, or be modifications of words in, their original language, forming, with perhaps a few forms of grammatical expression, the only remains of it that have been preserved, with the exception of some doubtful terms found in Sinhalese.

Strange to say, the Kandian Sinhalese and the Wanniyas apparently imitate the Vaeddas while they are hunting in the forests, and also when engaged on ceremonies at their threshing-floors, and use another series of expressions or nicknames for many of the same animals, to the exclusion of the usual names for them. They have acquired a belief that unless a special dialect be employed while they are in the forest, they cannot expect to meet with any success or good luck in seeking honey, or hunting, or in avoiding dangerous animals.

This dialect of the forest is termed Kaele-basa, * Jungle language/ It consists of the employment of new words not only for animals but also for a few other nouns, and for verbs used to denote acts most commonly performed on such trips. In addition, all negative (that is, unlucky) modes of expression are totally debarred from use on such occasions, as well as the words

123*»..*

124 ANCIENT CEYLON

meaning ' insufficient' and * too much/ which are inauspicious as indicating dissatisfaction with the number or quantity to which they are applied.

As it appears to have some bearing on the connexion between the Kandian Sinhalese and the Vaeddas, I give a list of the words of this dialect (but not the words used at threshing-floors), together with the Vaedi words, the colloquial Kandian expressions, and their English equivalents.

Where the letter N is suffixed the word is taken from Mr. NevilTs writings. In the cases where no word is given in the Vaedi dialect or Kaele-basa, both the Vaeddas and other hunters employ colloquial Sinhalese words, often slightly altered in pronunciation by the Vaeddas, as by using c for s, etc. It should be noted that in these lists the letter o is pronounced like the English ch in church. The Vaedi vocabulary which I append is extremely deficient, but that of the Kaele-basa is nearly complete.

The whole of the Kaele-basa vocabulary is not employed in one district. Many words are common to all Sinhalese districts, including even the extreme south of the island; others are found only in special localities.

It is interesting to note that in the north, when a Kandian hunter is addressed in the forest the title * Vaedda ' should be suffixed to his name ; for instance, a person called Banda would be addressed by his hunting comrades as Banda-Vaedda. Among the Wanniyas it is often the forest custom, when telling others to perform actions,, to prefix the interjection Ho, as Ho! Yamalla, e let us go *; Ho I WariM, '; Come ye/ When inviting persons to eat food they similarly prefix Ha, as H&l BaUgalla, * eat ye/ These interjections are not so used by Sinhalese.VOCABULARIES

125

English. Sinhalese Kaele-basa. Vaedi.

ANIMALS,

Ant-eater Kaballaya Potta, /. Pott! Malinyawa

Kabaellaewa Agya Gal-munda, N.

Talkola-pettiya Gal-gawara, N.

Axe-man Poro-kctraya Wa4^wa ?

Bear Walaha Uyangowwa Keriya

/. Waelihini Tadiya, /. Tadi Bala

Kaluwa, /. Kalu- Haecca, N.

Waelihini Wala, N.

Gamaya, /. Garni Araci, N.

Baraci, N.

Keri-bota, N.

Beast Srpawa ? Bota

Sip'a, N. 'Hipa,N.

Hatura, N.

Bee, honey MI maessa ? Ku Ja Maehikeli *

Ku4a MaeyikeH

? humble Bumbaeli- ? Penda-uli, N.

maessS.

? large Bambara ? Ma Maehikeli

Ma Maeyikeli

small Ku$a MI- ? Hin Maehikeli

maessa Hia Maeyikeli

Bird Kurulla ? Cappiya, /. Cappi

Sappi, N.

Buffalo Miwa Ambaruwa Manya, /. Manyi

GawayS, WalMannya,N.

Pimbinn§., N.

Cattle Harak BoilatN. Mannya. .N^.

Child Lamaya ? Kaekuia, pL Kaekul-

peto

Amml, N.

BiEnda, N.

Civet Cat UralaewS Appala-baetaeya - ?

Hotaembiliy§,

H6tambay3,

Cobra Naya B6yi-sattey§. ?

Cow Eladena ? Gonadena, N.

GonI, N,

Crocodile Kimbula Gamaya ?

Deer, Axis JPit-mti w& Ambaruwa Ke^tila

Ht-paelaeIM Kabara-bota.,, N.

Muwa-kewara, N.

? Mouse , MiminnS. Kekka, /. Kekk! Lembu, N.

Yaka^ayi Kewura, N,

DangarayS,, N.

M Red. Waeli-muwa Rat-ba£aya, N. . ?

ir; I

126

ANCIENT CEYLON

m

English. Sinhalese. Kaele-basa. Vaedi.

Deer, Sambar Gona Ambaruwa Gawara Mahagalla

Karakolaya Hela-kata,

Polla, /. PollI Keri-gona

Gawara, 1ST.

Gal-gawara, N.

Kata-kaebaela, N.

Magal, N.

Deity Deviya ? Hura "

E og Balla Aedura Kukka, f. Kikk!

Bandinna, /. Ban- Balo, N.

dinni Kotawa, N.

Hatara-bagaya

Hatara-baga-aetta

Elephant Aliya Uhalla Bota, /. Boti

Aeta (tusker) Usalla, /. UsaUi Bota-kanda

UsangaUa Goinbara-UhaUa

Gajjara, N. Aet-bota, N.

Bota-Kabala, N.

?"* Kada, N.

Kota baebela, N.

Mola, N.

Father Hya , ? Appa-latto, N.

- Appa

Fish Malu Mas ?

? smaU Ku ja-massan GembS ?

Uriyan

TittayS

Hare Hawa Yakdessa Yak-kadayS, N.

Yaddessa

Hornbm KSendetta Aeta-raetiya ?

(Tockus) Aettiriya

Raetiya

Hornet Debara Patarambaya Dembara-maecca, N.

' Iguana * Goy£, /? goyi KaeraeUa Munda, /. Mundl

(Varanus) Kapurala (aged) Go-kanda, N.

Manda, /. Mandl Go-munda, N.

Goy-bota, N.

Munda,*N.

? tied in a ? Mandu-walalla ? ?

circle for car-

rying

Kingfisher PilihiKJhiWci ? . Takan, N, (Tamil)

Kite, Brahinin y Ukussa Maralu ?

Lapwing (Labi- Kiram ? ' ? Aeti

vanellus)

Leopard Kotiya Diviyi Kapuru-balia

Siwupawa PoUecca

Baedi-mnta, N. Mita, N,

Raenaya, N. Mita-bota, N.

little one \ PettiyS, *? . Pecca, /."PeccI

VOCABULARIES

127

English. Sinhalese. Kaele-basa. Vaedi.

Lizard (Calotes) Katussa _ Kat-tomba, N.

Lord Swami Himi Hura

Man Miniha Hobaraya Min§, pi. Minu

Minna, N.

Monkey (Sem- Wandura Gas-gona Uda-kelinna

nopithecus) \ Gas-gona Keri-bandura, N.

Ma-bandura, N.

Monkey (Thev- Rilawa Kandan-paninna Hil-bandura, N.

sites] Patagahapu-eka Hossa, N. (?)

Nilawa, N.

Rila-pataya, N.

Mungus Mugatiya Appala-baetaeya ?

Ho-tambaya

Kg t}ra Hota-baraya, Dola, /. Doll

/. Hota-bari Hocca-dikka

Tadiya,'/. TadI Gal-bota, N.

Tel-kaliya Hossa-ulla, N.

Porcupine Ittaewa Ittaeya Iddaeya, N.

Katuwa Hicca, N.

Nittaewa, F.

Snail (land) Gombaella ? Tomba, N.

Snake Sarpaya Isa-gora-satteya Tola-gaecca

Sara-bota, N.

Urasa, N.

Turtle Ibba Biya-kukula, N. ?

Wife Hire ? Naena, N.

Woman Gaeni ? Mini, pi. Minila

MaeH, N.

Youth Ilandaraya ? Kaekula, /. Kaekuli

GENERAL NOUNS.

Areka-nut Puwak-ge Ji Golaeya (unripe) ?

Kahata-ge^i

Kaeppiti-ge Ji

Arrow Igaha : ' ------- Itala, N.

Moriyan-kecca, N.

Axe Porawa Kotanna At-barawa

Poro-wa^uwa Gal-raekki

Wa Juwa Porawa-pecca

Bark bag MaHa . ? Riti-malla

Patta-malla

? basket Pettiya ? Patta-pettiya

? of branch Atu-potta Rana-potta __

Betel-leaf Bulat kola Bol __

Bol-kola

Bol-pana

Daehaet

Paengiri

Billhook Kaetta Awudae Adinata

Kaeti-wa^uwa Mala-koti

128

ANCIENT CEYLON

$*???

m

English. Sinhalese. Kaele-basa. Vaedi.

Billhook, golden Ran-kaetta __ Manna-kaetta

Blade Tale ? Kaecca, N.

Blood Le Tuwala __

? track Le-para Kiri-para ?

Body Sarira ? Bo-pata, N.

Bow Dunna ? Dun-ga, N.

? string Dunu-lanuwa ? Dunu-^iya

Cave Gal-ge ? GaHge

Lena Rukkula, N.

Chillies Rata-miris Serapalu ?

Clan Gotra ? Warige

Waruge, 1ST.

Clouds Walahula -, ? Gal-periya, N.

Cold Sita ? Ehna, N. (?)

Cross-roads Man-handiya ? Ruppe.

Earth, (ground) Bima Bumi Polawe-waeli

Bimi, N.

Eye Aeha ? ; Eccel/N.

Fat Tel Kapparu, N. ?

Fire Gindara Ratta. Ginna

Gini Ginname

Ginna (body- Pimbinata

heat)

Fire-wood Dara Pirumbu ?

Flesh Malu Higaman, N. ? ? ;

Mas

Flock Raela K§ya, N. ?

Food (cooked) KSema Pas ?

Forest Mukalana Hlmala Baedde, N.

ITyana Gala, N.

Found things _^ Taegi (within ?

lebenawa)

Friend Yaluwa __ Kimbune, N.

Gourd Labu-ge4i Kalu-gaetaeya ?

Suraya

Titta-gaeti

Gourd, for water Labba Dabaraewa Panliya

PandSwa

Hair (of head)* s Isa-kes ? lea-Kola, N.

Halting-place Wagliya ? . Ruppe

Head Isa ? lea, N.

Oluwa Isi, K.

lya, N.

Heat Ushna ? ? Gina, N.

Hive (of Mi- Miya Gaba4aSwa Maehiketi-gama

maessa) Gaba J§,wa

Honey j Ml-paeni Bora Meccan, N.

j Micca, N.

Mijja, N.

, Mijjaka-peni, N.

- Mlaji-paeni, N,

VOCABULARIES

129

English. Sinhalese. Kaele-basa. Vaedi.

Honey-comb Mi-wada Padaya, N.

Horns An Aratu ?

Todii

Wakata«padaya, N.

House Ge ? Gal-gama

Tun-ge (hut)

Rukkula, N.

Jungle Kaele Himala ? .

Wai Uyana

Ladder (cane) We-wael ini- ? Ran-kendiya, N.

maga

Leg Kakula ? Danis patizl, N.

Lightning WiduHya ? Gini-waetunu, N.

Loin-cloth KSnama ? K6nam-po4ge

Millet Kurahan Alu ?

Jf cake Roti Alu-lelli ?

Alu-pota

Alu-piiwa

Gini-puwa ?;"

? flour Kurahan-piti Alu-ku4u -----

Moon Handa ? Pana-pojja, N.

Noise Sabda ? Anda, N.

Nose Nahe ? NayaN.

Piece Kaella ? Kate, N.

Porridge Talapa Kula ?

Rain Waessa ? Diya-pompa, N*.

Warusawa Waecca, N.

Waecca-wata, N.

Rice, cooked Bat Arambu ?

Bubbara

? uncooked mi Lawal-aeta Pubbnra,. N.

Wata-kuru,N.

? unhusked VI ICiri lawal-aeta Depatulan, N»

Depurul, N,

River Oya ? . Diya-gama, N.

Sea Muhuda ? Sawn, N.

Sky Ahasa ? Dewu-ula, N.

Stage (of jour- Hatakrna ? ' ? Pilluma

ney)

Stream Aela ? Diyaganaa^ N, l "'

Snout Hossa ? , Hocca

Sun Ira Suriya Ira-poj ja, K.

Thorn Katuwa Panaraela, N. ; . ?

Thunder Ahasa gaera- ? Dewu-ula ande, N,

wima

Tobacco Dun-kola Paepol-ro4« _

Sera-kola

Touchwood Tree Gaha Pirumbu Hongla Ga, pl/G&yi

Hon |e Guye (?), N.

Hon^eya

K130

ANCIENT CEYLON

English. Sinhalese. Kaele-basa. Vaedi.

Troop Raela Kaya, N. __

Village Gama ? No word

Water Diya Dira Diya

Watura Gangula, N.

Honcla

» Paedi

Paeheyiya

Paen

Wind Hulanga Wataya ; - ...y

VERBS, AND OTHER WORDS.

Arise, to (from Naegitinawa Sun elagannawa ? ?]

sleep)

Before Issarahata ? Iccata, N.

Behind Passata ? Paccata, 1ST.

Call, to Andagahanawa ? Kilapa, N.

Anda-talapa, N 1

Cannot Bae Issarata (with dative] ?

Passe puluwani

Catch, to (with Allanawa (Ata) gahanawa ?

hand)

Chew, to Kotanawa Kotabanawa ?

Chop, to (with Kotanawa Wa ^ulanawa ?

axe) Wa Jtiranawa

Cohabit, to Sar/wasaya ? Penna-ganna, N.

karanawa

Cold, to be Sita-wenawa _ _ Hita-gana, N.

Hita-pura, N.

Come, to Enaw§. Bandinawa, N. Enna (Fut. ist pers»

Pahasu-wenawa sing. Ennamma)

Cook, to (broil, UyanawS, Tambanawa ?

etc.)

Cut, to Kapanawa Man-karanawa, N.

Die, to Maerenaw§. ? Miya, JM.

Miyala inda, N,

Miyala ya, N.

Sala ya, N.

Dig the ground, Polawa Bin-dinawa ?

to haranawa

Distant i Agta ? Naekat, N.

Do not ? Issara^a ?

Passe puluwani

Drink to Boaawi, Balanawa Diya ka, N.

"Uka, N.

Eat, to Kknawa Balanawa Ba$a,-banda, N,

Kapa, N.

j KSpat-ena, N.

VOCABULARIES

English. Sinhalese. Kaele-basa. Vaedi.

Go, to Yanawa Baendenawa Mangaccana

Haerenawa (Fut. ist pers. sing*

Issarahaeta bahi- Mangaccan.

nawa Imp. 2nd pers. sing.

Karana karanawa Mangaccapa).

Issara balanawa, N. Palacca, N.

,, to jungle Kaeleta yan- Himala wadinawa ?

awa

Here MeM ? Metta

Hold, to Allanawa Bandinawa ------ ;,

Hot, to become Unu-wenawa -- Radaga, N.

Hungry, to be Batfagini allan- ? Balaga, N.

awa

Insufficient Madi Boyi ?

Journey, to Gaman-karan- ? Gamana

awa Mita-gana, N.

Jump, to Paninawa ? Bota-dama, N.

Duwa, N.

Paena, N.

Paran-dena, N.

Kindle (fire), to Pattu-karan- Mahi-karanawa Awuccana, N.

awa Pawatanawa, N.

? (by cross- Mandinawa Mandinawa ?

friction)

? (by twirl- Gini-ganawa Gini-ganawa Gini-gana (

ing-stick) Gini-gahen ginna

ganna

Large, great Maha Gaja Ma / !

Little Kuan ? Taenak, N.

Punci

Long Diga ? Dikka

Mend (fire), to Bo-karanawa Mahat-karanawS ?

PawattanawS

Meet with, ob- Samba-wenawa Haeppenawa ?

tain

No NSS Boyi Kocjawa1

Not Naeti . . ? Ko^awa

Pudendum (/.) ? ? Hantana patiliya, N.

Mala kabala, NT.

(*0 ? ? Hantana pole, N.

Quickly Lahi ? Numa, N.

Wigahata Peramat, Bf.

Bain, to Wahinawa Gangula-banawS, N. Diya pompa, N.

H.un, to Duwanawa ? Paena ucca, N.

See, to DaMnawi. ? Reviga, N.

Shoot (withi ar- Widinawa ? Penna-ga, N.

row), to Wida-bacca, N.

Wida-ga, N.

1 Compare Sin. nokatfanta,, always, ever (lit. unbroken); without the negative particle kadamt would thus mean * never.*132

ANCIENT CEYLON

English. Sinhalese. Kaele-basa. Vaedi.

Skin, to Hama-gahan- Honda-karanawa ?

awa

Sleep, to Budi-wenawa B5yi-karanawa Nida, N.

Ninda-gan-

nawa

So much Occara ? Witara, N.

Speak, to KiyanawS Galu-karanawa, N. Bana, N.

Guwa-kara, N.

Guwa, N.

Katan-kiya, N.

Kewili-la, N.

There Ohe ? Otta

? (distant) Aha ? , Aramba

That side Oya petta ? Oba

Obba, N.

Otaka, N.

This side MS-petta ? Meba

Mebaga

Mobba, N.

Tie (an Iguana), ? Manduranawa ?

to

Walk, to Ewidinawa Tamananawa Gaman ila gana, N".

Gaman ina, N.

Mita gana, N".

Where Kohe ? K6de, N.

Whet, to * Madinawa Honda-karanawa Koyde, N,^

'Why Eyi ? Ayi

Yes Ha ? Ha

£h? Ita, N.

Own

You To (thou), pt. ? Toba, N".

Topi Toban, N.

Umba Topa, N.

Topan, N.

Young Lapati *_ Eiama, N.

Xaruna

IV

THE RELIGION OF THE VAEDDAS AND

WANNIYAS

THE DEITIES

THE Vaeddas are not Buddhists, but a few who live in the villages of the interior sometimes visit a Buddhist temple and offer flowers there, if there is such a place in their neighbourhood. Neither the Forest nor Village Vaeddas pay any regular worship to the superior Indian Gods under their present names, although some of the latter make occasional offerings to Skanda, as the god of Kataragama.

The northern Tamil-speaking Vaeddas to the south of Trincomalee periodically visit and present offerings at the Hindu temples of their vicinity ; these are devoted to Skanda, who is known as Kumara Tevan, and Ganesa, called by them Pulikara Tevan, * the God Pilleiyar/ They also make offerings in the jungle to seven Goddesses, termed * the Seven Kannimar/ or Maidens, whose names they do not know.

The Wanniyas of the north-eastern part of the North-central Province describe themselves as BuddMsts, and sometimes pay visits to the Buddhist temples near them; and the older men at times take the eight vows, or * Atasil/ of the Upasakas or lay-devotees. These are the vows of adherence to the prohibitions against murder, theft, falsehood, drinking intoxicating liquor, and unlawful sexual intercourse, forming the * Pansil/ or five vows that all devout Buddhists should keep ; and three additional ones against nocturnal eating, the personal use of garlands and perfumes, and sleeping on anything but a mat laid on the ground. Persons who adhere to the Atasil are expected to be regular in their attendance at the

153I34 ANCIENT CEYLON

temples on the Poya days, at the quarters of the moon, and also must particularly avoid taking life of any kind.

With possibly the sole exception of the Wanniyas, all Village Vaeddas alike, according to my information, when cautiously interrogated, acknowledge that their chief deity is the God whom they know as the Gale Yaka, although some who afterwards showed me a temple erected to him in their village at first denied all knowledge of him when I questioned them regarding him. The Vaeddas of the interior, including those of the Madura-oya valley, stated to me that all worship him. One man remarked to me, "He is the greatest of all Gods." As the ancestors of many who are now classed as Village Vaeddas were certainly true Forest Vaeddas three-quarters of a century ago it would appear that the Forest Vaeddas cannot be ignorant of this deity, but most probably have the same belief in him. I mention this as Dr. SeUgmann, who has recently investigated the religion of the Forest Vaeddas, has informed me that he obtained no information regarding" Mm from the Madura-oya villages.1 After my own experience of a refusal of Village Vaeddas to divulge their worship of him to me I am inclined to believe that their knowledge of him was intentionally concealed.

Literally the name means c the Demon of the Rock * ; but as their male deities are all termed ' Yaka,' whether beneficent or malevolent, the true signification of the expression, as it was explained to me in Sinhalese by them, is ? the God of the Rock/ a name identical with that by which he is known and worshipped by the Kandian Sinhalese, who call him the Gale Deviya,

The Tamil-speaking Vaeddas term him Malei Pei, ' the Hill Demon'; Kaflu Pei,, ' the Rock Demon '; Maleiyam, ( He of the Hill'; and Maleiya-swami,' Hill Lord/ I give an account of his worship in a separate chapter ; it appears to be the primitive cult of the island. It was apparently this deity

1 He is mentioned by name as the * Indigollae Yaka * in one invocation of the Forest Vaeddas which Dr. Seligmann was good enouglx to me, and another was apparently addressed to him as the * King-of theTHE DEITIES 135.

who was known as the Vyadha Deva, ' the Vaedda God/ in the time of Pandukabhaya, that is, in the fourth century B.C. ; and possibly he was the God mentioned under the name of Puradeva, ' the Ancient God/ in the time of Duttha-Gamini (Mah. i, p. 100), as having a temple at Anuradhapura.

The Village Vaeddas also greatly reverence a Goddess known as the Kiri-Amma, which means in Sinhalese ' Grandmother/ but in reality in its use by the Vaeddas is equivalent to the Indian and Sinhalese word Devi, '? Goddess/ which is not employed by them.

Mr. Nevill considered her to be a form of Parvatl, the wife of the God Siva, who, as the mountain-born goddess, the daughter of the Himalaya personified, is called in Tamil Kiri-Amman, ' the Hill Mother/ kin being the Tamil form of the Sanskrit word gin, hill or mountain. So far as the name and position of the goddess are concerned, this identification appears to be quite satisfactory. On the other hand, it is to be observed that the expression ' Kiri-Amma' is applied by both Vaeddas and Sinhalese to other goddesses who are not connected in any way with hills. For this reason the name itself cannot be accepted as proof that the goddess in question is really Parvatl, unless there is some confirmation of her personality from other directions. It is also noteworthy that the Tamil-speaking Vaeddas do not recognise her, although they worship other Hindu gods.

In the villages of the interior, where the ancient traditions and practices are better preserved than near the coast, where Tamil influences have affected some of the religious notions of the Vaeddas, this goddess is held to be the most important deity next to the God of the Rock or Hill, whose wife she is supposed to be. This would tell strongly in favour of her being Parvatl if the God of the Rock were Siva. Yet we do not find her worshipped on the hills like the Gale Yaka ; she is chiefly, if not entirely, a Forest-Goddess, and it is to her that the Village Vaeddas of the interior especially appeal for protection and good-luck in hunting, their chief occupation. She is known as the Indigollaewa Kiri-Amma?her husband the Gale Yaka being also called the Indigollaewa Yaka?or*..__

\f I

136 ANCIENT CEYLON

the Kukulapola Kiri-Amma. The last place is a village of the Vaedi-rata (as the district inhabited by the Vaeddas is termed), on the west bank of the Madura-oya; while Indi-gollaewa is in the North-central^ Province, near Kalawaewa, and there is a temple at it devoted to her and her husband.

According to the tradition of that district she is not Parvati, but Mohini, the beautiful incarnation of Vishnu, who was also a Hill-God,1 and who, according to Indian authorities, took her form temporarily so as to enable the Gods to cheat2 the Demons at the celebrated Churning of the Ocean by the Gods and Asuras (or demons) in order to produce Amrita, the Liquor of Immortality. The admiring and unsuspecting demons agreed that the lovely Mohini should divide the Amrita between them and the Gods. She separated them into two rows, and then distributed the whole to those in the Gods' row, to whom she gave it first.

One Asura sat among them, and thus obtained a share of the precious drink. The Sun and Moon observed this and pointed him out to Mohini, who promptly cut off his head; but the inagic liquor had already conferred immortality on him, and therefore Brahma, who always found a way of dealing with apparently insuperable difficulties, transformed the two parts into heavenly bodies. The tail or body became a comet, and the head a planetary sign called Rahu, which as a heavenly dragon endeavours by way of revenge to swallow the Sun and Moon, and thus causes eclipses. On account of the unfair treatment of the demons by the Gods on this occasion, there has been undying feud between the two classes of supernatural beings from that day.

Some confirmation of the identification of the Kiri-Amma as Mohini is to be seen in the fact that throughout the interior of Ceylon, Ayiyanar, the son of Mohini, is everywhere considered to be also a Forest-God, who specially guards travellers in the forests and jungles when they appeal to him for pro-

1 In the Rig Veda, i, 154, 3 (Griffiths' translation), he is called

4 The Bull far-striding,, dwelling on the mountains.'

1 Her name is deitved from the word Mohani, delusion, fascination, and in South India ahe is always colloquially termed Mdhani.ET.

3"

P ?f.

£

THE DEITIES 137

tection?and is certainly not a Hill-God. On the whole evidence, therefore, and especially since the Vaeddas do not treat her as a Hill-Goddess, I am inclined to accept the only native explanation of the identity of the Kiri-Amma which I have been able to find, and to look upon her as a form of Mohini rather than ParvatL

Seven other Goddesses, who are also termed Kiri-Ammas, are revered collectively in the south. They are stated to have been originally influential chieftainesses who have been deified, possibly in comparatively recent times. Their names are given as : (i) Miriyabaedda Kiri-Amma, the most important of them; (2) Pusmaraga Kiri-Amma; (3) Unapana Kiri-Amma ; (4) Kosgama Kiri-Amma; (5) Bowelagedara ? Kiri-Amma ; (6) Balagiri Kiri-Amma; and (7) Ginigal Devatage Kiri-Amma. The last one evidently belongs to a different class from the others, and is clearly the Sakti or female manifestation of the minor deity called Ginigal Devata. There is some doubt regarding the class to which Balagiri Kiri-Amma belongs. These are all beneficent deities, that is, Goddesses.

Of the same class, according to Mr. Nevill, is Bowala Deya, who may be connected with the fifth one in the list just given. He is beEeved by Mr. Nevill to be a late instance of the propitiation of a local chief who became an evil spirit after his death. He resembles the numerous Bandaras of the Kandian Sinhalese, by whom more than one hundred are enumerated, some having protective powers though all are ranked as demons. Panikki Vaedda, already mentioned, is included among them; Dr. Seligmann informs me that a person of this name is a spirit deity of one group of the Forest Vaeddas. He traced the belief in him to a Vedarala, or village doctor, of mixed descent. The older men do not recognise him.

According to the description of them supplied to Mr, Nevill, a group of deities called the Uda Yako, ' the Upper Yakas/ or ' the Yakas who live Above/ are the most important gods of the Vaeddas. Their individuality is stated by him to be ill-defined; they occupy the position of superior nebulous spirits who are ' like little children/ and who apparently neither do much good nor much harm to mankind. The

138 ANCIENT CEYLON

expression evidently merely corresponds to the collective term, Atala Deviyo, of the Sinhalese, ' the Gods of the Upper World/ in contradistinction to the Patala Deviyd, the Gods of the Lower World/

After the Uda Yako, Mr. Nevill places the Bilindu Yako, literally ' the Children Demons/ said to be a father and son, the mother of the latter being thought to be the Kukulapola Kiri-Amma, while the child is supposed * to have died shortly after birth, and to be now separately gifted with divine powers/ Some state that the Bilindu Yak5 are two brothers ; others say that they are seven in number. It is apparent that their identity is uncertain. I have no knowledge of them, but perhaps it may be assumed that the first account of them, being the most definite, is the most likely to be correct.

The younger one is said by Mr. Nevill to be the Ilandari Devata of the Coast Tamils, who is also one of the deities of the Wanniyas, and is known to the Kandian Sinhalese as a son of the Kiri-Amma, thus confirming the relationship.

The elder one is identified by Mr. Nevill as a Mude Deyiya, or * God of the Sea/ of the Coast Vaeddas, but I am not aware for what reason, as the Sea God is not described by him as being the husband of the Kiri-Amma. As her consort, the elder Bilinda Yaka would appear to be the Hill God. The functions of these deities are said to be warding off disease, granting food, and generally protecting their worshippers from unseen dangers; they are thus similar to those of the Hill-God.

Mr. Nevill also mentions a Ma Yakini,1' the Great Goddess/ who appears to be the Indigollaewa Kiri-Amma; (2) Una-pana Yakini, who is evidently the Unapana Kiri-Amma; (3) Kino-mal Nacci,2 * the Blue-lotus (coloured) Lady/ whose symbols are an arrow and a bowl of water, and who is the protectress of hunters, and therefore apparently another form of the Indigollaewa Kiri-Amma ; and (4) 'a Baedi-Maeli, ' the Woman of the Jungle/ who has similar functions and

1 Yakini, or Yaksani, the usual colloquial, expression, Is the feminine form of Yaka.

2 The letter C is pronounced as Ch in all transliterations.THE DEITIES 139

may be the same person. These deities are beneficent, and therefore are Goddesses.

I also heard of one called Gdmbara Nacci Yaksani, * the Freckled Lady ' ; she is said to be the wife of the Gale Yaka. She must therefore be a form of the Kiri-Amma, all these deities being, like the Vaeddas, strictly monogamous.

So far as I learnt, there is only one other kindly deity, the Gange Bandara, ' the Chief of the River/ whose aid and countenance are sometimes invoked in the interior when the Vaeddas are about to proceed on a hunting trip. He is known also in the north-western Kandian districts. The Vaeddas locate him at Yangala, near the Mahawaeli-ganga, but he has no temple there.

King Maha-Sena, who is worshipped by the Kandians and the Wanniyas under the title of Sat Rajjuruwd, is not known to the Vaeddas, either in the form of a deity or otherwise.

Next come the malevolent deities also called Yakas, a title which in their case means { demons/ They cause the various evils that afflict the Vaeddas, and their position is exactly the same as that of the evil spirits whom the Sinhalese denominate Yakas. Sickness or misfortune of every kind is especially attributed to them; but some of them also exhibit their spite by throwing down rocks from the cliffs when people are passing by, and by frightening them in the night by strange cries and noises, the latter including the clapping of hands.

These other Yakas, who are thus generally harmful, belong to two categories, those known by names found only in the interior tract of the Vaedi-rata, and those recognised in other parts,

Of the former, Paeraet Yaka is one of the most powerful; he may represent the Daedimunda Yaka of the Sinhalese districts, if, as is probable, his name contains a reference to his being defeated (paeraeddund) by Buddha in his conflict with the forces of Mara. According to the Sinhalese accounts, this demon alone did not run away when Mara's army was defeated, but crept under Buddha's throne ; he was therefore called by Buddha " Daediya," 'the Resolute One,' a title140

ANCIENT CEYLON

that does not strike one as being particularly applicable under the circumstances. Daedimun4a is connected by a legend with the construction of Alut-Nuwara or Mahiyangana, apparently an early Vaedda settlement, which has been already mentioned as the site of one of their early battles. He is also known by the Sinhalese as Devata Bandara, ' the Codling Chief/ and Alut-Nuwara Bandara, * the Chief of Alut-nuwara/ As Daedimunda, I was informed by them that he is considered to be an ' Aemaptya/ or minister, of Vishnu, and the son of a demon termed Manawaka or Manoka, who was the spirit of a Mana or Black Stork (Dissura episcopa], which was caught by a pandita named Widura,1 yoked to a plough, and forced to plough a field for him. Daedimunda is considered by the Sinhalese to be the most powerful of all demons, and to rank next to Wessawana, or Kuvera, the Overlord of all Yakshas. It seems possible that the Vaedi name l Paeraet' has suggested his identification with a Yaksha of the Sinhalese. His connection with Mahiyangana, which he is said to have built, may indicate that he was an original deity of the Vaeddas.

Among the Yakas of this class mentioned by Mr. Nevill, the Kumbe Yaka is noted by him as being the most dreaded by the Vaeddas, and as being a very powerful and vindictive demon. I have no other information regarding him, and not having a full account of his attributes and special powers I cannot state if he is a demon of the Sinhalese. Their Saeda or Haeda Yaka is said by some to hold the next rank to Daedi-muruja. His name, which means * strong ' or * cruel/ shows that his character is like that of the Kumbe Yaka; but it is uncertain if the two are identical.

There is also a possibility that the name of the powerful Kohombe Yak& of the Sinhalese, and of Southern India, * the Demon of the Margosa (tree) * may have been contracted; or perhaps it may be derived from the Vaedi name Kumbe Yaka. There are reasons for believing that the Kohombe

1 The hero of the Vidhura Paadita Jataka (No. 545). It was as a punishment for his cruelty to the unfortunate Stork that the demon Punnaka subsequently ill-treated the saint.THE DEITIES 141

Yaka is an original demon of the Vaeddas ; he has as subordinates three ' Vaedi Yakas/ and also twelve, others who are termed ' Vaedi Kadawarayo/ and who are of inferior rank under him. The Kumbe Yaka is not likely to be the zodiacal sign Kumbha (Aquarius), which is represented by the Sinhalese as a demon with a human body and the head of a dog.

The Gini Rahu Yaka, ' whose worship seems a little mixed up with a Goddess known as Alut Yakim, the New Demoness/ who is represented as his wife, is thought by Mr. Nevill to be a form of Agni, on account of his name, and his symbol, a burning torch. This identification seems to me to be doubtful, partly on account of the character of the Saktl, who is said to be connected with water, and apparently to have been produced in the sea ; and also because no worship is paid to Agni in any form by the Sinhalese, so far as I am aware.

He is more likely to be a form of the demon Rahu already mentioned in connection with Mohim, than of Agni. The Sinhalese consider that Rahu the dragon has the form of a snake, and that he rides on a horse, holding a fish as his symbol.1 This indicates some connection with water. Offerings are made to Rahu by the Sinhalese, as well as to other planetary signs, which are all considered to be evil demons who afflict mankind.

There are only three deities to whom fire-worship, termed Gini Mangalya, is paid by the Sinhalese; in order of importance these are Pattini Devi, an incarnation of Durga, Vira-munda Deviya, and Devol or Devel Deviya. All are importations from Southern India. The Tamils of India and Ceylon have fire ceremonies in honour of Virabhadra, Draupadi and Devi-Amman, who is either Pattini or a form of Saktl, the daughter of Daksha and wife of Siva. The Gini Rahu Yaka does not appear to be one of the two male Sinhalese

1 The description of Rahu in Hindoo Castes is taken from Ward

as follows :?' This god is painted black ; wears black garments ; rides on a lion ; has four arms, in which he holds a scimitar, a spear, and

a shield, and with the other hand he gives a blessing." The Vaedi

name may mean (the Fiery Rahu Demon/142 ANCIENT CEYLON

deities, who are termed Gods and not demons. Torches are used freely in many dances in honour of demons, and in themselves are.no proof of any connection with fire-worship.

If the Wana-gatta Yaka is, as Mr. Nevill said, the deity whose power extended over the forests, and who 'is propitiated only when great want of success follows the huntsman's toils/ his character among the Vaeddas is very different from that of the Forest-Deities of the Sinhalese and Wanniyas-Mr. Nevill described this demon as ' a dreaded spirit, and the sacrifices and incantations are regarded with intense terror* and faith, and only resorted to in extremities/ * He stated that the offerings made to him are ' clothes and blood/ and he therefore considered him to be ' clearly a form of Bhairava/ Blood, however, is offered by the Sinhalese to many demons, if not to all. Clothes are not presented to any demons by them, I believe, and in any case would appear to be an unnecessary and inappropriate gift to beings of this type. Have the irresistible forces of civilisation begun to affect their ideas of propriety ?

Bhairava is not a Forest God; in Ceylon he is known as Bahirawa Deviya, and as a form of the God Siva he is a deity of the Underworld, or Patala. In Ceylon his special function is acting as guardian of sacred edifices such as dagabas or wiharas, and treasures, and everything underground (see Fig. 160). He has eight forms, termed the Ashta Kali Bahirawayo, to whom, collectively, worship is paid by the Sinhalese. In another aspect he is a truculent demon called Bahirawa. Yaka, whose duty it is to punish those who break into temples, or digabas, or who open the ground (as for wells or mining* work), or excavate treasure, without first obtaining the per-of the BaMrawa Deviya.

The characteristics of the Wana-gatta Yaka, as above-noted, do not, if he is a Forest Deity, point either to Bahirawa, or to the usual Forest-Gods, whose functions are decidedly

1 TetprQbaniant Vol. i, p. 196 ; at p. 183, however, he is identified

by Mr. with the 'Wanni Deva/ that is, the Wanni Deviya

fWanui of the northern part of the North-central Province),

h by to be Ayiyanar, but who possibly may be Bilinda

Yak5, lie is not a demon, but a Guardian Forest God.THE DEITIES 143

*

protective, and who are appealed to on the slightest pretext by all travellers in the jungle, whether under the name of the Wanniya Bandara of the Wanniyas, or the Wanni Deviya, or Ayiyanar, or Skanda as the ' Kataragama God/ of the northern and north-western Kandians.

Regarding Kalu Vaedda Yaka I have no knowledge. Mr. Nevill identifies him as the Kalu Yaka, ? the Black Demon * of the western Sinhalese, apparently only because both are termed ? black' ; but there are several who are similarly described, and it is very unsafe to trust merely to similarity of names for an-identification of these obscure deities.

All Sinhalese recognise ' Vaedi' Yakas to whom offerings are constantly made on account of the illnesses of men and women, but chiefly the latter, and especially after childbirth. If his functions are similar, the Kalu Vaedda Yaka may be one of these demons, the Kalu Yaka not being known to the Sinhalese as a 'Vaedi' Yaka, but as an Indian prince of Madura, who became a demon after his death. The latter's usual designation is Kalu Kumara Yaka or Bandara, the ' Black Prince Demon' or ' Chief ; and King Gaja-Bahu I (113-135 A.D.) is believed to be an incarnation of him because of the cruelties traditionally attributed to him during his invasion of Southern India, in revenge for the deeds of the Soliyans in the ' War of the Short-horned Buffalo/ during his father's reign.1

Kande Yaka, 'the Yaka of the Hill/maybe, as Mr. Nevill says, the Gale Ban Jara, 'the Chief (or minor deity) of the Rock/ of the Sinhalese ; but this assertion is not conclusive in the absence of some further basis than the resemblance of the names. He is said to haunt precipices, and only to be invoked locally, whereas the Gale Bandara is appealed to throughout a large tract of country in north-western Ceylon, A Kande Bandara is found in a list of deified chiefs and other deities contained in a very old manuscript of the Kurunaegala district, and he may be the Vaedda deity. There is a possibility that the name may be another term for the Gale Yaka,

r 1 This demon is illustrated in Callaway's Yakkun Nattanawd, 1.829, p. 4, and is there dressed as a Kandlan Chief.144 ANCIENT CEYLON

t

whom I have mentioned as the highest god of the Vaeddas, of whose status Mr. Nevill appears to have been unaware.

These rock and hill deities are difficult to identify without acquaintance with their attributes and functions. The Sinhalese recognise three if not four special forms of them: (i) the Gale Deviya, ' the God of the Rock/ who is undoubtedly the Gale Yaka of the Vaeddas; (a) the Gale Bandara, ' the Chief of the Rock/ an Indian demon who arrived on the southern or western coast with others, in a boat made of stone, and is by some confounded with No. i; (3) the Kande Bandara, * the Chief of the Hill/ about whom I have no information; and (4) Kuranaegale Postima Bandara, ' the Chief Postima of Kurunaegala/ a local demon of the Kurunaegala district, who is the spirit of the Prince Postima who was thrown down the precipice at the Kurunaegala rock.1

Mr, Nevill states that Mara, the personification of Death, is largely invoked by the Coast Vaeddas ' as a later conception of Kumbe Yaka as the God of Death, the opponent of youthful vigour/

Among the southern Vaeddas, I found that offerings are made to a second series of sickness-causing demons, several of whom are identical with those of the Sinhalese, from whom they appear to have been borrowed. These latter are as follows, each having a Sakti, or female manifestation of the same name, who is considered to be his wife, this being a local development, and I believe unknown to the Sinhalese.

NAME OF DXMO*. NAX* OF SAKT!. FUNCTIONS.

T. Sana! Yak& Saxmi Yaksanl Causes Sanni-roga (sickness

accompanied by convulsions, such as cholera), and other illnesses, and is the worst of the group,

2, Htalyaa Yak!,* Hfiniyan Yaksani Causes the ills that follow

curses, magical spells, and the glance of the Evil Eye.

1 See the account of it in Ehmn Years in Ceylon, by Major Forbes. Vol. i p» 194.

* in p. jo. In the plate he rides

en a^ a in the left hand and fire in a vessel in

the He Is by coil about his limbs

andTHE DEITIES

145

NAME OF DEMON.

3. Malia Sona

Yaka or Sam-aya Yaka *

4. Wata Yaka

5. Siri or Riri Yaka

6. Maralu Yaka

7. Abimana Yaka

8. Bills Yaka

NAME OF SAKT!. Maha Son! Yak-sam

Wata Yaksani Siri Yaksani

Maralu Yaksani Abimana Yaksani

Bille Yaksani

FUNCTIONS.

Causes fever, head-ache, and pains in the body.

Afflicts the stomach, limbs,

and eyes. Causes fever, loss of appetite,

and pains in the limbs. Causes ulcers and tumours. Spits at men, and takes away their appetite. Causes head-ache, tears, etc.

In addition to these there is the Kumara Yaka, ' the Prince Demon/ who is most likely the Kumara Bandara of the Kan-dians and the Sinhalese of the western coast, a son of a King of Madura and brother of the Kalu Kumara Yaka. He cannot be Skanda, the War God, who is also called Kumara, as he has distinctly demoniacal traits, and among other things causes fever and swelling of the body?probably dropsy.

There are other evil female deities : Siri Kadawara Yaksani, and Madana Siri Yaksani, whose names indicate a belief in the respective male forms Siri Kadawara and Madana Kadawara ; and at least three termed Giri, the feminine form of Gara, are known in the south, where they afflict women and children.

There are no less than twelve demons who are especially called * Vaedi' Yakas. Unfortunately I could not obtain their names, as my informants, although Vaeddas, were unacquainted with these details. They are said to be extremely malignant, so much so that if they strike one recovery is impossible. Mr. Nevill refers to apparently the most important one under the name * Maha Vaedde Yaka '; another is clearly the Kalu Vaedda Yaka already mentioned, and possibly some of the others whose names have been given belong to this class.

The Kandian Sinhalese also recognise twelve Vaedi Yakas ; two of them may be the demons termed Pudana Vaedi Yaka

1 Illustrated in Yakkun NaUan&wa, p. 7, as a black demon with a beat's head, who carries a* spear in his left hand, and an elephant (which he is about to devour) in his right. He rides on a pig,

L146 ANCIENT CEYLON

and Hella Vaedi Yaka, who, however, are not so deadly as the Vaedi Yakas of the Vaeddas. I have already stated that there are three Vaedi Yakas who are subordinate to the Kohombe Yaka, and twelve others under him in an inferior position who are termed Vaedi Kadawarayo. Very little appears to be known about them. Lists of these attendants and other particulars for which I am indebted to the kindness of Mr. Codrington of the Ceylon Civil Service, do not throw much light on the subject, as they do not seem to contain the names of Vaedi deities.

Lastly, there is Kurumbuda Yaka, ranked by the Sinhalese as a very malignant (wasa napuru) demon. He is the minister and attendant of the Gale Yaka, whom he accompanied from India.

Doubtless there are others of whom nothing is yet known, and especially an immense array of nameless minor local demons, who are found throughout the whole country, inhabiting rocks, and pools, and trees, and waste grounds. Dr. Seligmann has stated at a meeting of the Ceylon Branch of the Royal Asiatic Society that he found the Forest Vaeddas largely worshipping a number of spirit deities, who are the spirits of deceased persons, their relatives. He will doubtless give a full account of them in his work on the Vaeddas. He terms them the Nae Yakas, * the Kinsfolk Demons * or deities.

On a review of this formidable list of the deities, beneficent and malevolent, of the Vaeddas, it will be seen that the close connection between their religious ideas and those of the Sinhalese, and especially the Kandian Sinhalese, ex-elusive of Buddhism, is very prominent. In fact, the chief line of divergence lies in the deification of the Kiri-Animas and the recent Nae Yakas. Sacrifice of some kind is by the Kandians or the Sinhalese of the western coast districts to practically all the other deities of the Vaeddas, with the exception of Bowala Yaka, and possibly of Kurabe Yaka, and Wana-gatta Yaks. I have no acquaintance with the the God or of the Sea, of the Fishing

or ; he may be a deity borrowed from the Sin-THE DEITIES 147

halese or Tamil fishermen with whom they come in contact,, or possibly is aboriginal.

The point of particular interest is the supreme position assigned to the God of the Rock, or Hills, as a beneficent deity. Nominally, at least, he does not hold this rank among the north-western Kandians, who term him merely one of the most powerful of the demons, and one who does not often trouble himself with the affairs of men, although as a matter of fact they often appeal to him for assistance in case of the outbreak of epidemics or great want of rain. He is quite unknown to the Sinhalese of the western coast. I think there can be no doubt that he is the Hill God of the wild tribes of the South Indian hills. The legend regarding his arrival in Ceylon, and the particulars of his worship as it still survives among the Village Vaeddas and Kandians are given in the next chapter.

The statement of the Vaeddas that the Goddess known as the Indigollaewa Kiri-Amma is his wife finds confirmation in the North-central Province, where the same temple at Indigollaewa is the local centre of the cult of both deities. If the Kiri-Ariima is really Mohini, as I was informed, we have here a cult that has been to some extent developed independently of India, and that perhaps may be connected with the legends respecting the conquests of the Asuras by Skanda and Ms half-brother Ayiyanar. People who found themselves surrounded by such a numerous band of evil spirits as those of Ceylon naturally would be inclined to pay honour to deities like Mohini, who had proved themselves their outwitters.

After the demons had been cheated by the gods at the great Churning of the Ocean, Mohini regained the form of Vishnu, and left the scene. The God Siva subsequently heard of the incident, and proceeded to Vishnu, who to satisfy Ms curiosity resumed the shape of the fascinating beauty. The susceptible Siva was overcome by her charms, and the result was the birth of a son after Vishnu had returned once more to his own form. We"1 learn this from the Bhagavata PurStaa.

148

ANCIENT CEYLON

The infant made his appearance in the world from the back of the God's right hand, and received the name Eiyanar (in Tamil), or Ayiyanar (in Sinhalese), or, according to the Tanjore temple authorities, who may be taken to represent South Indian opinion, more correctly Keiyanar, 'He (who was born) from the Hand/ He is also known as Nayanar, and as Hari-Hara-putra, the son of both Hari, or Vishnu, and Hara, or Siva. His colour is dark blue or black, and in Ceylon his Vahana, or the ' Vehicle' on which he rides, is a white elephant. In India he rides both the elephant and the horse.

The vahanas, an elephant, a bridled horse, and apparently another smaller animal, part of the head, of which appears behind the horse, at the side of a giiardialn deity carved in high false relief on a pillar at one of the wahalkadas, or ornamental altar-backgrounds, at the Jetavana dagaba at Anuradhapura, show that the figure may represent Ayiyanar. He wears a cloth from the waist downwards, and has the usual heavy ear-rings, two jewelled necklaces, and large armlets and bangles. He holds an upright cross hanging from the fingers of his uplifted right hand. On the other side of his head flames emanate froiii a chalice-like object carved in relief on an upright slab or stele. In the panel below him is represented one of his wives, who carries a large flower or bouquet in her left hand, and has a long twist of hair hanging down on her right breast (Fig. 37). This carving probably dates from the early part of the fourth

century A.D.

In the reliefs on the gopura of the eleventh century A.D. at the Tanjore temple he is represented as a chM on a diminutive elephant, and has one face, but twelve arms (Fig. 39). Pushkala and Purna or Purana were his two wives.

A different account of him is given by Dr. Burgess, in describing the rock-cut temples at Badami in the Dekhan.1 It is taken by Mm from Foulkes's Legends of the Shrine of Harihara, and agrees with the Badami carvings, which may be four or five centuries earlier than those at Tanjore. In this story Harihara had a different*origin and mission. It was 1 The Indian Antiqnary, Vol. vi, p. 358.

ores P

tfTHE DEITIES 149

Siva who assumed the form of Vishnu in order to destroy an Asura called Guha, who by his austerities had obtained powers from Brahma which enabled him to conquer the Gods, and turn them out of their paradise. Siva killed the demon with his magical arrow.

In accordance with this legend the statue at Badami represents a figure who is half Siva and half Vishnu. He has one face and four arms, and carries as a battle-axe the crescent-shaped 'Keteriya' of the Sinhalese, round the handle of which a cobra is curled. He wears a high crown, the frustrum of a tall cone, decorated with symbols of the God Siva on the right half.

According to a Sinhalese tradition Ayiyanar came to Ceylon from Madura. His name seems -to show that his place of origin was in a Dravidian country. The honorific title Naya-nar, * the Nayar.' appears to indicate that he was originally a deity of the Nayars. If so, he was an early South Indian god, and his high position as the special Forest Deity of Ceylon may be due to his introduction by the early Nagas. If he were not aboriginal it is unlikely that he would be thought so important in the forests of Ceylon.1 The story of his wonderful birth must be a later invention in order to bring him into the Hindu pantheon.

After the account of his birth had been generally accepted it would be logically concluded that if the father, Siva, and the son, Ayiyanar, deserve worship, so must also the mother of such a son, especially as she was an incarnation of Vishnu, and had acted so successful!^ against the demons.

The Vaeddas, who were in close contact with the Sinhalese of the interior, probably acquired from the latter their knowledge of this goddess, and adopted her either as suitably filling an unoccupied place in their pantheon, or as being identical with some pre-existing goddess of theirs. This Mdhini worship must therefore be a later addition to their cult. Neither the Vaedda nor Sinhalese traditions which are given in the next

1 In South India Ms special function consists in his acting as night-watchman of the villages. In performing this duty he rides on a horse.$

*'*' '"&'

fell

ilfti

150 ANCIENT CEYLON

chapter mention the Kiri-Amma as having accompanied the Hill God when he came to Ceylon from India, but on the contrary state definitely that he had with him only one attendant minister. This is strong evidence against Mohmfs being an aboriginal goddess of the Vaeddas. In the Sinhalese districts her star has paled before the brilliance of a later goddess, Pattini, who was introduced from the Pandiyan Madura; and offerings, in addition to those of the Vaeddas, are now made to her only in the north-central districts, and in Uva and the south-east part of the island.

Ayiyanar, the son of Mohini, divides with his half-brother Skanda, the God of Kataragama, the attentions of all Sinhalese travellers in the forests and jungles of the interior of Ceylon. So far as the Kandians are concerned, they are the deities, above all others, whose powers are specially protective in such places. In the south-western part of the North-central Province it is stated that the first-mentioned deity and the Forest-God termed Wanni Deviya1 are identical, and this latter deity is manifestly the Wanniya Bandara of the northern parts. A Kandian invocation shows that he may be Bilinda and not Ayiyanar,

In view of the close connection between the religions of the Vaeddas and the Sinhalese, it is probable that the worship of Ayiyanar exists in some form among the Village Vaeddas, at least, if not the wilder Vaeddas of the forests. There is nothing to show that one of the Bilindu Yako is Ayiyanar, excepting his relationship to the Kiri-Amma; the identification of the younger one as the Ilandaji Devata precludes his recognition as Ayiyanar, unless these two are the same deity under different titles, which the names do not support.

With regard to the other Kiri-Ammas, the local titles attached to them confirm, at any rate as regards five of them, the statement of the Vaeddas that these are their deified female chiefs. In the list of chiefs of the seventeenth century

1 An incantation of the Forest Vaeddas which Dr. Seligmann has allowed me to see throws some doubt on this identification. According to it this Forest-deity may possibly be the Gale Deviya, But the Vaeddas may have confounded the two gods (see p. 159).THE DEITIES 151

the name of one important female leader is found; and there is a distinct and unmistakable reference to one in the story of Pandttkabhaya?the * horse-faced ' YakkhinL The worship of these Kiri-Ammas proves the exalted position held by the women in former times?a sure mark of at least a certain amount of civilisation?and such deification is rendered the less unlikely by the existence among the Sinhalese of a custom permitting queens to rule over the country at various times from the first century B.C. down to the sixteenth century A.D.

It is strange that an entirely different group of seven Iviri-Ammas are worshipped by both Kandians and Low-country Sinhalese. They are described as seven manifestations of the Goddess PattinL Pattini is never treated as a Hill Goddess, but is venerated only in her aspects as the Goddess of Chastity and the Controller of Epidemics. The worship of these seven Sinhalese goddesses seems to be an independent cult which has borrowed the nomenclature of the older one, and has ousted it in some districts.

The commanding position of Pattini among the Sinhalese is doiibtless chiefly owing to her being an incarnation of the Goddess Durga, the wife of Siva, a great foe of the demons collectively called Asuras. She has so entirely supplanted the terrible Indian Goddess Kali that it is now considered that the 'Ashta Kaliyo/ the eight forms of Kali, have sunk into the position of mere attendants on her, a clear proof that she is a form of Durga.

I know of no trace of Pattini's special cult among the Village Vaeddas, and she is not ascertained to be included among the seven Kannimar, or ' Maidens/ to whom offerings are made by the Tamil-speaking Vaeddas, who, however, were unable to furnish me with their names. A list of seven Kannimar published by Mr. Nevill1?possibly a different set of deities?shows them to be chiefly evil aspects of Kali, whom these half-Tamil Vaeddas may have merely taken over from their Tamil neighbours. By the Tamils of the southern part of the Eastern Province, and perhaps elsewhere in Ceylon,

1 The Taprobantan,- Vol. ii, p. 146.J

152 ANCIENT CEYLON

as in India, Pattini is worshipped as a Goddess, under the name Kannakei Amman.1

To what extent the Vaeddas borrowed their religious notions from the Sinhalese, and the latter from the Vaeddas, must be partly a matter for conjecture. The legends regarding the Hill God, or God of the Rock/prove that he at least was an original deity of the aborigines; and his cult must have descended from them to the Kandian Sinhalese.

That the twelve or fifteen demons called Vaedi Yakas or Vaedi Kadawaras were originally primitive evil deities is at least extremely likely. It is regrettable that practically nothing is known about them. The two or three uncertain demons are perhaps included .among them.

The origin of the custom of deifying important spirits such as those of chiefs, male and female, or of special ancestors, is doubtless very ancient; it appears to be widespread in India as well as in Ceylon, the worship of the Manes being well known. The Buddhist monks of Ceylon are of opinion that the spirits of some deceased persons become Yakas. In the Jataka story No. 545, it is stated of the Kuru King Dhananjaya, * his mother in his last existence but one before this was Ms guardian deity/ In the story No. 544, Angat^ King of Mithila, inquires of Narada, " I ask thee this matter, O Narada; give me not a false answer to my question; are there really gods or ancestors?is there another world as people say ? " Narada answered: " There are indeed gods a%d ancestors,2 there is another world as people say/' In the Jataka story No. 512, it is jelated that the spirit of the chaplain of the King of Benares supplied the latter with fruit daily after he had become an ascetic.

In Southern Indian Tamil districts and in Ceylon it is believed that a person who has been inordinately fond of his house and its surroundings becomes a spirit termed in Ceylon the Gewale Yaka, 'the Yaks at Houses/ and in India, Muni. Although considered to be an evil spirit, his love of Ms old home induces Mm to act as its protector to a certain extent ;

1 Tfo Tapfobanian, Vol. iii, p. 16,

2 That is, ancestral spirits.THE DEITIES 153

and he is supposed to remain in its immediate neighbourhood. It is considered to be a lucky thing to have such a Yaka about the premises, since his care of them and the inmates brings good fortune and prosperity; but, on the other hand, if the residents neglect him and do not make offerings to him he afflicts them in various ways. This is not done through vindictiveness but because, as it was explained to me by a Kapurala, "One must live, and this is the only way in which he could make people give him food." As I understand the position, this nearly coincides with the ideas of the Vaeddas concerning the spirits of their deceased relatives and chiefs.

Such Yakas as this one and the Bandaras mentioned below are generally believed to notify their position in the spirit world by appearing to persons in dreams and saying, "I am now a Yaka "?or a Bandara, as the case may be; but some do this by performing supernatural feats, and then informing the people through an authorised person?a soothsayer, when 'possessed'?that they have caused them, and that they require offerings. One Yaka who resided at Jaffna is stated to have left that place in disgust, and come to the North-Westem Province, because he was half starved there, he said. This was the Kambili Unnaehae, mentioned below as a Forest Deity.

The earliest instance of such deification in Ceylon is that of the Solian king who invaded Ceylon in the second century A. p.

The next person is King Gaja-Bahu I, who lived in ttie second century, and who is believed to have been in reality an incarnation of a demon of Madura. Nila, a chief of his time, about whose prowess in the invasion of India some stories have been preserved, is also now a deity called Kalu Kumara or Devata, apparently one of the ' Five Gods * of the Wanniyas.

After him comes King MahaSena, of the third century A.D., who is still worshipped as a Forest Deity by both Sinhalese and Waimiyas.

In the ranks of the deities termed Bandaras, who are the spirits of important chiefs, or heroes, or ancestors, are found154 ANCIENT CEYLON

six chiefs of a King Wijaya-Bahu; but whether he was the first who bore that name, and who reigned from 1065 to 1120 A.D., or a later one, I am unable to say. It is not probable, however, that the king was one of the later rulers of that name, who were all unimportant personages, although he may have been the second one, who was at war with the Tamils in Ceylon for a short time in the thirteenth century.

Among these Bandaras there are also included Postima Baridara, who has been already mentioned as a prince who was thrown over the precipice at the Kurunaegala rock, about 600 feet high, in the thirteenth or fourteenth century; and Panikki Vaedda, a Vaedda Chief of the fifteenth century. Another of these deities who can be dated is a chief who lived under a son of Wimala Dhamma Suriya I or II, in the seventeenth century.

An addition was made to the list in the time of Kirti Sri (1747-1780), this being a chief called Kirti Bandara: and during the early part of last century the cruelties practised by the last king of the Kandian territory, Sri Vikrama Raja Si^tha (1798-1815), caused him also to be enrolled in the ranks of the evil spirits.

Lastly comes a folk-story of a man of the Western Province, who, being thought to be dead, was taken for an evil spirit or Yaka when he returned home late one night. Every one in his village refused to open a door and admit him when he knocked at each house in turn, and informed the inmates who he was. It ended in his accepting the situation, and demanding abundant food, which was deposited for him nightly. I understand that he is now enrolled among the regular Yakas, and that offerings are still made to him.

Nearly all the other deities of the Village Vaeddas are undoubtedly of South Indian origin. Whether the belief in any of them was introduced into Ceylon by the Nagas, or the knowledge of all was acquired at a later date, possibly through . the intermarriages of the royal families of Ceylon and Madura, or through their introduction during some of the various Tamil invasions, there is no evidence to prove, excepting a doubtfulTHE DEITIES 155

Sinhalese tradition of the Kapuralas that their cult originated in the time of King Panduwasa Deva, that is, in the fourth century B.C. ; and other stories which state that many of the demons landed on the western or southern coast at some unknown time.

So far as these researches have extended, the result, as regards, the'primitive cult of the aborigines, may be summarised in a few words. The original religion of the Vaeddas appears to have been this : They worshipped one beneficent deity, the Hill God of Southern India, who provided them with food, sent them rain, and checked their illnesses and epidemics. They also believed in the existence of at least twelve evil deities or demons, who caused the ills that afflicted them. They may have had a Sea God also ;. but respecting this the evidence is insufficient. Probably, also,, they shared with the other inhabitants of India a belief in the existence of ancestral spirits, to whom offerings were made, and whose functions were partly hurtful and partly protective.

One thing at least may be remarked with confidence respecting the Vaeddas?that their religious conceptions contain no beliefs that tend to show any connexion with other aborigines than those of Southern India. I am not aware that there is any adoration of the sun, or planets, or astral bodies, or the powers of Nature, nor apparently is there any snake worship by them. How far their magical ideas extend is unknown; that some of the more settled of them must be acquainted with many of the practices of the Sinhalese is proved by their faith in the existence of the Huniyan Yaka, the demon whose special function is to give effect to curses, and magic, and evil spells.

The full story of Ayiyanar's miraculous birth may be contained in the Skanda Purana, which has not appeared in English. I translate the legend as it was related to me in a Sinhalese village near Indigollaewa ; it is in close agreement with that which I heard at the great Saivite temple at Tanjore. Some explanatory additions are inserted in brackets.156 ANCIENT CEYLON

THE ORIGIN OF THE SONS OF SIVA.

Great Vishnu [Ma Vis Unnanse] having taken the appearance of a woman whose name was Surangana [' Celestial Nymph/ that is, Mohini] was rocking in a swing. At that time Basma-sura * was a servant of the God Iswara [Siva]. The Goddess Umayangana [ParvatI] was married to Iswara. While Bas-masura was employed under Umayangana she went alone to the river to bathe, and taking off her Abarana [insignia] placed them near the river. Leaving them alone there she pulled up a small quantity of Singarael [a plant] and created from it a prince, and instructed him to remain beside the Abarana [to guard them]. She then entered the water. A tale-bearer went and falsely told Iswara that Basmasura had gone to watch the Goddess bathe. Then Iswara being angry mounted his elephant, and taking his sword proceeded to the spot. [Seeing a person sitting on the river bank] he cut off the prince's head, which fell into the water.

The Goddess thereupon came out of the water, and said to Iswara, " Why did you behead the prince whom I have created ? " Iswara replied, [" I thought he was Basmasura]. If you can create another prince, do so/' Then the Goddess said, " If you will cause the prince whom I created to come to life again, I will create [not one, but] seven more.0 Iswara agreed to this. But he was unable to find the head [pluwd], and he therefore entered the water, and created a white lotus plant [okt-gaha]. He then cut off the head of Ms elephant, and fixed it on the neck of the beheaded prince, and named him Gana2 Deviya [Ganesa],

Then the Goddess created a prince from a kind of grass J she made another from Singarael; a third from a piece of cloth ; a fourth from leaves ; a fifth from sand; a sixth from creepers ; and a seventh from a kind of fruit that had fallen from a tree. Those seven remained in one place. The God

1 A Rakshasa in Indian legend. The name is a compound of Bhasma, ashes, 4- A sura, demon. Siva Is called Bha$ma-priya ' Fond of Ashes.'

2 Evidently deriving the word from the Skt. gkma, slaying.THE DEITIES 157

Iswara, saying, " I am going to eat1 my sons/' clasped his arms round them, and all seven were caught, but one escaped beneath his hand and fled. The other six were crushed together, and became the God Kanda-Swami [Skanda], with six faces and twelve hands, who rides on a peacock. The prince who had escaped became the Kadawara 2 Devata [' the Celestial who escaped/ generally considered to be a demon in Ceylon].

After this, Iswara entered the river to bathe [to purify himself], and handed to Basmasura the arrow3 which he held in his hand. Basmasura thinking " I will kill Jswara with this arrow, and marry the Goddess Umayangana/' made off with the arrow. The God Iswara, being afraid, ran away, and got hid under the swing in which the Goddess Surangana was swinging. When Basmasura came up with the arrow Surangana asked him, " What are you doing ? " "I am running to kill the God Iswara," he said. " Why ? " she asked. " In order to marry Umayangana/' he replied. "What is [the use of] that; should I be a bad match for you [mama nara-kada ?]," she asked. Basmasura answered, " It is good" [expressing his approval of the match]. Surangana then said, " We will swear an oath never to give each other up." " Yes/' rejoined Basmasura [forgetting that her name Mdhim means 'Deluder'], "what oath shall we swear?" "Take your right hand and put it on your head," answered Surangana, '* and I will take my right hand and put it on my head/' At that time Basmasura having become foolish through the sentiment of Kama [love], without giving up the arrow placed it on his head in his hand. Thereupon he was burnt up by [the magical properties inherent in] the arrow. Then [Iswara came out from under the swing and embraced Surangana,

1 This Is the correct translation of the Sinhalese words, mage putto kanta ; but the last word, like the whole story, is taken over from the Tamil, in which it must mean * embrace.* Compare Skt. kantha-graka, an embrace. In Tamil, kamt&m or kandam is a neck.

8 From Tamil kada, to step aside, or escape from, + varar, celestials.

* I have a very short Sinhalese spear or assegai, which the owner termed an * arrow * ; the arrow mentioned in the story appears to have been a stabbing weapon, as no bow is referred to.158 ANCIENT CEYLON

and] Basmasura, through his love for her was conceived in her womb. Afterwards Surangana resumed her male form [as Vishnu]. Ten months being fully accomplished, he split his right hand, and took out the prince, who received the name of Ayiyanar Bandara. Surangana's name became Kiri-Amma.

This story was told to me in order to explain exactly who Ayiyanar was, and his relationship to the Kiri-Amma. Thus it is plain that by the Sinhalese, at any rate, the latter is believed to be M5hini, and not ParvatL

A different version of this story is current in Maisur, and is given in the Indian Antiquary, Vol. ii, p. 50. In it Siva handed to Basmasura his middle fiery eye, the glance of which consumes everything on which it gazes. Vishnu in the form of a bewitching female came to Siva's assistance, and the wicked Rakshasa was burnt up himself. The interview with Mohim resulted in the production of three Lingas.

Among the Wanniyas the chief deities x are (i) The Wanniya Bandara, (£) the Five Devatas, and (3) the Hat (or Sat) Raj-juruwo.

The Wanniya Bandara is the Wanni Deviya, the Forest God who is said by the Sinhalese of the North-central Province to be Ayiyanar. Invocations collected by Dr. Seligmann indicate a possibility that he may be the Gale Yaka, the special reason against this identification being the facts that the Wanniya Bandara or Deviya of the Sinhalese is not a Hill God, and the Gale Yaka is not a special Forest God. The name of the Wanniya god means the * Forester Chief'; he is not worshipped on hills like the God of the Rock.

The five Devatas are said to be Ilandara Deyiya, Ayiyanar Deyiyi, Kalu Devatawunar, Kadawara Devatawunar, and Mangalya Deyiya. The Hat Rajjurawo is, as already stated, merely a title of King MahSsena. All these act as guardian deities.

If the names of the Devatas are given correctly, Ayiyanar

1 1 made no inquiry as to their knowledge of the God of the Rock or the Kixi-Axnm&j having myself no information regarding them at that time (1885).THE DEITIES 159

reappears among them in a minor position, where one would not expect to find him. This considerably strengthens the probability that the Wanniya Bandara is the Gale Yaka. On the other hand deities often reappear with varying names under different aspects, and various persons hold divergent views regarding them.

Ilandara Deyiya is known to the Coast Vaeddas and the North-western Kandians ; he is said by the latter to have been a chieftain under King Maha-Sena, and to have resided at Minneriya tank. The Kandians of the North-central Province state that he is a son of the Kiri-Amma. The Kalu Devata is also reported to be a special deity of the North-central Province. Kadawara Devata has been already mentioned as the prince who avoided the squeeze of Siva which compressed the six others into one.

Although it seems clear that Ilandara and the younger Bilinda are the same deity, the identity of the Wanni Deviya is less certain, notwithstanding the information given to me that he is Ayiyanar. Such statements always require sifting carefully. The name of the latter deity in both Tamil and Sinhalese is an honorific form meaning ' Elder Brother/ an expression that would not be used unless there was a younger brother, that is Ilandara, ' the youth/ or Bilinda, * the child/ Thus the latter deity and Ayiyanar appear to be the two Bilindu brothers of whom Mr. Nevill heard.

In a poetical Sinhalese invocation addressed to ' the Twelve Gods/ references made to the Wanni Deviya (also called in it Wanni Bandara) indicate that he is Bilinda, and not Ayiyanar. It states that he, the ' God of the Wanni Country/ went to Kataragama in order to receive offerings, and that napuru ayayi w%e puraia aewidin yak menne waran laeba, ' the wicked elder brother having come to this city obtained power like a YakSL.' Dr. Seligmann heard that Kande Yaka killed Bilinda; perhaps reference is made to this in the poem. From it we also learn that the Wanni Deviya was bom at a place called Kiwiyaluwa, and was king of the Vaedi country, and king of Bintaenna, who promenaded round Sorabora tank. He protects the people of jungle villages, rides wild elephants,160 ANCIENT CEYLON

and always carries a goad; and it ' comforts his heart' to see Sambar and other deer, and to visit the Uda mta, the Upper-country. It is not unlikely that some sovereign or chief of the Vaeddas has been canonised (or possibly two legends have been united), and identified as a son of MohinL

Mangalya Deyiya evidently is the God Mangala, whom Mr. Nevill mentioned1 as being anciently worshipped in the Trincomalee and Batticaloa districts. He stated that when offerings were presented to him invocations were addressed to a number of deities called by Tamils ' The 160 Vatanamar/ who were subject to his orders. He remarked that ' Mangala was specially invoked by elephant-hunters and by wild buffalo hunters/ He learnt from the hymns to him that ' both they [the chiefs] and their subjects, these votaries of Mangala, evidently came from the Malabar coast, and followed the Nayik custom of inheritance of ancestral property in the male [female] line/ Here is another practice which points to the early Nayar connection with the people of Ceylon. Mr. Nevill considered this god to be e a personification of the influence of the seven or nine planets conjointly' ; but on reading over the two hymns which are given by him it seems to me clear that the God Mankala in whose honour they are composed is shown by his name Nayinar in them, and his vahana, an elephant, to be Ayiyanar, as an avatara or incarnation of a part of Vishnu. One of the names of Ayiyanar is Nayanar, according to Winslow's Tamil dictionary. Vishnu, as the husband of Lakshml, who is termed Mankalei in Tamil, is called MankaJanar, a title given to his son also in these hymns.

The second hymn begins :?" I sing the sacred story of the glorious Mankalars [Vishnu and Ayiyanar]. O Sankara [Siva] of vast Kayilasa ! graciously aid me to sing the praise of Narlyana [Vishnu] on the ever-writhing snake-couch/*

The first hymn relates the capture of the elephant which "became Ayiyaaafs vahana, and part of it is so interesting as giving a description of the god that I quote it:?

* Hankalai&r came and rose forth from a beauteous lotus flower, at the white-lotus flowering, b!iss~gi\ring, great city 1 The Taprofatnian, Vol. i, p. 54.THE DEITIES 161

called Kasi, called Miilam. At that very time was the Ava-taram of Mankalanar. When Nayinar was born at the auspicious hour, all came and worshipped the Mankalar.

' When twelve years fulfilled had passed by, wearing thrice three jewels [the nine gems], and assuming the triple-twisted, cord [the sacred thread] Nayinar was seated, with bow, arrow,, javelin, strong cord, axe, naga-like-cane like a goodly circlet, with girdle, indescribably-flowered clothing, girt with a curved, club, and wearing a gem-set ring and ear-rings, with goodly coat and hat; wearing all these Mankalanar was seated !'

The capture of the elephant by noosing it is then described,, and the hymn ends, ' they bathed it for beauty in the white-lotus-flowered pond. Placing on its feet bangles, on its neck bells, on its body spreading white cloth and so forth, they brought it. Wearing a coat, wearing a cloth, wearing a hat putting on a crown, like the Katpakam grove surrounded by the Ganges, surrounded on all sides, he was pleased to become seated/

In this story Kasi, that is, Benares, is mentioned as the birthplace of Nayanar, and the account of his origin is quite different from those previously given.

I made no inquiry into the demonology of the Wanniyas ; but it may be accepted as certain that, like the Sinhalese and Vaeddas, they are acquainted with along array of evil spirits.

1CRELIGIOUS CEREMONIES

THE commonest religions ceremony of the^ Village Vaeddas is performed on the occasion of their setting out on one of their hunting excursions, of which it, or an allied ceremony, is the invariable preliminary. This is a prayer for protection and success in hunting, offered to a deity, usually the Indi-gollaewa Kiri-Amma, whose powers are supposed to be specially manifested in the forests. It is always accompanied by offerings of food made according to a simple fixed ritual. For this purpose a trained intermediary or priest is not needed ; any member of the hunting party who knows the form of prayer which is necessary temporarily undertakes the office.

The Vaeddas of the interior villages prepare a covered shrine resting on four sticks, under a large shady tree. The bottom of the shrine is made at a height of about four feet from the ground ; it is nearly two feet square, with a roof arched over from side to side and the*front open, but not the back. It is covered with grass on the top and has the walls enclosed by leafy twigs. The inside is lined with white calico, if available. Similar frames or temporary shrines are erected by the Sinhalese for holding offerings to evil or beneficent deities.

Inside this are laid on separate green leaves the best foods they possess?fresh meat of some kind, Rice boiled with Coconut-milk if obtainable, small cakes of Rice-flour or Millet, some bits of Coconut if they have them, and a little water in a piece of Coconut shell. A few red flowers of the Ratmal tree (Ixora cocdnea), or Eramudu tree (Erythrina indica) are also placed inside the shrine before the offering; and in front of all a wick made of a bit of rag saturated with fat is fixed in anything available?often a piece of Coconut husk. Lustration of water is made in front of the shrine, and obeisance is performed to the offerings with the hands raised in front of

162RELIGIOUS CEREMONIES 163

the face and the palms touching each other. This is the dedication of the offerings.

The wick is now lighted, and the officiator turns aside until it expires. He then steps in front of the shrine, and dances there slowly, holding an arrow in his hand, after which he says loudly " Ayb5 ! Aybo ! Aybo ! Indigollae Kiri-Amme ! Hail! Hail! Hail! (literally, * may life be long '), O Indigollaewa Kiri-Amma ! Eat. Drink. Give us livelihood. Give us meat got by hunting. Do not cause us to meet with the Elephant. Do not cause us to meet with the Bear. Do not cause us to meet with the Leopard. You must make us a livelihood by (means of) the Pangolin. You must make us a livelihood by (means of) the ' Iguana/ You must make us a livelihood by (means of) the Monkey. We must meet with the Sambar deer. We must meet with the Pig. To the end while going, to the end while corning back, you must promote and give livelihood and protection, O our esteemed Goddess."

When the offerings are made to the Gange Bandara, the River Deity, the hunters first fix a day for them, and give notice to a professional devil-priest, termed a Kapuwa, who purifies himself by bathing on the three days preceding the ceremony. The offerings consist only of Betel-leaves, Areka-nut cut in pieces, and a little water. In fact, the deity is merely offered the usual ' chew' of Betel?the eastern form of the Stirrup-cup?before the hunters set out. The shrine is erected as in the preceding case, but lined with a torn cloth. TMs ceremony may take place anywhere; there is no fixed site for it. After presenting the offering, and lighting one wick, and dancing, the Kapuwa repeats the same prayer as before, simply changing the address to the Gange Ban4&ra instead of the Goddess.

When going only to collect honey they usually say aloud for the information of the deities, " We are going to cut a hive for the Yakas/* In their own interests the deities are then expected to see that the men are successful On their return, some present part of the honey they have obtained, and cooked rice if they can provide it, to the Kataragama God. Many offer the honey to the Kiri-Amma; but if they have found164

ANCIENT CEYLON

im i

II!

M1

very little they punish her by withholding this offering, apparently without fearing any act of reprisal on her part.

Village Vaeddas who know more Sinhalese, as well as the neighbouring Sinhalese villagers, place the offerings in a similar shrine, lined with a white cloth, and use a longer prayer to the Kiri-Amma. It is made?so I was informed, but this maybbe a mistake?before the lights expire, two being set in the shrine, one for the Goddess and the other for her husband, the Gale Yaka. The prayer was repeated for me as follows, the utterer holding an arrow in his hand while saying it :? " Aybo! Aybd! Aybo! 0 Indigollaewa Kiri-Amma, who became famous through splitting the Sapphire Gem at the Sapphire Mountain1 in the country o'f the Seven Seas, and even the country beyond it! While you are looking at this beautiful cooked food this is our supplication, telling you to give a good Sambar deer, having caught it with this Vaedi arrow. O my Kiri-Amma ! This is our supplication asking you for a good Horn-bearer2 [Sambar deer], for a Speckled One [Axis deer], for a Fat-maker [Pig], for a Meat-bearer [Buffalo] to be daily placed for us."

When the same people pray for honey an identical formula is used up to the word' food/ after which it runs :?" This is our supplication, telling you, O Indigollaewa Kiri-Amma to grant your loving favour, giving us a bee-hive until our eye is blind." The concluding part is a common saying among village Sinhalese ; its meaning here is 'even more than we can want/

In order to avoid dangerous animals, and the difficulties of the path, they say after ' food/ " This is our supplication telling you, 0 Indigollaewa Kiri-Amma! to grant your loving favour, beating and driving afar Leopards, Elephants, enclosing and filling up hollows, blunting pointed stones,

1 Alaka, on Mt. Kailasa, the home of the Yaksha sovereign Kuvera. In the Meghadnta (Onvry), v. 76, the exiled YaJksha says, "On the

battles of this (lake) is a beautiful mountain for sport, whose summit, composed of Sapphires, is worthy to be seen on account of its being enclosed by golden plantains."

? 2 These names of animals belong to the Kaele"~basa, and not the Vaedi dialectRELIGIOUS CEREMONIES 165

beating and driving afar Serpents, giving livelihood and protection in the forest, and [making up] all deficiency/'

Among the southern Village Vaeddas, and in the adjoining Sinhalese districts, offerings of the first-fruits obtained by hunting are made in a similar way, with one light, to the Kukulapola Kiri-Amma. They consist of fresh meat and honey,

To the seven Kiri-Ammas of the south, a single offering is made in the same kind of shrine when the men are about to leave on a hunting expedition, and also when children are sick or fretful. If they are procurable, it consists of milk-rice (rice boiled in Coconut milk), Jak-fruit, the flower-bud of the Plantain tree (which is used in curries), Betel-leaf and sliced Areka-nut, Sugar-cane, and a little Sandal-wood.

In this case, the shrine is subdivided into seven compartments in which seven leaves are placed on a white cloth, one for each Goddess ; and on each of them a small portion of each kind of offering is laid. Water is sprinkled over these articles, and in front of the shrine, and the offering is also purified by incense (a resinous gum which exudes from the bark of the Dum tree), which is burnt on a fire-stick, and waved round it. A wick is then placed near each end of the offering and lit. After the lights have expired, the offerer takes a Betel-leaf in his right hand, between the first two fingers, and waves it from side to side in front of the shrine, and then, still holding it, makes a long prayer to the seven Goddesses, which I had no opportunity of writing down.

When children are ill, and the parents do not possess things suitable for giving to these seven deities, or the time is inauspicious, or there is not an opportunity of doing it (as in the case of a sudden^violent attack), they make a vow to present an offering to them ; and hang up a bare, a visible token of the forthcoming sacrifice. There is no magic, as some have supposed, in this act; a bare, which has various forms according to the personage to whom the offering is to be made, is like an engagement ring in Europe, and is invariably necessary among all Sinhalese when a vow to present an offering has been made. It must not be removed on any account166 ANCIENT CEYLON

until the vow has been accomplished. In the present instance it consists of a Mango-leaf and a strip of Palmira-palm leaf, strung on a thread, which is then tied across the doorway of the hut. On one side of the Palmira leaf are written the words * Patta-Giri, Bala-Gin, Molan-Giri/ probably to indicate that the child is to be specially guarded against the evil actions of the female demons who bear those names.

Among the Kandian Sinhalese, Girl is the feminine form or Sakti of a class of demons, twelve or more in number, called Gara (plural Gaerae or Garayo), who afflict only women and children. The word gam means sickness or disease, and is derived from the Sanskrit root gmh, to seize ; these demons are tjius personifications of certain diseases.

In this case, the offering, as described above, is usually made after the child has recovered, or as soon as the requisite articles for it can be procured; but sometimes, as when an infant has been fretful in the night, it is presented on the following day, if possible.

Respecting the ceremonies used in presenting offerings to the Bilindu Yakd, Mr. Nevill merely remarked,' The offerings are those, omitting rice, still used in India and Ceylon at the festival of Pongal, in honour of the January sun. The symbol used is the arrow/ I

Mr. Nevill observed concerning the religious ceremonies of the Vaeddas: ' In almost all their religious rites the arrow is used; it receives worship as an emblem, or is waved in the hands of the celebrant, around the sacrifice,

' They leave tiny babes upon the sand for hours together, with no other guard than an arrow, stuck in the ground by their side. Their belief in the efficiency of this has received no shock. They never knew such a child to be attacked by wild beasts, pigs, leopards, jackals, etc., or harmed. They say, " Are we not the children of our Gods, and if we leave our child under their care will they not watch over it ? " The arrow being the God's symbol, they themselves are practically, as Ms children, the iya-vans, or " sons of the arrow/' and this fully accounts for the name Yaka or Yakkho/ [that is, as 1 The Taprobanian, Vol. i, p. 195,RELIGIOUS CEREMONIES 167

I have already stated, iya-ko, arrow-persons. I have previously expressed my opinion of this derivation; I may add that the Vaeddas never claim to be called either Yakas or lyakas].

Offerings of food are made by the Tamil-speaking Vaeddas to their deceased relatives, excluding infants. As the manner in which they make the gifts was certainly not derived from their Tamil neighbours, this ceremony may have been developed by their ancestors in ancient times, although its absence elsewhere shows that it is not a primitive one. The offering is first made seven days after the death, and subsequently once a year, after the Hindu custom.

A shrine like that already described is erected under a shady tree in the jungle. It is from three to five feet high and is usually, but not always, arched over; it is lined with a white cloth. The foods placed in it |ire cooked meat if available, cooked rice, Betel-leaves, Areka-nuts, and Plantains. Of course they are purified by lustration. In front of the offering one wick is lighted inside the shrine. Tom-toms are beaten loudly to attract the notice of the deceased person, and a dance is performed in front of the offering in his honour, after which the officiating relative merely says, " Lord, eat and go." The party then return home.

The ceremony of the offering to the Seven Kannimar is said to be quite similar, the shrine having, as in the case of the seven Kiri-Ammas of the South, seven compartments, one for each Goddess, whose share of the food is thus given separately. I do not know the prayers addressed to them. An arrow is certain to be held by the dancer.

Regarding the manner in which offerings and. invocations are made for propitiating the various evil demons of the Vaeddas, I regret that I have no information, my visits to their district having been too short to permit me to collect the particulars. Those who live in the forests informed me that they are accustomed to place offerings of food for them in shallow hollows in the surface of rocks. This matter has now been investigated by Dr. Seligmann.

When the Wanniyas are about to set out on a hunting trip they first purify themselves on the preceding day by bathing168 ANCIENT CEYLON

in their little village tank, and then perform the following ceremony to ensure success in their expedition, no women being allowed to see it.

Under a large tree at the foot of the embankment of their tank, one of the party, who becomes the temporary officiator or priest, cleans, by pounding, four quarts of paddy (rice iu its husk), and boils the rice so obtained. Others fix in the ground three sticks in a triangle, with a platform in it well above the ground level. The boiled rice is placed in a new earthenware pot, or '* chatty/ which is deposited by the celebrant (who alone performs the whole ceremony), on the frame, resting on the tops of the sticks ; and a little saffron is sprinkled on the rice.

On the little platform below the pot seven Betel leaves are next arranged in a circle, with their points together, and an Areka-nut is put on each; a pinch of Camphor is also placed near the outer end of each leaf, a light is applied to it, and it is burnt.

Water is then taken in a washed gourd, or new * chatty' (or pot) and with the hand a little is sprinkled three times on the ground in front of the offerings. Dummala incense is next laid on a fire-stick, and while it is burning the stick is waved round the platform and the pot of rice. The officiator now steps back, and with his palms joined in front of his face pays obeisance to the offering by bowing to it three times. This completes the dedication.

He next takes three wicks soaked in fat, one of which he fixes on the end of an upright stick set in the ground in front of the frame, so as to be higher than the pot of rice, and the other two he arranges on the ground on each side of the taller one. The three are then lighted. Before they expire he walks aside, and turns his face away until they have completely ceased burning. After this he returns, and again sprinkles water three times in front of the frame.

He now raises his joined hands, and standing in front of the offering three times repeats the following invocation :? " You, O Wanniya Ban Jara [? Ayiyanar] are required to take the offering of a feast of cooked food. Wanniya Ba$ jlara, weRELIGIOUS CEREMONIES: 169

must meet with the Royal Great Hive; we must meet with Horns; we must not meet with an Elephant, a Bear; we must not meet with a Snake ; we must meet with livelihood."

The celebrant then removes the offering, of which all the hunting party partake.

On the return from their expedition the same offering is repeated in the same manner; but the officiator merely says, *' Wanniya Bandara, we met with a livelihood."

Hunting parties of the Kandian Sinhalese of the North-central Province perform a ceremony which is very similar to that of the Wanniyas and Vaeddas, when about to leave their village on one of their expeditions in the forest. Under a large shady tree they prepare a maessa, or small covered shrine, which is raised about three feet off the ground, and is open only in front; it is supported on four sticks set in the ground.

In this they offer the following articles if available, or as many as possible of them :?One hundred Betel leaves, one hundred Areka-nuts, Limes, Oranges, Pine-apples, Sugarcane, a head of Plantains, a Coconut, two quarts of rice boiled specially at the site of the offering, and silver and gold. Also the flowers of the Areka-nut tree, the Coconut, and Ratmal tree. All are purified by lustration and incense, as usual, and dedicated.

They then light a small lamp at the front of the offering* and remain there watching it until it expires, differing in tkis respect from the practice of the Wanniyas, who must never see the light go out.

Before the light expires they perform obeisance towards the offering, and utter aloud the following prayer for the favour and protection of the Forest Deities, which must also be repeated every morning during the expedition, after their millet cake, gini-piiwa, has been eaten, before starting for the day's hunting :?

** This is for the favour of the God Ayiyanar; for the favour of the Kiri-Amma, for the favour of theKataragama God [Skanda]; for the favour of the Kalu Devata; for the favour of Kambill170 ANCIENT CEYLON

Unnaehae *; for the favour of Ilandari Devata Unnaehae ; for the favour of Kadawara Devata Unnaehae ; for the favour of Gale Barbara ; for the favour of the Hat Rajjuruwo.

" We are going to your jungle (uyana); we do not want to meet with even a single kind of [dangerous] wild animals. We do not want to meet with the Tall One [Elephant], the Jungle Watcher [Bear], the Animal with the Head causing Fear [the Snake], the Leopard. You must blunt the thorns. We must meet with the Horn-bearer [Sambar deer], the Deer [Axis], the One full of Oil [Pig], the Noosed One [Iguana], the Store-house [Beehive]. We must meet with about three pingo [carrying-stick] loads of honey. By the favour of the Gods. We ask only for the sake of our bodily livelihood/'

The first, fourth, sixth, seventh and ninth personages are included among the deities of the Wanniyas previously enumerated. The Kiri-Amma is the Goddess of the Vaeddas. I have already stated that the Gale Bandara is not the God of the Rock, but a deity who landed on the south coast of Ceylon, with others from Madura. Kambili Unnaehae is an evil deity who is well known in the North-western Province. Thus these deities of the Forest are a miscellaneous gathering of Gods and evil deities.

In the hunting-prayer of the Wanniyas the first thing asked for is the Royal Beehive, the wish to find which occupies a foremost place in the hearts of all these northern hunters, though the Vaeddas have no knowledge of it. From the WanniyaKona, whose grandfather was said to have taken one, I obtained the following account of it.

The entrance to the nest of the Bee-King is always at the foot of a large hollow tree, up the inside of which it extends. It is surrounded by seven other nests, which are those of his seven Adikarams, or Ministers ; and they also have their entrances at the foot of their respective trees. The Royal Hive is the largest; it extends up the hollow tree higher than a man's height; but all honey above the level of the chin must be left for the Bee-King. From each Adikaram-miya or Minister's hive there will be got two gourd-f ills of honey, and 1 Unnaehae is equivalent to our * Mr.*RELIGIOUS CEREMONIES 171

from the Raja-miya or Royal hive fourteen gourd-fuls and seven large chatties or pot-fuls.

Whoever may first discover the nest, no one but a Wanniya can cut out the honey, and that only after a solemn ceremony, otherwise the bees of all the hives would attack and kill him. This result is said to have nearly occurred on one occasion when some Sinhalese villagers were rash enough to attempt to take one themselves ; the boy who found it was badly stung, and would have died had not the Wanniyas been summoned in time to save his life by their prayers and magic spells, which appeased the bees, and enabled the hunters to get the honey.

The taking of the Raja-miya being an event of such extreme importance, a special offering is necessary as a preliminary. For this there are required one hundred ripe Plantains, one hundred Limes, one hundred Oranges, one thousand Areka-nuts, one thousand Betel leaves, seven quarts of unhusked rice (paddy), and seven Coconuts.

The first Wanniya who sees the Raja-miya must make the offering, and conduct all the proceedings on the occasion ; and for the time being he is called the Wanniya Kapurala, or demon-priest. For the seven days prior to cutting out the honey he must bathe, after anointing his head with lime-juice. He must continue to wear the same cloth all the time, and on each day he must wash it, whether it be an old or a new one. These must appear to be very unusual purifications to persons who rarely perform such acts in ordinary life, and they evidently indicate the extraordinary character of the occasion.

On the day when the honey is to be taken, the party proceed to the site of the Royal Hive; and there, within the circle of the Ministers, the offering is presented to the Forest-Gods, to ensure the success of the undertaking.

The Kapurala first pounds the paddy, and having arranged seven new cooking-pots in a row in front of the Raja-miya, he boils the rice in Coconut milk separately in each. He next spreads a new white cloth on the ground in front of the Raja-miya, and places on it in a row opposite the pots sevea large leaf-plates, which must consist of either pieces of Plantain172 ANCIENT CEYLON

leaf or leaves of the Halmilla tree. After these have been lustrated the rice is deposited on them, that from each pot being put on the leaf adjoining it. The Betel leaves, Areka-nuts and fruit are then laid beside the rice, and sprinkled with water; and Camphor is burnt on one spot in front of all. Incense burning on a stick taken out of the fire is now waved round the cloth, and a triple obeisance completes the dedication.

A wick soaked in fat is next fixed opposite each offering, and lit; and, as in the ordinary ceremony, the Kapurala turns aside until these lights have expired. He then returns, and makes a lustration three times round the cloth.

He now stands facing the offering, with palms joined in front of his face, bows thrice to it, and says,'' Wanniya Bandara, Five Devatas, Hat Rajjuruwo ! This is for the favour of the Gods. To cut the Raja-miya came we." The others thenrespond in the Kaele-basa, or jungle dialect, ( Kapurala, Gabadaewa wadulapan," (cut the hive,) whereupon he proceeds to cut out the honey, and place it in the gourds and pots which have been brought for the purpose. Lastly, he removes the rice, and shares it with the party, who all eat it up at the spot; after this they carry home the honey.

When northern Kandian Sinhalese meet with a hive on their honey-collecting expeditions, the person who cuts it out is addressed as "Waduwa," Axe-man, in Kaele-basa. Under no circumstances, however, must he take out and divide the honey among the party. This can only be done by a second person, who is addressed as " Purawanna," '"He who fills * the receptacles brought for it.

In the North-central Province and the adjoining part of the North-western Province, Ayiyanar is not merely a forest deity; he also exercises a general supervision over the village tanks.

In the former district Mr. R. W. levers stated 1 that when

a village tank has filled, the elders of the village perform a

ceremony called Mutti Mangalya or 'Cooking-pot Festival.8

They proceed to a special tree at the tank, and a salute of

1 Manual of the North-central Province, p. 109. Note A.RELIGIOUS CEREMONIES 173

two guns is offered there to the deity. The chief elder, a ' Gamarala/ then steps forward, and announces to the God that the tank is becoming full, that cultivation will now be commenced, and that after the harvest the festival will be celebrated. A bare is then deposited in the shape of a few copper coins wrapped up in a piece of rag coloured yellow with saffron, which is tied to a branch of the tree. The ceremony is ended by the Gamarala's commending the tank, the village, its residents and their cattle to the protection of the deity.

After the harvest is finished the villagers at a public meeting appoint a day for the fulfilment of the promised ceremony, called the Mutti Mangalya. The nearest Anumaetirala (the title of a dancer in honour of a god and not of a demon), is invited to conduct it, and notices are issued to the washermen whose duty it is to supply the necessary white cloth, and to the tom-tom beaters who must take part in the ceremony. The Gamarala directs that every shareholder in the rice-field should contribute to the feast.

On the appointed day these contributions are collected; they consist of rice and other materials for curries, Coconut oil, cakes, sweet Plantains, and Betel and Areka-nuts. The food is then cooked, and at the evening the assembled people eat it.

The meal being over, the Anumaetirala, accompanied by all the people and the tom-tom beaters, proceeds in procession with two new earthen pots to the tree on which the bare was hung. Under it on a raised altar of sticks (y^hana) overhung with cloth and erected earlier in the evening are placed the two pots (mutti) after being purified with water and incense as usual, and marked with saffron ; Betel and Areka-nuts are also deposited on it [and no doubt a light also]* The deity is then addressed by the Anumaetirala [presumably thanking him for his favours and requesting him to accept the offering], and ceremonial dancing by him to the strains of the tom-toms continues till dawn. The two pots are then removed from the altar and laid on the stumps of two branches on or under the tree.174

ANCIENT CEYLON

By the mouth of the Anumaetirala the god now makes it known that he has accepted tlie offering, and that the tank, the village, its inhabitants and their cattle are taken under his protection for a period of one, two, or three years, as the case may be.

The people then return to the village, where the Anumaetirala again dances, and the tom-toms are beaten until the mid-day meal is cooked. This is eaten up at noon, after which all disperse. Mr. levers stated that a somewhat similar ceremony is performed in case of an epidemic among either men or beasts. Thus in that part of the country it is clear that the place of the God of the Rock, the Gale Deviya, whose worship and functions are described in the next chapter, is at least partly occupied by Ayiyanar. He is said to have fifty names, each one indicating a different function or power possessed by him.

The villagers catch the fish when their tanks are nearly emptied, by wading out in the water and suddenly dropping in it a conical wicker basket or creel (karak-gediya) without bottom, and with a small opening at the top through which the arm can be inserted for removing any fish that have been imprisoned. In the case of tanks infested by crocodiles in the eastern part of the North-western Province these fishers appeal to Ayiyanar to protect them while so engaged. They break some leafy twigs, and hang them on a horizontal branch or creeper, and say, " It is for the favour of the God Ayiyanar. Do not permit any living creature whatever in the tank to bite us." After this, the crocodiles are said never to molest them even when close to them (I have seen them only a few feet distant from the men), provided they do not defile the tank in any way? including expectoration in the water.

When they are travelling through wild forest which is believed to be infested by wild animals, or possibly robbers, or by evil demons, both Kandian Sinhalese and Wanpiyas are accustomed to make a very simple offering to one of the Forest Deities, who is usually Ayiyanar, accompanied by the prayer, " It is for the favour of the Gods/' or " the God Ayi-yanSr." The offering merely consists in hanging a leafy twigRELIGIOUS CEREMONIES 175

across a horizontal creeper or branch at the side of the path, and usually under a large shady tree. In some places where no suitable creeper or branch is available two sticks with forked tops are set firmly on the ground, and a horizontal one laid across them, on which the offerings are hung. These sometimes accumulate through the action of successive travellers until they form a large heap of such twigs. In the last chapter I mentioned that the Wanniyas made this offering to the Hat Rajjuruwo when supplicating him to stop the approaching rain-storm. I do not remember noticing these twig offerings in the districts of the Vaeddas.V THE PRIMITIVE DEITY OF CEYLON

IN the account of the religion of the Vaeddas it has been mentioned that their chief deity, the Gale Yaka, is probably identical with the Hill God of the aborigines of Southern India. It may be assumed that the knowledge of him was either brought to Ceylon by the first comers, or was acquired by them at an extremely early date, as nothing is known of him by the Sinhalese of the coasts, or the northern Kandian Sinhalese, or by the Tamils of Northern Ceylon, If his worship had been introduced at a later date, after these races had arrived in the country and had occupied all the coast-line, some, at least, of them would be acquainted with it. This god is regularly propitiated in the interior, however, by the Kandian Sinhalese in part of Uva, and the southern half of the Eastern Province, and especially in the tract to the north-east of Kurunaegala. The residents in these districts may have acquired a knowledge of him from their ancestors the Vaeddas.

Although he is known by the name of Gale Deviya, * the God of the Rock/ In these latter districts the Sinhalese consider him to be a powerful demon, and state that they apply the expression * God J to him merely as an honorific title calculated to please Mm. It will be seen, however, that It has a much more honourable meaning. It Is an excellent illustration of the degradation of an ancient deity into the position of a demon. I have already given the names by which he Is known to the Vaeddas, and pointed out that was a deity at Anurldhapura?* the Vaedda God '?? In the fourth century B.C., who appears to have been this one.

A regarding his arrival la Ceylon from India

Is current the Vaeddas the ; and all

177178 ANCIENT CEYLON

who are acquainted with him agree that he came from a country called Malwara-desa or Malawara-desa, ' The Country of the Hill Region/ which can be no other than some part of the Malayalam tract, our Malabar, for which, however, a separate expression, Malayala-desa, is now commonly used in Ceylon.

The tradition of the Tamil-speaking Vaeddas, which is very definite, is that accompanied by his minister Kurumbuda, * in the olden time ' he landed from a vessel on the east coast, at a place called Periya-kaduvei-karei, close to Valeichena, which is twenty miles north of Batticaloa. A temple was established there for his worship, and it was in existence down to comparatively recent times, when the residents of the place having died out or left, it was abandoned, and the

__ !?

FIG. 40. Kokka-gala.

site became overgrown with jungle. They state that he did not remain at this spot, but went to some place in the interior with which they are unacquainted. " When he came/' they said to me, " he told us the names of things, trees, and animals, and how we should make offerings and dance to him when going into the jungle to hunt, and at other times. He told us everything we know." Such teaching is distinctly a characteristic of only a primitive deity.

The Vaeddas of the interior state that the Gale Yaka came over the sea, and alighted on two hills of their district in succession, on which dances are still performed in his honour; one of them is called Kokka-gala, and I believe the other is Omun-gala. Omun may be the Sanskrit word oman, * favour'; the name would then appear to mean the rock on which the God granted favours?probably a translationTHE PRIMITIVE DEITY 179

of an ancient Vaedi name for it. Somer thought that this deity afterwards proceeded to Kataragama; I have already referred to the probability that this place was a site where a Vaedda deity was worshipped in early times.

The Sinhalese who inhabit villages in the-same district carry his movements a step onward, and repeat a tradition that he and Kurumbuda went to some place further inland, where they killed a number of Buddhist priests, and took possession of a cave in which they resided. Neither the name of the district to which he proceeded, nor the site of the cave is known by them.

To follow up the God's travels it is now necessary to move to the Kurunaegala district of the North-western Province, where the latter part of the story is much more definite, and is related as follows. The Gale Deviya, attended by Kurumbuda Devata, alighted from Malwara-desa on Ritigala, the hill called Arishtha in the Ramayana, and thence came to Maenikpaya-kanda, the upper part of Rana-giriya, called also Deva-giriya, * the Hill of the God/ a steep forest-clad rocky hill near Nirammulla, about fifteen miles north-east of Kurunaegala. Some say that they landed first at Wilbawa, two miles from Kurunaegala, before going to Ranagiriya; others believe that they came direct to Ranagiriya from India,

They were pleased with the general convenience' of a large Buddhist cave-wihara or temple which they found established under a rock on the slope of the hill, and wished to take possession of it; but the sixty monks who occupied *jt refused to hand it over to them, and began to chant * Pirit/ or sacred stanzas, for protection against evil in general and demons in particular, as a spell to keep them out. If they could persevere in this course, and continue the chanting without intermission for seven days and seven nights, demons would have no power over them. In the meantime the Gale Deviya could not harm them while the magical verses were being repeated. So he said to his minister Kuranibucja, " Kill these monks for me/' But the monks went steadily on with the Pirit, and Kurambucja could not touch them unless he could make them stop. It must have been an interestingi8o ANCIENT CEYLON

spectacle. Six days passed, and the demons had made no progress whatever. At last, on the seventh day, the resourceful Kurumbuda threw down into the midst of the holy men the quarter of a bull, at which all the monks started, and raised their hands higher than their shoulders in astonishment, and said with disgust, " Ish " ! It was a little word, or hardly a word, but it was fatal to them. The Pint was stopped for an instant, and in that instant Kurumbuda plucked off their heads, and drank their blodd.

The Gale Deviya then took possession of the cave and the hill, which has ever since been his headquarters in Ceylon, his ' Mula-gala.1 He wanted to live at this place because it was in the great Pallekaele Forest, nine gawus long and nine gawus broad,1 without a village in it. Here he could live undisturbed by the busy world around. Over this forest he placed his minister Kurumbuda Devata, in charge as Mura-karaya, or Guardian, with his residence at Kurumban Kanda, a hill in the northern part of it.

By some, the Gale Deviya is spoken of as the Demala Yaka, the Tamil demon, all South-Indians being collectively called ' Tamils' by the Sinhalese villagers.

As at most of the detached metamorphic hills of Ceylon, there are several large natural caves, due to weathering and flaking of the rock, on the sides of Ranagiriya, which retain evidence of their former occupation as residences of Buddhist monks, or temples ; but all are now abandoned to the forest, and to the bears and leopards which sometimes take shelter in them. It is not definitely known which of them was the scene of the legendary contest, but it is supposed to be one of the higher caves. There is a small ruined dagaba, or solid dome-shaped relic-mound, built of brick, near some lower caves, and over a cave close by it the following inscription has been cut in the earliest form of letters, with the bent r, which shows that it is probably of not later date than.the second century B.C. :?

Gamika sita sala Paramaka Tisasa ca.

The cool hall (of) Gamika and of the Chief TIssa.

1 Thirty-six miles square.THE PRIMITIVE DEITY 181

Gamika being a feminine form may be presumed to be the name of the pious headman's wife, who evidently joined her husband in causing the place to be prepared for the reception of the monks. The inscription proves that the legend of the expulsion of the monks cannot have originated before the second century B.C.

The bricks used in the enclosing wall of a room formed at another cave, termed the Uda wihara, the upper temple, are 1275 inches long, 8-12 inches wide, and 2-75 inches thick, the contents being 285 cubic inches, and the product of the breadth multiplied by the thickness, 22*3 square inches. According to the table given in the next article, these dimensions point to about the third or fourth century A.B., as the approximate time when they were moulded. Thus the monks were still on the hill at that period. The good state of preservation of the plastering on the wall may be taken to indicate a tenancy of the cave extending to perhaps the thirteenth or fourteenth century, or even later. Therefore it would appear that the monks were never driven away from the hill up to comparatively recent times.

In view of this, it is strange to find that the villagers living on the eastern side of the Central mountains, who are totally unacquainted with this district, have preserved the same tradition of the contest for one of the caves. It is strong evidence of the antiquity of the story ; and the presence, from early times, of a temple to this God is also indicated by the names of the hill itself. A possible meaning of Rana-giriya is ' the hill of the battle/

By way of explaining the legend, it may be surmised that while in early times an upper cave was utilised as a dewala, or demon temple (literally, a god's-residence) for the Gale Deviya, the Buddhist monks occupied the lower ones, and wished to get the dewala removed. If the death of some of the monks occurred through an epidemic disease, or in any mysterious manner, it would certainly be attributed to the malicious action of this so-called demon ; and in this manner the outlines of the story may be accounted for, so far as this hill is concerned. The dewlla of the Hill Godi8a ANCIENT CEYLON

has long ago been transferred to a more convenient site at the village of Nirammulla, nearly two miles from the hill.

The legend evidently contains a reminiscence of a conflict between the two religions, Buddhism and the worship of the Hill God, in which the latter was victorious. As a matter of fact, in spite of the overwhelming position of Buddhism, the belief in the power of the Gale Deviya has survived down to the present day in considerable vigour throughout a large tract of country surrounding his headquarters at Nirammulla, even while all the inhabitants also adhere to Buddhism. It is doubtless due to this faith in Buddhism that the God has been relegated by the Kandians to the ranks of the demons of the island, although he must originally have been a deity friendly to them.

I

FIG. 41. Vaedda Temple of the Gale Yaka.

Evidence of this is to be seen in the fact that notwithstanding their present opinion of his character as a demon, the Kandian Sinhalese of the district where his cult prevails still attribute beneficent actions to him. When unfavourable seasons ruin or seriously damage their crops, it is to him that a group of villages will unite to make offerings, and appeal for suitable rains or better times. In wide-spread outbreaks of malarial fever, or in serious epidemics affecting man or beast, the people of the whole country-side equally turn to him collectively for alleviation of their misfortunes. His commanding position is shown by their very rarely or never asking him to exert his powers in the case of minor evils, or those affecting single families.THE PRIMITIVE DEITY

183

In spite of the Buddhist story of his killing the monks, it is undeniable that the functions generally credited to him by the Kandians are those of a superior beneficent God, and not those of a maleficent evil spirit. This is nearly the position that he occupies among the Vaeddas, who, however, are on more intimate terms with him, and in some parts even expect him to attend to their little hunting requirements, and undertake the provision of game for them, like the Kiri-Amrna. In sickness, too, he is the benevolent deity to whom each Vaedda family turns for assistance and medical aid, and who protects their districts from epidemics and misfortune.

FIG. 42. Rock Temple of the Gale Deviyau

The general character of the edifices constructed in honour of such a powerful and kindly deity certainly leaves much to be desired. Among the Tamil-speaking and other village Vaeddas, the building erected as a temple for him is an extremely simple and economical oblong structure, a mere hut, consisting of only one room, with an entrance at the middle of one end. It is roofed with grass, and has the spaces between the sticks of the walls closed by leafy twigs, like their own houses, which, in fact, it closely resembles. The service to the God is sometimes conducted in front of the entrance.

In the north-western Sinhalese ? districts the Hill God's temples, termed dewalas, like those devoted to all minor deities or demons, differ in no respect from the latter structures* As a general rule, they are dedicated to several of these godlings184 ANCIENT CEYLON

or demons, as well as to the service of the God, who thus finds himself in a somewhat mixed company.

They all consist of two rooms, one being a small rectangular chamber with ' wattle-and-daub' walls, plastered over with mud, in which the Abarana, or symbols of the deities, are stored, as well as any lamps required for the services, a copper or bell-metal vessel for containing sandal-wood, and at least one earthenware cup used for holding lustration water (Figs. 51-53). Attached to this is the Dig-ge, ' the long-house/ an oblong shed extending longitudinally in front of the sanctum, in which part of the services are held, and tom-toms and pipes are played. At one side, a small structure called the Mulutaen-ge, the 'kitchen-house/ is built for use as a kitchen when food is cooked for the gods and demons. In some places the dewala is a small cave-shelter under an overhanging rock, with the front enclosed by a wall of brick or dried clay. The reader is referred to the illustrations of both kinds of dewalas devoted to this deity, including a celebrated one at Nirammulla, which is held to be the leading one in Ceylon provided for him.FIG. 43. The Nirammulla Dewalaya,CEREMONIES

WHEN the Tamil-speaking Vaeddas are about to leave their village on a hunting-trip they erect under a Velan tree, the tree whose wood is always used for arrows, a little shrine termed kuddram, like those of the Kiri-Amma, a rectangular structure supported by four upright sticks set in the earth, with the floor of the shrine about three feet off the ground. The top is usually arched over, and the inside is lined with white calico. In it are laid the bright yellow flowers of the Ranawara bush (Cassia aunculata) and those of the Red Lotus (Nelumbium speciosum), with two wicks soaked in oil or fat in front of them.

On the ground immediately in front of the shrine, resting on a white cloth, are placed a small * chatty * or earthen pot, holding rice boiled in Coconut milk, and in two circles round it seven small earthen vessels (kuncatti] and seven larger ones (mancatti) containing the other usual cooked food as offerings?meat, cakes, etc. Round these are thirty nuts of the Areka palm, laid on thirty leaves of the Betel vine. Water is sprinkled over these articles and in front of the shrine, obeisance is paid to. the offering, and the dedication is complete.

The wicks are now lighted, and the officiator, an ordinary villager acquainted with the service, dances in front of the offering, holding a flower-staff, called mugura, in his right hand. He is dressed in a good white cloth, and has a headdress of the same yellow and red flowers as those in the shrine, which covers the whole top of his head. The mugura is a stick, eighteen inches long, covered over with the same kinds of flowers, and having a looped handle in the middle, so that it may be held horizontally without touching the flowers.

He then chants in a Sinhalese doggerel some words the meaning of which he, being acquainted only with Tamil, does not comprehend, although he knows that he is asking

186THE PRIMITIVE DEITY, CEREMONIES 187

for deer. The purport, as nearly verbatim as I could follow them, is:?

" Pata tantana thanum 1 Om 1 Tana tantana ! I have placed for you a gift, cooked food, Deity of the Country [Decamu Hura]. Tana tantana tana ! Stopping a Sambar deer at the place where it is to be killed, O Protector and Friend of Dharma, at the corner where the Sambar is shot with an arrow make it over to us. When the Sambar has fallen make over three more to us."

These men ask for no other animals but Sambar deer. After the celebrant has again danced before the offerings, he sits down cross-legged in front of them, holding the mugura vertically, and makes 'dabs ' at, or points it at, the seven smaller vessels one by one. He then sets it upright in the middle of the milk rice.

The God now says in Tamil by the mouth of the celebrant:? " Go ye! To whatever place you go, you will overcome the thing encountered." The food is then eaten by the hunting party at the site of the ceremony. On their return no further offering is presented.

In case of sickness, the Lord of tife Hills is again appealed to. The officiator, the interested relative of the patient, proceeds to the front of the temple, and stands erect there, facing the doorway ; in his outstretched right hand he holds horizontally an arrow, near the butt-end, just beyond the feathers, with its point towards the doorway. He then says, " Lord God (Andavana Swdmf) I Hill-Lord ! By this, as it is brought and held in the hand, health must go (to the sick person)." He at once leaves without awaiting any intimation of the result of this order?or perhaps request, as the word which expresses an imperative * must' is also used, as in Sinhalese, with the meaning * hoping that (something) will occur/

The Village Vaeddas of the interior appeal in a similar manner to the same deity for the cure of sickness, carrying with them their bow at the time. They admitted to me, with a laugh* that holding the arrow with its point towards the shrine had the appearance of a threat; but the God Is thought to be so powerful that probably this is not their intention.i88 ANCIENT CEYLON

Those in the south prepare a little shrine, like that already described, under a shady tree, and offer inside it cooked rice, pieces of Coconut if available, a small cake made of rice-flour, and a little meat, these foods being purified by lustration of water; one lighted wick is fixed in front of the offering.

The officiator holds an arrow upright in his right hand, and while repeating his prayer to the God for the removal of the sickness makes little cuts with it at the rice, believing that as he cuts it the sickness passes into the rice. At the conclusion of the ofering, a ceremonial dance is performed by the celebrant, in front of the shrine, holding the arrow in his hand.

Now comes the peculiar part of the ceremony. Although the sickness is supposed to have been communicated to the rice, it is thought that a ceremonial sprinkling of water over the latter will drive it out again. This is now done, and the food having been thus purified is divided among the persons who are presenting the offering and the patient, and is eaten together with the other things offered.

This ceremony is called the Gale-Yak-maduwe bat pujdwa, * the Offering of the Rice of the Gale Yaka's Shed/ the shed being the shrine. Similar dances and offerings are customary in the neighbouring parts of Uva.

The Vaeddas of the interior and their Sinhalese neighbours also dance to the Gale Yaka and his wife the Kiri-Amma, in order to avert apprehended epidemic disease, or misfortune. This is before the occurrence of the sickness or bad-luck; after the sickness has set i^the village becomes ceremonially impure, and it is held by them to be imperative that no dance to the God should take place within its boundary, which includes all the neighbouring jungle. The dance, which is usually performed once or twice a year, and preferably, by those who keep an account of the days of the week, on a Saturday or Sunday, is commonly executed under an Iron-wood tree (Na, Messua ferrea) when it is in flower, or a Banyan tree (Nuga, Ficus indica}.

The dancer, a professional devil-priest, or a Vaedda, significantly termed Deyiyanne Kapuwa, * the Devil-priest of theTHE PRIMITIVE DEITY, CEREMONIES 189

God/ is dressed in a white cloth, and has a red handkerchief wrapped over his head. He also wears several bead necklaces, and any kind of bangles that his small store can furnish. If a Sinhalese man, he holds in his right hand a small dwude, literally ' weapon/ a stick roughly shaped like an arrow; if a Vaedda an arrow is held by him.

He now becomes * possessed' by the God (mayanwela, in Sinhalese), after which everything he does or says is supposed to be the action or speech of the deity himself. While dancing in front of the shrine containing the offerings, he chants verses in honour of the two deities who are being worshipped. The usual food offerings made at such ceremonies are presented, the dance being begun after the two lights which are placed in front of the offerings have expired. Tom-toms are also beaten as an accompaniment by those who have them.

But the special place for such dances to the God of the Rock, for the Vaeddas particularly, and also for the Sinhalese who live near them, is on the summit of precipitous crags on or near the top of certain hills of the district, on which this form of worship has been performed from ancient times. On these they dance once a year to ensure the general prosperity of the district. The Officiator, the Deyiyanne Kapuwa, is accompanied in the Vaedda ceremony by any two men as assistants, who alone climb up onto the rock with him. Among the neighbouring Sinhalese the assistants are the washerman who washes the Kapuwa's clothes, and the smith who made the God's emblem; the former stays at the foot of the crag, and the latter alone goes with the dancer to the summit. The three persons wear no special head-covering, but each one has a handkerchief on it, and is dressed in the Sinhalese ceremony in a clean white cloth. In the Vaedda ceremony, I was assured that each one is dressed in an old torn cloth, and not a new one. It appears to represent the traditional dress of their ancestors.

The dancer carries up the symbol of the God, which is not

an arrow, as one would expect, but a short-handled bill-hook,

a Ran-kaetta, * Golden Bill-hook/ in Sinhalese, or

Manna-kaetta, in the Vaedi dialect. The latter word probablyigo ANCIENT CEYLON

means 'Bill-hook of Honour'; or it may be connected with the word mantra, prayer, spell. He holds this in his right hand while performing the dance.

No particular day of the week is selected by the Vaeddas for the ceremony, all days being alike to them. It usually takes place during the daytime, but occasionally at night, torches being then carried by the assistants. This appears to point to planetary influences over the hour selected for it, which in such cases will be determined by a Sinhalese Kapuwa.

The dance performed by the Vaeddas is an extremely simple one. The body is slightly bent forward, with the elbows near the sides, and the fore-arms extended horizontally. In this attitude the performer lifts up and lowers his feet alternately, turning round gradually to the right while chanting verses in honour of the God, to the air, which begins each stanza :?Tanan tandeni ta'na ne'e, the last part being sometimes varied to tadi na ne, when repeated at the end of a verse. As my informant, himself an officiator at these services, was unfortunately obliged to leave immediately owing to his child's sickness, in order to arrive at his home, far away in the forest, before complete darkness set in, I had not an opportunity of writing down the invocation which accompanied his dance; it consisted of eight or ten four-line stanzas of a very simple character.

For three days before the dance the Kapuwa must make himself ceremonially pure by bathing daily, and by not entering a house. For three days after it there is the same restriction against entering houses. Apparently time is required, as in ordinary devil-ceremonies, for the divine afflatus to become dissipated, and while it lasts the * possessed' person lives in a state of tabu. During his whole life the professional Kapuwa must specially avoid eating, under penalty of death inflicted by supernatural power, any part of certain animals which are ' unclean ' to him. These are the Pig, Ura ; the large Monkey, Wandura (Semnopithecus priamus) ; the Peafowl, Monara; the Shark, Mora] and a large river-fish called Magura.THE PRIMITIVE DEITY, CEREMONIES^ i

This prohibition appears to have no connection with Hinduism, or the common Brown Monkey, Rilawa (Macacus 'pileatus), would be included, and also the Rat, as the vahana of Ganesa, and the Turtle as representative of Vishnu ; or some of these. I have already referred to the significance of the inclusion of the Shark. The Pig must appear in the list for the same reason, that is, as an eater of dead bodies, which might be those of human beings. The Magura is probably added for a similar cause. The plumage of the Peafowl is generally thought to be auspicious, but I am not aware if this is the opinion of the Vaeddas. The primary feathers of the wing are always employed for feathering arrows, which it will be seen, by their use in the religious services, have something of a sacred character attached to them. I can offer no suggestion regarding the inclusion of the Wandura, unless it be on account of its human appearance. I think it is clear that there is nothing totemistic in these prohibitions.

In the district of the Vaeddas, the following are all the hills on which are found the so-called ' Dancing Rocks ' (natana gal), one on each hill, of which I could obtain information:? Omungala near Rugama tank, Henanne-gala, Kokka-gala, Dambara-gala, Unakiri-gala, Mawara-gala, and one near Diwulana tank. %

It is among the north-western Kandian Sinhalese, however, that the ceremonies in honour of the God of the Rock, as he is there called, have survived, or have been developed, in the most complete manner. Yet it is evident that even there the cult has seen its best days. The dance is no longer an annual event at several of the rocks devoted to it; occasionally intervals of some years elapse between two celebrations, and in a few cases it has altogether fallen into disuse. This is said to have been caused by the death of the officiating priests, and the want of successors, and not through lack of support by those who provide the expenses.

In these districts, in all cases the dance, which is a very important part of the proceedings, and indispensable in the complete ceremony, takes place on a high precipitous projecting crag near the top of a prominent hill, or on the summit192

ANCIENT CEYLON

of the hill if it is a single bare rock. These rocks commonly face towards the south, but not invariably; and I have stood on one that was on the northern end of a long hill (Dolu-kanda) which has a high vertical precipice on its top, facing, due east, that might have been selected for the purpose, if desired. This fact is of some value as almost necessarily indicating the absence, from an early date, of any connection with sun or moon worship, at any rate with adoration of the rising sun, which is further emphasised by the performance of the ceremony on all the rocks at or after noon, and never in the early mornings. The occasional dances at night by the Vaeddas also prove that the cult is quite unconnected with sun worship.

The following is a nearly complete list of all the hills in the Kandian districts on which the Dancing Rocks are situated, together with the names of the dewalas at which the subsequent proceedings are carried on, and from which a procession accompanies the performer to the hill where the dance is to take place. An asterisk is prefixed to the names of those hills at which dancing has now ceased.

HILL.

DEWALA.

Gopallawe-gala Ganemulla. ? G5pallawa.

,, Maedagampola.

Deviyanne-

kanda. Gokaraella.

*Ganemulle-gala Ganemulla. Devagiriya

(Kabara-hinna

and Kaeta-

gala-hlnna) Nlrammulla. Hummuttiya-

wa Galpaya, or

Aematiya

Galpaya, in ?

Pallekaele

Forest.

*Nelliya-gala. Nelliya. *Manapaya-

kanda. Yapagama.

Kandalawa-kan-

da (two rocks) Kandalawa,

HILL. Kombuwe-gala.

*B6gamuwa-

kanda

Aragama-kanda. *Dolukanda:

(Rasamiina-

gala). Madahapola-

kanda. Dekanduwala-

gala.

Nikawae-gala. *Rock on Mor-

atta-oya, Andiya-gala (N.

C. Province) Bamba-kanda

.(MState District)

DEWALA. KSmbirwa, Malla-waewa

(abandoned).

B5gamuwa. . Ganegoda.

Barube.

Madahapola. Kanduboda-

gania. Nika-waewa.

Kahalla.

Andlya-gala

(Indigollaewa).

Galewela.

. Bdgamuwa-kanda.

. Kand a la wa- konda

2

50. D o I u - k 0 n d a

Dancing Rocks of the Gale Deviya.

193194 ANCIENT CEYLON

All but the last two are in the eastern part of the Kurunaegala district.

The ceremony takes place in the months Aehala (July-August), or Nikini (August-September)/ sometimes on a Monday, but generally and preferably on a Wednesday or Saturday, and never on a poya day (the Buddhist Sabbath, which is kept at each quarter of the moon), but sometimes on the day following it. Wednesday and Saturday are specially devoted to demon ceremonies, and are the two most inauspicious days of the week, and as such are invariably avoided for beginning any journey or work.

The months in which the dance is performed are two in which the full force of the winds of the south-west monsoon is felt in this district, and the work of the dancing-priest is thus on some occasions excessively dangerous on such exposed sites, a few of which can be reached only by means of ladders. In one instance, at Aragama-kanda, it is stated that the dancer was blown clean away and never seen again; and that any dancer escapes unhurt is attributed to the protection afforded by the God.

For two days prior to the ceremony the dancer must not enter a dwelling-house, and he usually lodges at the dewala ; for three days after it he is subject to a similar ban. He then bathes, anointing himself with lime-juice, and the restriction ends. He must never enter a house in which a birth has taken place, until a month has elapsed after the event? some say seven days only?nor a house in which a death has occurred until three months have passed, and then only after it has been cleaned and purified by having new cowdung plastered on the floors.

As long as he lives he must not drink any spirituous liquor, and he cannot, on pain of death, eat the flesh or eggs of Fowls, Kukuld?including any bird which bears this Sinhalese name. Thus, the prohibition applies to the Jacana, Diya-Kukula (Hydrophasianus chirurgus) ; the Spur-fowl, Haban-Kukala (Galloperdix Ucakaratus) ; and the Ground Cuckoo, Mti~ Kukula (Centrococcyx rufipennis). Included in the prohibition are also Pigs, Urd, among which is reckoned the Dugong^THE PRIMITIVE DEITY, CEREMONIES 195

called in Sinhalese the Sea-Pig, Mudu-Ura (Halicore dugong) ; Sea-Turtles, Kaesba; and Eels, Anda. The * Iguana/ Goya, is added by some, but not the Peafowl.

No devil-priest may eat these animals, whether he dances or not; they are termed ' unclean ' (kilutu). If he has once officiated as a priest they are forbidden foods for the rest of his life, whether he takes any part or not in the services of either the God of the Rock or demons.1 From infancy, it is customary for the male children of the priests not to be permitted to taste them, as possibly they may become devil-priests, and it is advisable to guard them from unnecessary defilement. The girls are allowed to eat them. Only men of good caste, rate-minissu, can officiate at the ceremonies in honour of the Gale Deviya; but the caste has no influence affecting these food-prohibitions, which are equally applicable to low-caste dancers who take part in the services in honour of certain demons.

We have here a different set of animals from those considered unclean in the Vaedda districts, with the exception of the Pig. The Fowl may be forbidden partly as a bird whose blood is offered to demons, and perhaps also as a household bird, the cock being often kept at Buddhist temples to awake the monks early in the morning. Many Buddhists think it wrong even to eat the eggs. The Eel may be included on account of its feeding on garbage, and because it resembles a snake, which has protective powers.

I know no reason why Sea-Turtles, and no other kind of Turtles or Tortoises, are forbidden food, unless it be the Turtle-incarnation of Vishnu, who was at least very highly respected in Ceylon in pre-Christian times, and even now is permitted to have his statue in the Buddhist wiharas. It was he who took the precaution to tie charmed threads on the arms of Wijaya and his comrades, in order to preserve them froiij the Yakshas.

The story of the incarnation goes back to the time of the

1 In Southern India, the Pusaris, the officiators at demon offerings, are also forbidden to eat Fowls, Pigs, Peafowls, and Turtles, as well as certain grains and pulse.196 ANCIENT CEYLON

great Churning of the Ocean in order to make Amrita. The Gods and Asuras or Demons agreed to work together for this laudable purpose, which was effected like the production of the sacred fire. They took a mountain, Mandara, for a twirling-stick ; and the King of the Serpents, Vasuki, allowed his body to be utilised as the cord to be passed round it once, and pulled at the ends alternately, the Gods holding it at the tail, and the Asuras at the head. At first the effort failed; the mountain sank in the water or mud by its own weight, until at last Vishnu transformed himself into an immense Turtle, and permitted it to rest and turn on his back. It is called a Tortoise in the translations of the legend; but being in the sea it must have been a Sea-Turtle.1

It is almost needless to remark that such a restriction must have been originated among a race who knew and ate the Sea-Turtle. It cannot have come from an inland district where the Sea-Turtle would be unknown. This excludes all inland tracts, as none but fishers or those living near the sea would be affected by it. It appears to date from ancient times; at the present day and for more than two thousand years the people in Geylon who are chiefly, or almost entirely, influenced by it in the case of the ceremonies in honour of the God of the Rock have never seen such an animal.

On the other hand, not one of the Vahanas of the Indian gods is prohibited as food ; even the Bull may be eaten by Sinhalese Kapuwas, as well as the Peacock and the Rat. It will be observed that these are Saivite 'vehicles/ and do not belong to the worship of Vishnu.

The dance on the rock takes place about noon, or in the afternoon, and it sometimes lasts for nearly an hour. The day is fixed some weeks in advance, in order to allow time

1 In the Ordinances of Manu, v, 18, the Tortoise is included among the animals which * the wise have pronounced eatable ' ; tame Cocks and tame Swine are excluded (v, 19). Monkeys come under the category of animals with five toes, which are forbidden. Peafowl are not

expressly excluded, but it is ordained (xi, 136) that the slayer of a

Peacock or an Ape must present a cow to a Brahman?the same fine as for killing wild carnivorous animals (xi, 138).THE PRIMITIVE DEITY, CEREMONIES 197

for the necessary provisions to be collected for the offering and feast which follow the dancing.

On the appointed day, the dancing priest, who is termed Anumaetirala,1 and not Kapurala (the ordinary title of a good-caste devil-priest), dons at the dewala the traditional dress of the God,2 consisting of a many-flounced coloured skirt or skirts, an ornamental jacket with puffed out sleeves reaching to the elbows, and especially a tall tiara-like conical white hat (toppiyama), made in three tiers or sections, as well as a jingling anklet, salamba (in Tamil, silampu), and any other usual ornaments, bangles, etc., of his profession.

He also takes in his right hand the Abarana, or symbol of the deity, a Bill-hook, nearly sickle-shaped, with an ornamental handle about two feet long, to which is tied a much smaller one, having its blade immediately below that of the large one. This latter is the symbol of the God's minister, the redoubtable Kurumbuda Devata. In his other hand he sometimes holds some flowers of the Areka-palm, merely for the dance, and not as offerings.

The Anumaetirala now becomes ( possessed ' (mayanweld) by the God, and henceforward his actions are no longer under his own control but are those of the deity. Holding the symbols of the deities, he marches to the Dancing Rock, generally about a mile from the dewala, accompanied by his two assistants, the smith who made the weapons, and the washerman who washes the Anumaetirala's clothes ; he is preceded by tom-tom beaters, and followed by an indiscriminate crowd of villagers. Meanwhile, others have collected near the hill which is his goal, selecting vantage grounds whence a good view of the proceedings is obtainable. The crowd often numbers several hundred people.

The Anumaetirala and his two assistants alone ascend the hill, the former being assisted by the other two to mount the crag, if necessary. At some places all three climb onto the

1 From Anumdtiya, sanction or command. Thus he Is the person who acts under the God's command, that Is, because he Is compelled

by the God to do it. It is supposed to be involuntary on his part,

2 See the Frontispiece.io8 ANCIENT CEYLON

.?

Dancing Rock, but generally only the smith goes onto it, in order to be ready to render him any assistance which may be needed on account of the wind ; in such a case the washerman waits for them at the foot of the crag.1

On this wild and often extremely dangerous platform, on some hiUs a mere pinnacle, usually hundreds of feet above the plain below, and in one case?(Dolukanda)?more than a thousand feet above it, and in full view of the spectators gathered there, the Anumaetirala now performs his strange dance, like that of all so-called devil-dancers. He chants no song in honour of the ancient deity (according to my information), but postures in silence with bent knees and waving arms, holding up the Bill-hooks?the God himself for the time being. In a rough outline drawing, representing this dance, on the wall of the ancient temple of the God at Mallawaewa (Fig. No. 42)?from which I obtained the cup and lamp illustrated below?the deity is drawn with a triangular beard, and he holds a Sword instead of a Billhook. When he begins to feel exhausted the performer brings the dance to an end, but sometimes his excitement makes it necessary for his assistant to seize him, and forcibly compel him to stop. He then descends from his dizzy post, assisted by his henchmen, and returns to the dewala with the tom-toms and the crowd.

While the party are absent at the Dancing Rock, the women of the village at which the dewala is established cook at the house of another priest, called the Mulutaen (kitchen) Kapurala, a feast consisting of cakes, milk-rice (rice boiled in coconut milk, that is, not the liquid found in the coconut when it is opened, which is never used, but some made by squeezing grated coconut in water until the latter acquires the colour of milk), rice, and also curries made of three kinds of vegetables termed tun-main.

The whole party from the rock come to this place, and the Anumaetirala examines and expresses his approval of the food, which has been laid ready for his inspection in earthen

1 At some sites it is said that only the Anumaetirala goes on to the rock on which the dance takes place. *:THE PRIMITIVE DEITYi CEREMONIES 199

vessels deposited on a clean white cloth on the floor, each covered by a small round grass mat. He then proceeds to the dewala, in which he replaces the Abarana, and removes the hat and the other habiliments of the God, being then no longer possessed by him. After this, he comes back to the house of the Mulutaen Kapurala, and he and all the rest of the people eat the food that had been prepared. I was informed that on one occasion seven and a half bushels of

Fio.Sl. Block Earthenware Cup,

Fia.52,Beaten Copper Cup. Fic».53. Black Earthenware Lamp.

Ancient Utensils of tlie Gale-Deviya,

rice were thus disposed of, making a good meal for five or six hundred people.

During all this time the deities have been left without food. After the conclusion of this feast and the indispensable chew of betel which follows it, the Anumaetirala, the Mulutaen Kapurala, and some of the people return to the dewala, and there the Mulutaen Kapurala himself cooks, in the Mulutaen-ge or kitchen, some of the same foods for the Gods. The Anumaetirala never takes any part in these duties, but200 ANCIENT CEYLON

when the cooking is finished it is he who makes the offering of the food. At other times, in the case of other Gods and demons, this is done by a third priest, called the Tewawa Kapurala, all having strictly defined duties.

On a yahana, an oblong stand or altar, with a flat top, prepared previously in the dewala, the Abarana, or insignia, of all the deities to whom the dewala is dedicated, are separately placed on a white cloth by the Anumaetirala, after sprinkling in front of it saffron-water, that is, water in which a piece of saffron or turmeric had been placed. Incense laid on burning charcoal deposited in a censer is also waved in front of it. This always follows the lustration with the saffron-water, and is invariably a part of all such purifications.

He now once more assumes the dress of the God, and other ornaments of devil-dancers, but not the tall hat, which is reserved for the dance on the rock, of which it thus indicates the special character. Its place is taken by a white cloth which covers his head and shoulders. The Mulutaen Kapurala now brings into the dewala the food which he has cooked, and the Anumaetirala offers it to the deities.

As a good illustration of this service, I shall take the case of a dewala at which the insignia of five other deities are kept. These are Pattini Devi, whose symbol is her hollow jingling bangle or anklet, called salamba ; Devata Bandara or Daedimunda, already mentioned in connection with the Vaeddas, whose symbol is a yakada-awude, ' iron weapon/ a thin rod of iron, thicker at the butt end than the other, which is bent over to one side ; Kumara Bandara, an Indian evil spirit or demon, the son of a king of Madura, whose symbol is a straight sword, kaduwa; Dahanaka Bandara, whose symbol is a bill-hook, and Yapawu Bandara, whose symbol is also a bill-hook, these two last being deified chiefs. There is also the sword of Kurumbuda Devata.

The Anumaetirala lays on the altar, on a white cloth extended in front of the Abarana of the seven deities, seven sets of pieces of fresh green Plantain leaf, to act as dishes, each consisting of two pieces, laid one over the other. These, like everything else used, are purified by sprinkling themTHE PRIMITIVE DEITY, CEREMONIES 201

with saffron-water and fumigating them with the incense.

He then takes four strips of Areka-palm flowers, and lays them on each plantain dish, so as to form a hollow square on each; and in the centre of each square he places a little of the newly-cooked rice and the other food, after first purifying it, and reverently offering it with both hands towards the Abarana. Great care is necessary to apportion the rice equally among the seven deities, so that none may be offended at receiving less than his neighbours ; this is a detail regarding which they are unduly sensitive. A little sandal-wood, handun, is now rubbed on each leaf, and one or two little earthenware oil-lamps are lit in front of the offering. In some places a narrow-mouthed, round-bodied flower pot, called kalasa, filled with Areka-palm flowers, is placed there, with an earthenware lamp resting across its mouth, on the flowers.

A yata yahana, or 'Lower Altar/ is now formed by a mat laid on the floor in front of the God's altar. On this six sets of Plantain-leaf dishes are first prepared for the reception of the meats, after the same purification as before ; and on them is divided the rest of the cooked food, after being first purified, and offered in the hands towards the Abarana, No light is placed on this altar. This food is said to be presented to six minor deities, called Devatawas, whose names are unknown. They are considered to be a form of demon; and judging by the practice elsewhere part of the food is doubtless an offering to all absent minor deities or demons collectively, who always expect to receive a small share when others are fed.

Invocation is now addressed by the Anumaetirala to the Gods of the upper altar, calling on them to come and partake of the feast prepared for them. After a short interval the foods are removed, and sliced Areka-nut laid on Betel leaves is deposited in its place, purified as usual, on the seven leaves of the upper altar, but none is presented on the lower one for the inferior personages.

Both before and after offering the food, tom-toms and other instruments are sounded as loudly as possible, so as202 ANCIENT CEYLON

to arouse the attention of the Gods, who are also held to feel satisfaction when this music is executed zealously and in an artistic manner. The instruments used for the purpose at the dewala in question are as follows :?one drum, dawula ; one common tom-tom, bera-gedi (not the long tom-tom used for demons, which Is called yak-bera); one double kettledrum, temmattan (used at wiharas) ; one horn trumpet, horandewa; one reed-pipe, naldwa; and one perforated conch-shell, hak-gedi. When all these instruments are emitting their loudest sounds it will be understood that the result is a fine medley of noise.

After offering the Betel, and while the instruments are being played, the Mulutaen Kapurala may dance, holding a sword and not the Billhook in his hand, and chanting at the same time verses in honour of the various deities ; but this is not an indispensable part of the ceremony. In the ?end, the Anumaetirala puts back the Abarana in their places in the sanctum, removes his habiliments, and thus brings the affair to a conclusion. He and the Kapurala, but no one else, may now eat the offerings, the deities being supposed to have taken as much of the essence of the food as they required.

The rock-dancing ceremony takes place not more than once a year in connection with each dewala of the God of the Rock; but this offering at the dewala itself is, in the case of this one temple, made every three months, and is devoted to all the deities of the dewala collectively, and not alone to the God of the Rock. The ceremony is the same at places where he and his minister are the sole recipients of the offering; in such case a separate altar is constructed for Kurumbuda at a lower level than that of his master, but well above the ground.

The whole service is considered to be in honour of Gods, and not to be# a demon ceremony. In all ordinary services for demons, meat in some form, or blood, is a necessary part of the food. It is clear that the Gale Deviya's aspect as a beneficent deity is alone kept in view in these proceedings.

To my mind, the detailed account which I have been ableTHE PRIMITIVE DEITY, CEREMONIES 203

to present respecting this god leaves no room to doubt that we have here the worship of the original deity of Ceylon, dating from pre-Buddhistic times. The oldest gods of the East were mountain deities. In Babylonia they were born on the mountains, and Enlil, the greatest of the early gods, was especially the God of the Mountain. Such also were the great Indian Gods, Indra, Vishnu, Rudra, and the Maruts.

What is the explanation of the curious fact that in most respects the God of the Rock resembles the Rudra of the early Aryans of India? It is unlikely that an aboriginal deity of the hated Dasyus would be elevated by the first Aryan invaders of India to the high rank indicated by his being addressed as ' the best among the Gods/ a term applied to Rudra. In any case, the earliest Aryans of north-western India could have no knowledge of the religious notions of the South-Indian hill tribes. Did the aborigines, then, take over Rudra from the Aryans at a later date ? This also seems improbable, as apparently the mountaineers, at all events, kept aloof from the invaders, and were little affected by their civilisation.

If they did not so borrow him, perhaps we may conclude that there was a wide-spread primitive belief in such a deity, extending not only throughout the hills of Central and Southern India, but also through the country from which the Aryans came. As some of the chief Aryan gods were hill-deities this must have been to some extent a mountainous tract. It could only be in such surroundings that a belief in mountain-gods could be developed.

Rudra was the parent and Lord of the forty-nine Maruts who were the deities of the storm, and who dwelt * on the lofty mountains/ So far as my information extends, he differs from the God of the Rock chiefly in being a Destroyer ; the later story of the destruction of the sixty t monks cannot be held to be sufficient proof that the original Vaedda God had this character.1 But Rudra's aspect as such is not strongly emphasised in the Rig-Veda, He was pre-eminently

1 Dr. C. G. Seligmann has, however, met with an invocation in which the Indigollae Yaka is referred to as drinking human blood.204 ANCIENT CEYLON

the beneficent kindly deity of the Aryans, the Health-Giver, the ' Best of all Physicians/ In this respect, Rudra and the Vaedda deity have identical functions and attributes, as the following extracts from the Rig-Veda, (Griffiths' translation,) show plainly.

Book i, 43.

i. What shall we sing to Rudra, strong, most bounteous, excellently wise ?

4. To Rudra, Lord of Sacrifice, of hymns, of balmy medicines, we

pray for joy and health and strength.

5. He shines in splendour like the Sun, refulgent as bright gold is he,1

the good, the best among the Gods.

6. May he grant health unto our steeds, well-being to our rams and

ewes, to man, to woman, and to kine.

Book i, 114.

1. To the strong Rudra bring we these our songs of praise, to him the

Lord of Heroes with the braided hair,

That it be well with all our cattle and our men, that in this village all be healthy and well fed.

Book ii, 33.

2. With the most saving medicines which thou givest, Rudra, may I

obtain a hundred winters. Far from us banish enmity and hatred, and to all quarters maladies

and troubles. 4. Let us not anger thee with worship, Rudra, ill praise, strong God !

or mingled invocation: Do thou with strengthening balms incite our heroes ; I hear thee

famed as best of all physicians.

10. Worthy, thou carryest thy bow and arrows ; worthy, thy rnany-hued and honoured necklace ;

Worthy, thou cuttest here each fiend in pieces ; a mightier than thou there is not, Rudra.

11. Praise him the chariot-borne, the young, the famous, fierce, slaying like a dread -"beast of the forest.

O Rudra, praised, be gracious to the singer; let thy hosts 2 spare us and smite down another.

1 In modem paintings the God of the Rock is represented with a skin of a golden yellow colour. t The Maruts, the Storm Gods.THE PRIMITIVE DEITY, CEREMONIES 205

Book v, 42.

ii. Praise him whose bow is strong and sure his arrow, him who is

Lord of every balm that healeth. Worship thou Rudra for his great good favour.

Thus, although their symbols differ at present, though perhaps not originally, one carrying a billhook, and the other the bow and arrow, both are great sickness-removers from man or beast, the givers of health to their worshippers ; and as such, both being gods of the mountains, they may have been in their origins the same deity. If so, the belief in such a god must extend back to a very remote age.

No reason has been discovered why certain hills only, sometimes in close proximity, were selected as the sole spots on which the dances to the God of the Rock should be performed. I could not observe that a specially sacred character is attributed to them by the Sinhalese, although I obtained some verbal evidence, possibly of little value, in favour of it; * but I believe the Vaeddas have some idea of the kind regarding those in their district, and Mr. F. Lewis heard this said of Kokka-gala.

Ritigala and many other prominent Kandian mountains are not known, according to tradition, ever to have been the sites of Dancing Rocks. Even a commanding rocky peak a few miles from Kurunaegala, known as Yak-dessa-gala, * the Devil-dancer's Rock/ is stated never to have been one of them, notwithstanding its suggestive name.2 This is

1 Mr. Bell states regarding the Dancing Rocks at Indigollaewa and Nikawae-kanda, ' The rocks are so sacred that no one dares venture near them, except on the pemhaem [procession] day; even hunters worship as they pass ! (Arch. 'Survey. Annual Report lot 1895, p. 5.)

In the Kurunaegala districts I have never known any one show any reluctance to go with me onto the hills, or, even onto the Dancing Rocks. I have also seen men go readily to cut grass on such a hill, and a party of villagers once spent a night with me at a cave close to one of the rocks.

2 In the work on the Kohomba (or Kosamba) Yaka to which allusion is made in the Preface, Vaeddas are mentioned as living at this hill and at Alagala, on the side of the railway to Kandy. In it the word * Vaeddas' is written Vaeddan and Vaeddhan ; the latter shows the transition from Vyadha.2o6 ANCIENT CEYLON

probably correct, as no Yak-dessa, a title applied only to dancers of two very low castes, could ever take any part in the services in honour of this supreme deity of the country. With the decline of the cult, the use of many rocks for the dancing ceremony may have been abandoned and forgotten. In early times its practice must have extended over the whole country.

From the foregoing statements it may be concluded that the God of the Rock is a form of the original Rudra, who was developed at a later date into the great deity Siva.

In his South Indian form he appears to have had two sons, Eiyanar or Ayiyanar, ' the elder brother/ and a younger one termed Ilandari, ' the youth/ in Tamil, or Bilinda, ' the child/ in Sinhalese. To the sectarian feelings of the worshippers of these younger deities may be due the story of their enmity. Both are Guardian Gods of the forest districts of Ceylon, and as such both may have the title Wanni Deviya, God of the Wanni.

Having these sons, the father must have had a wife, who would be the Hill Mother, Giri-Amma, a word which became Kiri-Amman in Tamil, and thence Kiri-Amma in Sinhalese. As the mother of Ayiyanar, she must be identified with the Indian goddess termed Mohini.

When Rudra had developed into Siva the northern form of th£ Hill Mother, Parvatl, supplanted this southern deity as the wife of the Hill God in Hinduism.

According to the Maha-Bharata the birth of Skanda was 'full of all mysteries/ and in one legend he was the son of Rudra, Thus, if the God of the Rock be Rudra it would seem that, as his son, Ayiyanar may be a form of Skanda; and in that case Ilandari or Bilinda may be a southern form of Ganesa. The latter's name in Tamil is PUleiyar, 'the child/ and both deities alike were killed while young, and revivified.Part II STRUCTURAL WORKS

PAGE

VI. THE ARCHAEOLOGICAL VALUE OF BRICKS . . 209

VII. ANCIENT ROCK CUP-MARKS ..... 221

VIII. THE LOST CITIES OF CEYLON .... 235

IX. THE EARLIEST DAGABAS . . . . . 261

X. THE EARLIEST IRRIGATION WORKS .... 347

w, v , *VI

THE ARCHAEOLOGICAL VALUE OF BRICKS IN CEYLON

THE references In the early histories of Ceylon to the construction of any very ancient structures that can be identified at the present day, with the exception of some reservoirs and a few special dagabas, are so rare and meagre that it is almost impossible to learn from the existing writings-anything of value regarding the ages of nearly all the remains of buildings of various kinds that are scattered over the whole island. Even in the case of such well-known edifices as the earliest and most celebrated dagabas there are many points of great interest to the antiquarian respecting which the histories are silent. As an instance of the neglect of the chroniclers to transmit to us a satisfactory record of buildings of even great importance, I may cite the absence of any reference to the addition of an outer shell enclosing the celebrated Ruwanwaeli dagaba which was built at Anuradhapura in the second century B.C.

By a long series of measurements and sketches taken whenever opportunities occurred for a period of more than twenty years, I endeavoured to ascertain whether the mouldings and decorations of the various edifices disclose any types of detail that afford a clue to the period when the buildings at which they are found were constructed or repaired. In Europe, each arch or moulding and almost any kind of decoration is stamped, as it were, with the approximate date of its construction ; and it seemed possible that some similar gradation might be discovered in Ceylon. It proved, however, to be nearly hopeless to expect to meet with much success in this research, since it was ascertained that nearly identical mould-210 ANCIENT CEYLON

ings were reproduced in buildings the erection of which is certainly separated by many centuries.

Probably the only conclusion to be arrived at, from an examination of these details, is that the simpler forms of outlines or decorations in buildings of a similar class may often indicate an earlier date of erection than the more elaborate ones. But this would be a very untrustworthy guide, as such details might be affected by the amount of the funds available for the work. There are also some peculiarities in the various modes of building with stone, and in the style of decorative design, that may point to earlier or later work. Thus, there are three entirely different types of holes cut for wedging stone, which indicate early, middle-age, and twelfth century work. But as general guides to the ages of structures It is evident that conclusions derived from such data are too vague to be of much use for practical purposes.

As a last resource, we are reduced to the bricks and brickwork. It has been already learnt that in some countries, as in Rome and Persia, the sizes of the bricks employed in buildings afford a valuable guide to the date of their manufacture. In India, although no definite scale of the dimensions may be possible, it is at least known that bricks of large sizes are a trustworthy indication of early work. I shall endeavour to show that in Ceylon, also, they may be utilised, in some instances, to a much greater extent than in India, but within somewhat wide limitations.

In order to obtain a satisfactory basis for such a conclusion the first requisite is a list of the dimensions of the bricks used in certain structures of varying ages, the dates of the construction or restoration of which are already known from the records found in the histories or preserved in the inscriptions that have been cut on them. It is in this respect that the chief difficulty lies. Even when a building is stated to have been erected during the reign of a special king, it may have been subsequently enlarged or repaired at some unknown time ; and we might thus be led to accept as bricks of a certain age some that were burnt centuries afterwards. It was also a common custom, in the case of slight repairs that wereARCHAEOLOGICAL VALUE OF BRICKS 211

executed at a later date, to utilise ancient bricks and cut stones brought from some pre-existing ruin of the neighbourhood.

But as knowledge of the different types of bricks and building work accumulates, the greater part of these difficulties tends to be overcome. For instance, when any extensive repairs have been carried out we always find a large proportion of broken bricks laid in the re-built work, and nearly invariably bricks of two or even three or four sizes, which have been moulded at the time when each repair was done, or additional work built.

The data on which the value of any tables of the sizes of the bricks must chiefly depend are as follows. The list of structures of various periods will probably be accepted as belonging to the dates here assigned. For the basis on which the first dates rest, reference may be made to the genealogical table of early kings at the end of the chapter on inscriptions, and to the remarks appended to it.

I. The large dagaba, called the Maha Saeya, at Mihintale, was constructed by King Devanam-piya Tissa in about 240 B.C.

2 and 3. The older dagabas at Tissa (or Magama) in the Southern Province, the Maha-naga dagaba and the Yatthala dagaba, date from the reign of King Maha-Naga, that is, the second half of the third century B.C.

4. The Sanda-giri dagaba, at the same place, belongs to

the time of Kakavanna-Tissa, who reigned in the first half of the second century B.C. (Pujavaliya, p. i6.)

5. The Miriswaeti dagaba at Anuradhapura, was built by

king Duttha-Gamini in 158 B.C.

6. The inner work of the Ruwanwaeli dagaba at Anura-

dhapura was built by the same king in 137 B.C.

7. The Lankarama dagaba at Anuradhapura was erected

by King Watta-Gamini, 76-88 B.C.

8. At MiUaewa-gala wihara in the North-central Province,

near Tantirimalei, an inscription in letters of the first or second century A.D,1 refers to the construction of

1 This inscription, a.facsimile of which is given in Fig. No. 153, is as follows :?

(i) Sidha. Wihara mawita puriha cata gowaka Sivaha lene (2)212 ANCIENT CEYLON

the wihara in the reign of King Naga, that is Ila-Naga, 38-44 A.r . or one of the earlier Nagas of the second century.

9. Hurulla Tank, in the North-central Province, was made

by King Maha-Sena, 277-304 A.D. (Pujavaliya, p. 27.)

10. Padawiya tank, in the North-central Province/appears to have been built by King Maha-Sena. According to my information, it is called in a rare manuscript (the Maha Jalanandana) Maha Ratmala tank, which is included in the list of those made by him (Mah. i,

P* 151)-

11. The Vannatti Palama, the dam on the river below

Padawiya, for utilising the water of that reservoir, is of about the same age.

12. Nirammulla Tank, in the North-western Province, on the Kimbulwana-oya, is included in the same list, as Kumbhila-vapi, and is of the same age.

13. The inner room of the building called the Dalada Mali-

gawa,' the Palace of the Tooth-Relic/ at Anuradhapura was probably built early in the fourth century A.D.

14. Topawaewa, the tank at Polannaruwa, was made by King Upa-Tissa I, 370-412 A.D. (Rajavaliya, p. 54).

15. The Gallery at Sigiriya was built by Kassapa I, 479-497 A.D.

16. The Rankot, ' Golden-Spire/ dagaba at Polannaruwa was built in the latter part of the twelfth century A.D.

17. The Thuparama Hall at Polannaruwa was built in the twelfth century.

18. The wall of the Fortifications at Polannaruwa also belongs

to that period.

ca(tu disa sa)gaya niyate. Naka rajaha rajahl mawita. Hall I Having built the wihara, the cave of Siva, Guardian of the City Dagabas, is assigned for the community of the four quarters. Constructed in the reign of King Naga.

The meaning of Cata is doubtful when qualified by puriha ; as the word often stands for cattya in inscriptions I have given it this interpretation. Siva or Siva may have been an official whose duty it was to undertake the repairs, whitewashing, etc., at the Anuradhapura dagabas.ARCHAEOLOGICAL VALUE OF BRICKS 213

The following table exhibits the contents and mean dimensions in inches of the bricks used in this series of ancient works. I add also, for comparison, the sizes of the bricks used in the large dagaba at Sanchi in India, which dates from the time of Asdka1 (263-222 B.C.), although the measurements are only given roughly by Sir A. Cunningham in his work on The Bhilsa Topes, p. 270.

Because of a reference to the name Cetiya-giri in Tumour's edition of the Mahavansa i, 49, as the place where Devi, the mother of Mahinda and wife of Asoka lived, and as there was a Buddhist heresy in the fourth century B.C. called 'the Cetiya schism/ Sir A. Cunningham stated his belief that this dagaba may be as old as 500 B.C.,2 but his argument is not satisfactory. In the first place, the corrected reading of the word in the Mahavansa is Vedisa-giri, Vedisa being mentioned in several of the inscriptions found there ; and in the second place, the name Cetiya-giri is not a proof that a dagaba existed there, the early meaning of Cetiya being merely {a religious building/ and only secondarily a dagaba. The Mahavansa says plainly that the wihara at Vedisa was established by Devi, the wife of Asoka, who himself left an inscription at it.

I look upon the Sanchi bricks as of the greatest interest. They are perhaps the earliest Indian bricks to which a date can be attached,3 yet their dimensions demonstrate, without any doubt, that others of the same shape had been burnt in India long prior to this date. The measurements bear no definite ratio to each other. The length of the original

1 The Bhilsa Topes, p. 124. 2 The Bhilsa Topes, p. 271, note.

3 They are far from being the earliest Indian bricks, of course. In addition to the Greek references to the use of bricks in the structures of the fourth century B.C., the employment of burnt bricks one foot square, for building the altar of Nirriti, the Goddess of Destruction, the mother of Fear, Terror, and Death, is enjoined in the Satapatha Brahmana (vii, 2. i) :?' They measure a foot square . . . they are unmarked?[an expression which shows that it was the custom to mark some other bricks] . . . they get baked by rice husks , . . they are black.' They were to be laid loosely, evidently not like other bricks? * he does not settle them?settlement being a firm footing?lest he give a firm footing to evil/ The date of this may be any time in the early part of the pre-Christian millennium.214

ANCIENT CEYLON

bricks of this type would almost certainly be fixed at the cubit, like those of Ceylon, and the breadth and thickness would be simple fractions of the length. Yet nearly two inches had already been struck off the length, and probably an equal amount off the breadth when the Sanchi bricks were burnt. I refer here to this Indian question as it is of some value in connection with possible early brick-making in Ceylon, where this art may have been practised by 300 B.C., if not in the first half of the fourth century B.C., as I have explained in dealing with the first irrigation works.

No Name of Structure. Date. *S toO 3 Breadth. Thickness. Breadth X Thickness. Contents. |

B.C.

Large dagaba, Sanchi . 3rd Cent. 1 6*OO IO-OO 3-00 30 480

I Maha Saeya, MiMntale N.C.P. 3rd Cent. 17-92 8-87 2-91 25«8 461

2 Mahanaga dagaba, Tissa S.P. 3rd Cent. 17-35 8-84 2-83 25 434

3 Yatthala dagaba 3rd Cent. 17-85- 8-64 2-9O 25 447

4 Sanda-giri dagaba, ? ? Early 2nd Cent. 17-14 8-67 2-8 1 24-4 418

5 Ruwanwaeli dagaba,

inner work . . . N.C.P. 2nd Cent. ? 9-67 2-79 27 ?

6 Miriswaeti dagaba,

Anuradhapura . ? 2nd Cent. ? 10-41 3-0 31-2 ?

7 Lankarama dagaba,

Anuradhapura . ,, Early ist Cent. 17-37 8-94 2-62 23-4 407

A.D.

8 Millaewa-gala dagaba . ? ist or 2nd IS'57 8-00 2-72 21-7 339

Cent.

9 Hurulla tank, central

sluice .... ,, Late 3rd Cent. ? 8-40 2-70 22-7 ?

Hurulla tank, high-level

sluice ..... ? »» ? 8-35 2-51 21 ?

10 Padawiya sluice . ? ,, 14-02 8-50 2-46 2O-9 293

ii Vannatti Palama . ? ,, 13-61 8-48 2-70 22-9 311

12 Nirammulla sluice . . N.W.P. »i ? 8-45 2-70 21-5 ?

*3 Dajada. Maligawa, inner

room . . . . . N.C.P. Early 4th. Cent. 14*10 8*AZ 2»C2 21*3 7QQ

14 Topawaewa sluice . ? Late 4th Cent. *"* T-3 7.40 * J* 2*43 f,t. ^ 1 8*OO JWV

15 Sigiiiya gallery . . ? Late sth Cent. I3-09 7.32 2*3* 16-90 220

16 Rankot dagaba, Polan-

naruwa .... ,, 1 2th Cent. 12-52 8-40 2-OO 1 6-80 2IO

17 Thuparama hall, inside,

Polannaruwa. . ? »j I2OO 8-36 2-OO 16-70 2O I

18 Fortification wall, Polan-

naruwa .... Jt »» I2-63 8-20 2-OO I6-40 207

! j

ARCHAEOLOGICAL VALUE OF BRICKS 215

Although when the bricks used in other buildings are examined the actual irregularities in the sizes become much more evident than this table shows, if the contents and area of the side be alone considered, it is quite clear that there is a generally-diminishing scale in the dimensions of the bricks from the earliest period down to the thirteenth century. But although the gradation is found to be often characterised by irregularity, this is not present in such an excessive degree as to prohibit the use of the dimensions?with a certain amount of discrimination?for determining the probable dates of the structures in which the bricks are found. For instance, the short table already given would at least enable any one to distinguish, by the bricks alone, a work of the tenth or twelfth century from one of the second or third century A.D., and the latter from one of pre-Christian date. Even if some exceptions occur in which the age of the construction is doubtful, or even with regard to which a dependence on such measurements might lead to an actual mistake in the time, they should not be allowed to outweigh or to throw much doubt upon the general advantage to be attained by the use of such an accessible method of ascertaining or corroborating the probable dates of structures.

In taking the dimensions of the bricks it is of less importance to measure a great number of lengths than to take a good series of thicknesses, so as to obtain a trustworthy mean thickness; the breadths occupy an intermediate position in value. The reason is plain. The average thickness is about 2| inches, one twenty-fifth of which is only one-tenth of an inch, a dimension that can hardly be correctly noted on ordinary bricks, which often vary to this extent, or more, in different parts of the same brick ; and thus it can only be accurately measured by taking the mean thickness of several bricks. In the case of the lengths of the bricks, however, one twenty-fifth of even the shorter bricks is half an inch, beyond which the error due to utilising the average size of only two or three bricks as the mean length is not likely to extend. Thus an error of half an inch in the length is of equal importance to an error of one-tenth of an inch in the thickness.2x6 ANCIENT CEYLON

A clear understanding of this fact is of practical service, firstly, because it saves time on the ground, especially in searching for more whole bricks than are necessary ; and secondly, because the accidents of time or the repairs of edifices have resulted in the fracture of nearly all the longer bricks, with the exception of those laid in buildings that were specially protected from damage by their mass or situation.

It is still more important in another way. In cases where no whole lengths can be found it permits the use, within certain limits, of a length calculated from the thickness or breadth, or derived from a comparison with other similar bricks, the error in such instances being probably often little more, in proportion, than that made when only two or three measurements of the thickness are possible. This is especially the case if the bricks have been moulded on the ground, and not on planks or tables.

On comparing a large series of measurements of bricks employed in various parts of Ceylon it is clear that the proportions varied in different periods. In the earliest times the length was commonly about six times the thickness, and the breadth was about half the length. Afterwards, the length was reduced to about five or even four times the thickness, though it never reached the English ratio of three times the thickness. The breadth also latterly varied between one-half and two-thirds of the length, but was commonly near the latter ratio.

The contents fell from a possible maximum of about 673 cubic inches 1 to 77 cubic inches; and the area of the side was reduced from 34 square inches to 7-7 inches. In actual dimensions, the length varied from a possible 19-8 inches to 8-2 inches, the breadth from 10-4 inches to 5-0 inches, and the thickness from 3-4 inches to 1*55 inches.

I have consigned to an Appendix a Table containing the sizes of the bricks measured at a large number of ruins in Ceylon,

1 The contents of the largest bricks of which the three dimensions were actually measured was 583 cubic inches ; but at three other sites larger ones appear to have been burnt, although unfortunately

their lengths could not be measured, no unbroken bricks having been

found (see Appendix).ARCHAEOLOGICAL VALUE OF BRICKS 2x7

which it is desirable to preserve for the use of local archaeologists, and in some instances for general reference. In preparing it I was confronted by the difficulty caused by the .absence of the lengths in places where no whole bricks could be found, or were not in a position where their lengths could be measured, as in some dagabas where the outer bricks are ?all' headers '. In these examples I have inserted in brackets in some cases an approximate length and contents, obtained by making the length a definite proportion of the breadth or thickness. Of course it will be understood that the figures so included in brackets merely indicate the probable length and contents, and nothing more.

As regards the periods into which the table has been divided, it is evident that their limits cannot be accurately defined. It may at least be asserted with some confidence that the first one, ending at about the Christian era, is nearly correct; and the same remark may possibly be applicable to the second period.

Owing to the slight change in the sizes from the sixth to the twelfth centuries there will always be considerable uncertainty as to the dates of that period. It is probable that by a reference to the table, however, the date of any bricks may be fixed at that time without an error of much more than one hundred and fifty years ; and prior to that time usually within the limits of about one hundred years.

I give a few examples of the application of the information obtained from the table in determining the ages of some structures. Many others will be found in the following chapters, where I have made full use of it.

The Thuparama wihara and dagaba at Anuradhapura were built by Devanam-piya Tissa (245 ? B.C.) ; but the latter was at least partly pulled down and rebuilt subsequently, and in any case the bricks used in it, which would doubtless consist chiefly.of those originally employed, are covered up by a coat of plaster. In a small building which still remains on the same platform, and close to the south side of the dagaba, .there are bricks measuring 19 inches by 9-15 inches by 3 inches, with a contents of 521 inches, and a side area (which in future218 ANCIENT CEYLON

will be termed Bt, that is, the average breadth multipled by the average thickness) amounting to 27-4 square inches.

From an inspection of the table it is safe to conclude that these bricks must belong to some period B.C., and possibly to a late date in the second century, or early in the first century. The building in question must have been a house for statues or relics, or both, no buildings intended for the occupation or personal use of the monks being ever permitted within the inner enclosure round a dagaba in Ceylon, according to information given to me by Buddhist monks. References to such a building occur in the histories, and are given in a later chapter on the dagabas. It is only the measurements of the bricks which prove that the building is of pre-Christian age.

The size of the bricks found in the Maenik dagaba at Tissa. shows that it was built at about the same time as the Sandagiri dagaba at the same city; or in other words that it was erected in the first half of the second century B.C.

The table also proves that there are many other works that in all probability date from the last three centuries B.C., and it may be also stated as a general rale that where any early bricks are found in rock-caves at which inscriptions in the earliest characters occur, some of them are of a size which indicates that they belong to the same period.

The great value of the bricks in assisting in the determination of the ages of some of the dagabas at Anuradhapura is apparent in a later chapter.

The bricks of the structures built at Anuradhapura in the second half of the second century B.C., and early in the first century B.C., exhibit great irregularity compared with those at both earlier and later works. It is not known why such variations were made in the sizes at this period. Perhaps it was a time of experiments, and trial was being made of some bricks of a larger size than had been used previously. The difficulty that must have been experienced in burning and handling such clumsy and heavy ones apparently soon led to the adoption of smaller and more convenient dimensions,

At Polannaruwa, the larger bricks used in the little Pabulm,,ARCHAEOLOGICAL VALUE OF BRICKS 219

' Coral/ dagaba are demonstrated by the table to be of pre-Christian age ; while those at the Gal-wihara at that city belong probably to the first century after Christ, and agree closely with the largest ones at the sedent Buddha at Tantirimalei Mr. Bell, the Government Archaeologist, has already drawn attention to the similarity of the carving at these two places, the works at which he attributed to the same period.1

Although the whole construction at the Gal-wihara is credited to King Parakrama-Bahu I in the Mahavansa, it is now shown by the bricks that his work there consisted only of some repairs at the old structure, and perhaps some additional rock excavation and carving. It is impossible to believe that the larger bricks found at these two sites were brought in each case from pre-existing ruins of their vicinity, at which, by some wonderful coincidence, bricks of practically identical sizes had been burnt more than twelve hundred years before. If the Tantirimalei bricks were burnt for use at the work at which we find them, so also were those of the Gal-wihara ; and it is certain from the table that they do not belong to the twelfth century.

In the case of the irrigation works, the table, as might be expected, yields some interesting information, part of which is utilised in later chapters. It will suffice here to draw attention to the evidence of the bricks at Pavat-kularn, Sangili-Kanadara tank, Batalogoda, the Alle-kattu dam, and Nuwara-waewa, the largest reservoir at Anuradhapura, as well as the channel for filling it. The construction of not one of these important works is mentioned in the histories. Other works not now described, such as theMamadu tank in the Northern Province, and Katiyawa tank, and Kltikadawala tank, and some others, in the North-central Province, appear to have been formed in pre-Christian times, although these also are not mentioned by the historians.

At the head-works of the channel for conveying water to the Giant's Tank, it is the bricks alone which prove the antiquity of this great irrigation scheme, as explained in the chapter

1 Annual Report, 1896, p. 8.220 ANCIENT CEYLON

on the ' Lost Cities/ Even without the inscription that was discovered at the reservoir, the dimensions of the bricks would enable any one acquainted with this table to affirm that the enlargement or improvement of this irrigation work was undertaken in about the twelfth century.VII

ANCIENT ROCK CUP-MARKS

AS the Vaeddas were unacquainted with the art of smelting iron, or making any metal tools, and appear never to have had any stone tools of their own manufacture excepting very rudely made arrows and ' scrapers/ etc., no early stone-cutting in any form can have been done by them. What knowledge of the art was ? possessed by the primitive Nagas is quite unknown, as no work that can be attributed to them has been discovered, nor, I believe, have any stone tools or weapons been found in the northern part of the island. Any early rock-cutting must thus have been done by the later immigrants or those who learnt the art from them, or by persons employed by them.

Some of the most ancient and undoubted work in stone which can be recognised at the present day in Ceylon consists of the cutting of the katdm, or drip-ledge, over the earliest caves, and the carving of the earliest cave inscriptions under it?(both, however, works that indicate a good prior acquaintance with the ordinary use of the stone-cutter's chisel)? in the second half of the third century B.C. For reasons which are given in a later chapter, there is a possibility that some excellent stone-cutting in some sluices may be of about the same date.

It is certain that the men who employed the tools for such purposes were not mere learners of the art of trimming stone. The cutting at the earliest cave inscription exhibits a freedom and accuracy of touch which are a clear proof of previously-acquired skill. It may be concluded with certainty that these stone-cutters had either been brought over for the special purpose from India?where the Bharhut and Gaya sculptures

221222 ANCIENT CEYLON

prove that stone-cutting had been practised for a long period prior to the reign of Asoka?or had possessed in Ceylon a long acquaintance with the art, even although any very early work done by them cannot be recognised now. , The Mahavansa (i, p. 149) refers to a ' colossal and beautiful stone statue' of Buddha which was carved during the reign of Devanam-piya Tissa, probably in about 235 B.C. (and which may be the large statue now at the Abhayagiri monastery) ; but a special work of this kind might be done by one or two men imported from India, and does not prove, like the other works, a general knowledge of stone-dressing in Ceylon at that period.

Neither dolmens nor circles of stone posts, on which cup-markings are sometimes cut in other countries, being known in Ceylon, the only cup-shaped holes that have been met with occur on the natural faces of rock masses or boulders. Some later outline carvings were also left on a few stone slabs or steps, evidently the work of the stone-cutters who were engaged in trimming them; or were cut on rocks as boundary marks of the various districts into which the island was subdivided.

The shallow cup or saucer-shaped holes such as are found on rocks elsewhere are uncommon in Ceylon, and I know of only a few places where they occur. One of these is the hill Ritigala, where Mr. H. C. P. Bell, the Government Archaeologist, informed me that he met with some, accompanied, I understood, by circles. He has not yet published a description of them, I believe.

Another site is a rock lying in the bed of the Kallaru, a stream in the Northern Province. As an ancient stone dam of unknown date, but probably pre-Christian, now called the Allekattu (Fig. 147), was built across this stream at the place, the work of cutting the holes appears to have been done by the men engaged in its construction. The holes are six in number, and in shape are excellently cut deep saucers, with well-smoothed sides and bottoms. The illustration, (Fig. No. 54), shows that after making seven wedge-holes in order to split off this stone for use in the dam, the masons left it untouched,at Ailekattu Dam

Boulder at

FIG. 56. -mediya- gala.

Shallow Cup-holes.224 ANCIENT CEYLON

evidently on account of the cup-holes in it. They are thus at least as old as the dam. Five holes are distributed nearly in the circumference of an ellipse, with a larger hole at its centre, an arrangement that at once recalls the mode of placing the food offered to the God of the Rock by the Tamil-speaking Vaeddas. The width of the whole design is n inches, each hole being from ij to 2| inches in diameter, and about i inch deep.

A third place at which I have met with these holes is on a nearly flat-topped boulder at the side of a path in the forest, at a place called Sigarata-hena, near Pulugannawa tank, in the Eastern Province. The plan (Fig. No. 55) shows their relative positions. Near them, on the same rock, there is a shallow circular channel one inch wide, half an inch deep, and i6J inches in diameter, with a tapering radial cut at one side, apparently to drain away water or oil. At a distance of 18 inches from this there is a peculiar sunk rectangle, measuring 23 inches by 15 inches, with a short curved channel cut from one corner. In a north-and-south line passing through these, and 7 feet 6 inches from the centre of the circle, there is also a hole of the larger and deeper type next described, 8J- inches in diameter at the mouth and 5J inches deep, with a rounded bottom. The boulder on which the holes are cut is 24 feet long.

In this case the holes are close to the site of an important early monastery, and they may thus be assumed to be the work of the men employed in building it. The brick fragments at the place are 2-5 inches thick and 8-8 inches wide ; thus Bt. (the mean breadth multiplied by the mean thickness) is 22 square inches, and the contents becomes 330 cubic inches if the length was six times the thickness, or 275 inches if it was five times the thickness. These dimensions point to the second, third or fourth centwy A.D. as the pencil; when the bricks were burnt, and the hojes may be of the same age.

A fourth site is on the sloping surface of an immense rock termed Gal-maediya gala, * the Stone-Prog Rock/ which rises high above the surrounding forest, two and a half miles above the point where the river Siyambalan-gamuwa-oya crossesROCK CUP-MARKS 225

the minor road from Kurunaegala to Anuradhapura. There are traces of a small building on the crest of the rock, and the remnants of a monastery on another rock termed Nelun-gala,

* Lotus-Rock/a short distance away ; while on the other side of the Frog Rock is the embankment of the Siyambalan-gamuwa tank, where a considerable amount of stone-work was used. The cutting of the holes may in this case also be attributed to some of the men employed at these works. The bricks used at the sluice of the tank and at the monastery at Nelun-gala, are of similar sizes, and indicate the second, third, or fourth century A.D. as the time when they were made. They average 2-56 inches thick, and 8*45 inches wide ; Bt. is 21*6 and the probable contents does not exceed 332 cubic inches.

The holes (Fig. No. 56), which are in roughly parallel rows that run north and south, are 67 in number. The lines of holes are in five pairs, each line consisting of either six or seven holes, with in three cases an additional hole at the end, and in one case a hole at each end. The holes are shallow saucers in shape, about an inch and a half in diameter on the average, and a quarter of an inch deep ; many of them are perfect circles in plan and beautifully hollowed out. They appear to have been intended for playing the game called in Arabic

* Mankala/ and termed Olinda Keliya, * the Olinda game' in Ceylon.1

It is strange that a site should have been selected in which the lower holes are on a part of the rock which slopes downward considerably, close to the edge of a precipice. The place is also quite exposed to the sun, and, as I myself experienced, the surface of the rock becomes greatly heated during the daytime, when one would expect the game to be played on such a site. I| would seem to be unnecessary to cut this large number of'holes in such close, proximity, and in lines almost parallel, if they were mferely' intended for a game at-which two persons require only one set of two rows. There are other parts of the rodriftat appear to be much more suitable for playing this game where it would be needless to crowd 1 For a full account of, It, see The Ancient Games.

0226 ANCIENT CEYLON

the holes together. The Sinhalese villagers who accompanied me to the rock, and who were well acquainted with the game, could offer no elucidation of the use for which the holes were intended ; they were unable to understand why any one should desire to play the game on such a site. Yet notwithstanding these reasons for doubt, and partly because of the holes next described I assume that these cup-holes were intended for the Olinda game. I may observe that for some reason which is unexplained it is possible that there may have been something of a sacred character in this game in Ceylon ; it appears to be specially connected in some way with the celebration of the festival of the New Year.

Another site was discovered by Mr. F. Lewis, at that time a member of the Forest Department in Ceylon, at Pallebaedda, in the Sabara-gamuwa Province, and I am indebted to his kindness for my information regarding it. The holes are cut in a rock immediately in front of three caves that were prepared in ancient times for the occupation of Buddhist monks. Mr. Lewis states, 'In front of the wihara cave is a rock of a * hog's back' outline, on the ridge of which are two well-cut [square] holes evidently to receive the wood-work of a shrine, A little to the right of them is what appears to be a sort of cribbage-board, in which there are 18 holes cut in the rock, ending with a crescent-shaped hole. The smaller holes are each about ij inches in diameter, spaced ij to if inches apart/ His sketch shows this to be an unmistakable Olinda board of nine circular cup-holes in each row.

The antiquity of the holes is indicated by the number of them, fourteen being invariably employed at the present day in Ceylon and Southern India. If the holes do not actually date from the period when the caves were being prepared for the monks it is probable that they belong to some time during the next few centuries.

As to the date of the work at the caves, there is definite evidence in the forms of the letters in the dedicator}7 inscriptions cut over them, below the kataras. One inscription copied by Mr. Lewis is, Tapasa D(e)ita$a lene sagasa, and in a second57 Ku.-dci Woe ro-gala. ,i

iff 6

58. Galpitiya - gala ///32

Kolo*wacwa Channel.,/,. 59. Western Group

60 Eastern Qroup.

61 Wei! an go 11 a

,

Plans and Sections of large Cup-holes.

227228 ANCIENT CEYLON

line a word which in his hand-copy is Tusasa. This may be in reality Gutasa, and the whole inscription then would be, ' The Cave of the Ascetic Deva; to the Community; (and) of Gutta (Gupta)/ In similar early letters, in which the rounded form of s does not occur, there is inscribed over another cave, * The Cave of Pusadeva/1 Both inscriptions may date from the second century B.C.

In addition to these sites, I was informed by the Vaeddas of the southern part of the Eastern Province that some small holes are to be seen on a rock called Lenama-gala, six miles from Haelawa.

While there are so few places at which these shallow cups or saucer-shaped holes occur in Ceylon, there are many peculiar and much larger and deeper holes of a different shape, which await some explanation. This at least can be said of them? that there are traces of early monastic buildings in the immediate neighbourhood of nearly the whole of them. It is probable, therefore, that they were cut by workmen who were engaged on the construction of the monasteries. I illustrate a few typical examples in Figs. 57-62 (in which all the sections of the holes are drawn to a scale of two feet to an inch), in addition to the single hole already noted at Sigarata-hena. The north points marked on the plans are only approximately correct.

A group of three holes arranged at the corners of an isosceles triangle with a base of 3 feet 6 inches and sides of 4 feet 10 inches, is cut on a low flat rock to the south of the above-mentioned Frog-rock. They may be of the same age as the OUnda holes in the latter rock. They are aE of one size, being 6 inches wide near the mouth, 6 inches deep, and 2 inches wide at the bottom, which is rounded.

Six holes have been cut in a group in a winding north and south line extending 40 feet 8 inches in length, in a rock called

1 There is nothing to connect this person with the Fhussad?va of the Mahavansa, the great Archer-Chief of Duttha-Gamini .On the other hand, it may be noted that the name is a most uncommon one ;

I have not met with it elsewhere. . There remains a possibility, but nothing more, that the inscription, was cut by orders of this Chief, but in that case one would expect to find him termed Parumaka, * Chief/ROCK CUP-MARKS 229

Kuda Waera-gala, ' the Small Dagaba-rock/ at Wambatuwa-gama, in the North-central Province (Fig. No. 57). Close by, to the south, there is a large rock on which are the remains of a monastery, including a small dagaba at which the bricks have a mean thickness of 2-75 inches, and a breadth of 8-8 inches, Bt. being thus 24-2 inches. The size exactly agrees with that of others found at an old ruined monastery of the immediate neighbourhood, where the length is 15-70 inches and the contents 382 cubic inches. This size indicates the latter part of the first century B.C. or the first century A.D. as the time when the bricks were burnt. As the holes in the rock are of such a character that they must have been made by skilled stone-cutters, it may be taken as certain that they were cut by the men who were engaged in preparing stone for the adjoining monastery.

The holes numbered 3 and 6 may be considered typical of the smaller kind of these holes. Of all the holes which I have examined I believe that No. 5 is the only one with an axis considerably out of the perpendicular. The dimensions in inches are as follows :?

Bottom Width Bottom Width

Top above the Top above the

Number. Width. Curve. Depth. Number. Width. Curve. Depth.

1. 9 2j 9f

2. gj- 2J- I2j

3- 6f 26

4- 9i 2i I0f

5- 9 ' 2i* I'°'i 6. 6J 2 4f

On another rock not far away, called Galpitiya-gala, three holes have been cut (Fig. No. 58). The hole No. 7 is n inches wide at the top, ij inches at the bottom, and 7! inches deep;

No. 8 is io| inches wide at the top, 2 at the bottom and 6

inches deep ; No. 9 is 6 inches wide at the top, i \ at the bottom, and only 2| inches deep.

At a rock in the jungle, on the side of the ancient channel from Kalawaewa (tank) to Anuradhapura, which may date from the end of the third century A.D., there are two groups of holes, which may have been cut by men who were employed on the channel works. No ruins are known near them, it is230 ANCIENT CEYLON

said. The larger group (Fig. No. 59) consists of six holes, of which the dimensions in inches are as follows :?

Bottom Width Bottom Width

Top above the Top above the

Number. Width. Curve. Depth. Number. Width. Curve. Depth.

10.

11. 8 3 6f

12. 8 2f 6

13. 6J (weathered) 3!

14. 8 ( ? ) 5* 15- 8J( ? ) 4*

The smaller group (Fig. No. 60) consists of only two holes in an east and west line, of which one is 9 inches wide at the top, 3 near the bottom, and 6 inches deep; and the other is 6J inches wide at the top, 2§ at the bottom, and 3 Inches deep.

At Wellangolla, in the North-western Province, two holes are cut in a sloping rock over which passes a track leading to some caves that were made over to Buddhist monks, apparently in the second century B.C. An inscription cut over one of them in two lines, in the earliest characters, with the bent | , runs:?(i) Supadu... lene sagasa, (2) Asiya Nagasa gapati Anurudi kulasa ca dine. ' The ' Very Pale ' cave of the Community; given by Asiya Naga, and by the family of the (female) householder Anuruddhi/ Three brick fragments at it average 2-8 inches in thickness, and one is 9 inches in breadth, Bt. being thus 25-2 square inches. If the length was six times the thickness, this being the usual proportion in pre-Christian bricks, it would be 16-8 inches, making the con" tents 423 cubic inches. The dimensions thus point to pre-Christian times, and possibly the second century B.C., as the period when the bricks were made. There is also a small stone flower altar 2 feet 6| inches square, but its age is uncertain.

The holes (Fig. No. 61) lie in a north and south line, which is not parallel either to the adjoining edge of the rock or to the path ; their centres are 4 feet 4 inches apart. Hole No. iSis 6 inches wide at the top, i inch at the bottom, and 2f inches deep; and hole No. 19 measures 6| inches in width at the top, and ends in a point at the bottom ; it is 8J inches deep. Holes ending in an actual point are very rare, and I have examined only one other hole which was of this type.ROCK CUP-MARKS 231

Near the Wellangolla holes a long inscription has been cut on the rock by' the great king Jettha-Tisa, son of the great king Maha-Sena/1 recording grants made to the monks. Jettha-Tissa reigned from 332 to 341 A.D., and the holes may have been cut by the person who chiselled his inscription, if they were not made when the caves were being prepared for the monks.

At Rugama tank, in the Eastern Province, there are three holes in a triangle, cut in the rock at the flood-escape (Fig. No. 62). At hole No. 20 the rock is broken away at the mouth ; below this it is 5 $? inches wide, and it has a total depth of 10 inches. Hole No. 21 is the pointed hole referred to above. It is also worn at the mouth, and is 6 inches wide below this part, and 9 inches deep. Hole No. 22 is 12 inches wide to the outer part of the curve at its mouth, and is 5 inches deep, with a flat bottom, a very unusual feature.

On a rock close to the cave called * Great Beautiful/ in the Eastern Province, at which an inscription was left by the great chieftain of the second century B.C., Nandimitta,2 there is one hole 6 inches wide at the top, and 6J inches deep, with a well-rounded bottom. It may have been cut at the same time as the inscription, or, as there was a monastery near it, at a somewhat later date.

Two holes are cut in a. north and south line below 153 steps chiselled out of the steep sloping face of an immense rock called Tumbulle Waehaera-gala, in the North-central Province. They are cup-shaped, one being 2i inches deep, 4! inches wide at the top/and 2 inches at the bottom, while the other is 3f inches deep and 6 inches wide at the top.

There are monastic ruins on the rock and part of a dedicatory inscription over a cave near its base, in letters probably of the second century B.C., by Sumana Tisagoia,' Sumana of the Tissa clan' (?). Bricks at this cave measure 15*70 inches by 8*80 inches by 2*75 inches, Bt. being 24-2 and the contents 380 cubic inches. They apparently are of a late pre-Christian date, or an early date in the first century after Christ. An inscription by a ?'Tisa Maharaja/ near the cave, belongs to the second or

1 No. 102 of Dr. E. Mullet's Ancient Inscriptions in Ceylon* * See The Earliest Inscriptions, No. 47.232 ANCIENT CEYLON

third century A.D. The steps and the holes may have been cut at the same time as this inscription, if they were not made by the person who prepared the cave for the monks.

At Nagadarana-gama, in the North-central Province, Mr. Bell met with one of these holes surrounded by two concentric circles cut in the rock. His account of it is as follows :? * Here, too, is one of the unexplained incisions in concentric circles not infrequently met with; the outer ring 3 feet in diameter by i inch in depth, the middle 2 feet 3 inches and 3 inches deep, whilst the central hole is cylindrical, i foot in diameter and depth.'* This is evidently of a different type from the holes above described.

In his Annual Report for 1891, p. 7, he mentioned a similar hole, a foot deep, at Tamara-gala. A circle 2 feet in diameter was cut round it, and outside that another 13 feet 6 inches in diameter, ' shallowly cut/

There are many other rocks in the North-central and Northwestern Provinces where holes similar to those I have described are cut, and in some instances single holes are found. The Vaeddas informed me that they have seen groups consisting of seven and even ten holes of this kind on rocks in the southern part of the Eastern Province. I have observed a group of, I think, seven such holes arranged in an extended line on a long low rock in the North-central Province, at the side of a path leading to a wihara. At that time I recorded no particulars of these holes, though I often met with them on rocks in the jungle. In the case of the row of seven holes, I was informed that these were utilised at festivals as lamps for illuminating the path, oil being poured into them on water, and a floating wick fixed on it, resting on four cross-sticks, in the manner often employed for hanging coconut-oil lamps.

However much the holes of this description vary in size they are always exactly circular in cross section, with the upper edges carefully rounded. The sides and bottom are always beautifully smoothed and in some instances almost polished. No chisel mark can be seen on any of them, and it is evident that the smoothness is due to much friction, 1 Arch. Survey, Annual Report for 1892, p, 8.ROCK CUP-MARKS 233

which must have been caused in the case of the deeper holes, if not in all, by turning round inside them a stone or iron implement of a special shape, or by constant rubbing. There is nothing to indicate if the smoothness was originally given to them by the men who cut them, or is due to long use of them for some purpose or other. The latter is the probable cause of it. At each group of holes there is usually at least one small one; but in some groups there are no deep holes.

With the exception of the single group now occasionally employed as lamps I have never met with any villagers who could even suggest any use for the larger holes just described. Their Sinhalese name is kowa, which commonly means * crucible '; and a word like it, but with a slightly different spelling (kowuwa), is used in inscriptions as the name of the stone flooring slabs laid round the dagabas at Anuradhapura. If the modern name indicates the employment of some of them as mortars for preparing medicines, this apparently can only apply to the shallower ones, and the mode of utilising the others is left unexplained, as well as their excellent finish and perfect shapes.

Although the holes are usually close to the sites of temples, I know of no purpose for which they could be required in connection with the services at them. I have been told by Vaeddas that they sometimes present offerings of food to demons in hoEows on rocks, and such a use might account for some saucer-shaped holes, especially those in other countries ; but it will not explain the reason for cutting the deeper holes.

Holes are sometimes cut in boulders or rocks near temples for use in pounding paddy (rice in the skin) ; but they are wider and of a different shape from these, being always cylindrical. Others in which money or valuables have been concealed are also cylinders, with an offset round the top into which the covering slab was inserted and cemented, earth being then sprinkled over the spot. Sockets for holding wooden or other posts are also cylindrical, or square in section.

It seems possible that some of the deeper holes may have been cut for expressing oil by hand labour for temple illumination, as an act of merit for the piously disposed, in the234 ANCIENT CEYLON

manner now practised for extracting coconut oil by a ' chekku/ that is, by means of a pair of bulls which turn a loaded wooden pestle in a large wooden mortar. The pieces of coconut are compressed between the pestle and the side of the hole or mortar, and the oil is gradually squeezed out of them. It is a practice of great antiquity, and notwithstanding its primitive appearance is said on the best authority to be as effective at the latest European machinery. Even this explanation does not satisfactorily account for some of the groups of holes, and especially for the rounding of their edges, although, on the whole, it is the best one that I can offer.VIII THE LOST CITIES OF CEYLON

TAMBAPANNI AND WIJITA

TN the account which has been given of the aborigines of 1 Ceylon, I have endeavoured to show that at the time when the written history of the country begins they held only the southern two-thirds of the island. The first capital of the Gangetic ancestors of the Sinhalese was establishedin some part of this district, and was believed by the early annalists to be in the neighbourhood of one of the coast settlements of the aborigines, as the extracts which record the fate of their princess Kuweni render quite clear. It was near to settlement or town, Sirivatthapura, that Wijaya and his followers were understood to have landed. It follows that the early writers were aware that the Sinhalese capital was close to the place of their debarkation, at a spot where the natives had an opportunity of collecting treasure-trove in the form ol the cargoes of vessels that were wrecked on the adjoining coast. This fact, about which there can be no doubt throws aside all the inland sites that are connected with the story by modern tradition. Such tales are not of the slightest value when compared with the written beliefs of the pre-Christian chroniclers from whichthe historians gathered their information Many different places have been selected by European aid later Sinhalese writers as the site of the first capital, but the early annalists appear to have had no doubt regarding its position. It was then known as the city of Tambapanni an early name of the island itself, apparently borrowed from Southern India, where there is a river of this name. The JJipa-vansa says (p. 162) of this place, 'Tambapanni was the first [Sinhalese] town in the most excellent Lankadrpa [Island of236 ANCIENT CEYLON

Ceylon]; there Vijaya resided and governed his kingdom. . . . The town of Tambapanni surrounded by suburbs was built by Vijaya in the south, on the most lovely bank of the river.' l The history being written at Anuradhapura, all sites to the north, east, and west of that city are at once excluded by this sentence.

In this story the tale about Kuweni is altogether omitted, but unless the new-comers had formed an alliance with some of the natives it is difficult to comprehend how they could acquire the supremacy over more than a small part of the country. What probably occurred was that for a long period antecedent to the appointment of a Gangetic prince as ruler, the Magadhese merchants had been accustomed to visit the island in ships that sailed direct from the mouth of the Ganges, or perhaps called at other trading stations on the way. At last an adventurous member of one of the northern royal families accompanied a party of these merchants to Ceylon, and by allying himself with some of the natives succeeded in acquiring the general sovereignty of the island in the districts where the influence and power of the traders were sufficiently extensive. Other parts of Ceylon probably retained their own rulers in a state of complete independence until at least the time of Pandukabhaya, the fifth sovereign, who by his wise policy of conciliating the native chiefs succeeded in inducing all to accept his control.

The annalists state that the first Gangetic prince (who is mentioned only as Wijaya, * The Conqueror') married a Pandiyan princess of the southern Madura, by whom he had no children. Shortly before his death he despatched messengers to his father's capital, Sihapura,' the Lion City/ in the Ganges valley, to request another prince of his own family to come to Ceylon in order to succeed Mm. His nephew, called merely Panduwasa Deva, 'The Deity or King of the Pale Race/ son of his elder brother, Sumitta, who had succeeded to the throne on the death of his father, accompanied the ambassadors to Ceylon, and became its second sovereign. One of the late king's ministers, called Upatissa, faithfully managed

1 Translation by Dr. H. Oldenberg, p. 162.THE LOST CITIES 237

the government during the year's interregnum that followed the death of Wijaya, early in the fourth century B.C.

It is the descriptions of the journey of this prince to Ceylon, and that of the princess who followed him afterwards in order to become his queen, which afford definite information regarding the place in the south of Ceylon at which the first capital was founded.

According to the Mahavansa, Wijaya died soon after despatching the ambassadors to SHiapura, and the Regent had settled down at another early town, called Upatissa, to the north of Anuradhapura, and on the bank of the Malwatta-oya, then known as the Kadamba river (Mali, i, p. 34). The returning members of the mission could not be aware of these facts, and evidently landed at the usual port near the old capital. The Mahavansa states (p. 36) that they arrived at the mouth of the Maha Kandara river, and Mr. Tumour has added, apparently from the Tika or Commentary, * at G5nagamaka tittha/ the ford or landing-place of Gonagama.

Following this prince, there arrived the princess who became his queen, who also landed at the same port of G5nagama, whence she also proceeded to Upatissa, the new capital. The Mah. says (i, p. 36), 'The ministers having already consulted the fortune-teller Kalavela, and having waited on the females who had arrived at Wijita [on their way to Upatissa] in fulfilment of that prediction, having also made enquiries there regarding them and identified them, presented them to the king at Upatissa/

Where was the town Wijita, to which these ministers proceeded from Upatissa, a city north of Anuradhapura, in order to meet the distinguished traveller from Gdnagama ? It has been long believed that it was at Kala-waewa, in the North-central Province, where a small Buddhist temple, called Wijita-pura wihara, exists to the present day. I have examined this place, and failed to find signs of any early works of importance. The best evidence, the dimensions of the bricks, is uncertain. Those accessible in the dagaba at the wihara are all more or less in pieces, and are of two sizes, averaging 271 inches in thickness, which it is possible may be pre-Christian,238 ANCIENT CEYLON

and 2 TO inches. There are also some worn fragments of inscriptions of the fifth or sixth century A.D., cut on the steps leading to the temple enclosure. Nothing but this monastery is locally known to have been constructed at this spot.

In the story of the re-conquest of northern Ceylon from South-Indian invaders by King Duttha-Gamini before 161 B.C., there is a long and fanciful account of his capture of a very strong fort at Wijitapura, with triple fortifications, the strongest fortress in the country next to Anuradhapura, which was at that time the capital; but no such place is known anywhere near Kalawaewa. The account of this campaign is fully related in the Mahavansa (i, p. 96 ff.). Duttha-Gamini, marching from Magama or Tissa in the extreme south-east of Ceylon, began it by capturing the town of Mahiyangana, an early settlement on the eastern side of the Kandian mountains ; after which he gradually made himself master of a chain of forts established by the invaders along the banks of the Maha-waeli-ganga. The history then states (p. 97) ' All those Damilas [Tamils] who had escaped the slaughter along the bank of the river threw themselves for protection into the fortified town called Wijita/ It is clear, therefore, that this town was not far from the lower section of the Mahawaeli-ganga; and, as we know from the journey of Panduwasa Deva's bride, was on a public road leading direct from the port of Gonagama to the northern capital. By holding it the Indian troops evidently hoped to check Duttha-Gamini in Ms victorious march on Anuradhapura.

A later historian who described the extensive works of King Parakrama-Bahu I (1164-1197 A.D.) at Polannaruwa, his capital, relates (Mah. ii, p. 201) how he formed three suburbs of the city :?* Afterwards the king caused three smaller cities to be erected, namely, the Rajavesi Bhujanga, the Raja Kulantaka [also called Sihapura on p. 259] and Wijita/ It then states that in the space between the palace and these three towns he built three wiharas, thus indicating that they were not far from the capital. At p. 260 reference is again made to ' the branch city, Wijita/ It is a constant habit of the later historians to use the word meaning to * construct'THE LOST CITIES 239

when the actual work done is a repair or re-construction ; and whether it was the case in this instance or not, it is at least proved by these records that close to Polannaruwa there was a Wijitapura in the twelfth century. Can it be the celebrated fortified city captured by Duttha-Gamini ?

When that king had taken it he next marched on a post termed Girilaka, the station of a chief called after it 1 Giriya, * the Giri person/ This may have been the place eight miles north of Polannaruwa now known as Giri-talawa, on the present road to Anuradhapura. The meaning of the name, ' Giri plain/ shows that it may be derived from the Girl village, where the chief Giriya lived.

From there the king proceeded to Mahela, which may be the village now termed Maha Aela-gamuwa, on the road from Dambulla to Anuradhapura. With these very probable identifications to confirm the line of Duttha-Gamini's march, I feel justified in assuming that the fort of Wijita which he captured was close to Polannaruwa, and possibly either an early name of that city itself, or a place at the site of Parakrama-Bahu's ' branch city/ It cannot have been a town on the north-western side of Kala-waewa, at the site of the Wijitapura wihara, which is completely out of the line of march to Anuradhapura from any point on the lower course of the Mahawaeli-ganga, and is also too far from that river to be a rallying-ground for troops who were blocking the king's advance on the capital.

We now return to the journey of Panduwasa Deva's bride from the coast to Upatissa. If Wijitapura where the king's ministers met her was near Polannaruwa we see at once that the meeting-place was nearly half-way on the great highway which passed from Magama to Anuradhapura and Upatissa, through Guttahala (now Buttala), and across the Mahawaeli-ganga at Dastota. This old highway, part of which is now called Kalu-gal baemma, * Black-Stone Embankment/ is still in existence, but overgrown with forest; and it is said that it can be traced from Buttala to the river. Where I examined it

1 ' Each village gave its name to the Damila chief in charge of it.* Mali, i, p. 97.240 ANCIENT CEYLON

near Nilgala, and at the present high road to Batticaloa it is well defined. Near Nilgala it runs on an earthen embankment which is about twenty feet high near some stream-crossings ; it is five feet high at the path from Nilgala, and one hundred feet wide at the base. The top of this bank appears to have been thirty or forty feet broad, or even more.

It is clear that if the true site of Wijita-purais even approximately fixed by me, the landing-place from which it was necessary to pass through it in order to arrive at the capital cannot have been at any point on the western coast, or even on the central part of the eastern coast. We are therefore reduced to southern and south-eastern Ceylon in which to find the port where the princess disembarked, the same Gonagama at which Panduwasa Deva landed.

The name of the river, Maha Kandara, is of little use in the quest without further cprroboration of its position, there being several Kandura streams in Ceylon. But the name of the port itself may now be utilised. Where is there a Gonagama landing-place in southern or south-eastern Ceylon ? This query is easily answered. Four miles inland from the mouth of the Kirindi river which runs past Magama or Tissa there is a natural pool still termed Gonagama-wila, ' the Gonagama pool/ I suggest that, taken with the other evidence, it proves that the mouth of this river was the landing-place in question. If so, the Maha Kandara river is the present Kirindi river, the old name of which in the Mahavansa was Karinda. The ' Sambar village' which gave its name to the pool, but may have been nearer the mouth of the river, has long since disappeared.

This, then, is the place, unlikely as it may seem at the extreme south-east of Ceylon, at which the two Indian travellers, one from the same country as Wijaya, and the other from its immediate neighbourhood, landed in Ceylon. It is a fair inference that this was the usual route of the early Gangetic traders, and that the journey of Wijaya was believed to have followed the same course.

That such is the case is confirmed by the Rajavaliya which says (p. 20), in unmistakable terms, that Pancjuwasa DevaTHE LOST CITIES 241

landed ' at the haven of Tammanna/ the Sinhalese name of" Tambapanni city. Thus Gonagama was the port or haven of Tambapanni.

It was at the same spot that Wijaya and his men landed :? ' When the ship made for land in the direction of Ruhuna [southern Ceylon], they saw the rock Samanta Kuta [Adam's Peak] while at the sea, and concluded among themselves that it was a good country to live in. Having seen the sea-coast they landed at Tammanna-tota, and rested beneath a Banyan tree' (Raj., p. 16). From the sea or the coast near Kirindi,. Adam's Peak is clearly visible in fine weather, as the writer" from whom this account was taken by the historian evidently was aware, or he would not have specially mentioned the factr which itself excludes every site on the eastern coast.

Having once found this landing-place of Wijaya, Panduwasa Deva, and the latter's queen, the fact that the first capital was so close to it that it was termed * the port for Tammanna' leaves no room to doubt that the later Magama, now called Tissa, at the side of the Kirindi river, and only six miles from its mouth, was the spot selected by the settlers as the first seat of government. All the early settlements of the leading chiefs are termed gama, ' village,' in the Mahavansa, and the capital became the Maha-gama, ' the Great Village' of the country. The appellation still survives as the name of a small village, Magama, on the bank of the river, between Tissa and its mouth.

The city was established along the higher ground on the left bank of the Kirindi-oya. In a slight hollow to the left of this again the Tissa reservoir or ' tank ' was made for supplying the place with water. Tissa appears to have been the name of a suburb on the eastern shore of the reservoir, where an inscription of about the second century A.D., cut on a pillar to record the suppression of a heresy, refers to it as Asatisa* rajakaya gdma. * Asatissa, the royal village/

It is not my intention to give a description of the present-state of the early cities. I shall be satisfied if I can succeed in identifying the sites of some of them, and thus clearing up certain difficulties in the early topography of the island.

?R242 ANCIENT CEYLON

Wijitapura is described as follows in the second century B.C., at the time of Duttha-Gamini's war: 'The fortress of Wijitapura was in this wise. It was girt about with three moats filled with water. Around it was a rampart of bronze closed by a gate of eighteen cubits. Amongst the fortresses reduced there was none like unto this. Except the city of Anuradhapura none of the other fortresses equalled it.* (Raj.,

P. 38).

According to the Mahavansa it was founded by a chief called Wijita, who accompanied Prince Wijaya to Ceylon, and it was then an ' extensive settlement' (Mah., i; p. 34). Panduwasa Deva subsequently removed there from Upatissa-nuwara, and made it his residence in the early part of his reign (p. 37), and the brother of his queen also lived at it, probably as ' Governor/ like other princes mentioned in the histories. Thus it was evidently one of the most important towns in the country at this time. It was then abandoned by the sovereign in favour of Upatissa, and it does not re-appear in history until the war of Duttha-Gamini; nor after he captured it is it again mentioned until the twelfth century A.D.

The measurements of some of the bricks still to be found at Polannaruwa prove that buildings of pre-Christian date existed there or in its immediate neighbourhood; but beyond this meagre evidence which they have preserved nothing further is known of the early settlement at this town on the greatest highway in the kingdom. Its position on this route was too commanding, however, for it to be totally given up ; and in all probability the new city, Polannaruwa, merely supplanted the old one/

UPATISSA

In the MahUvansa (i, p. 34) it is stated that the chiefs under Wijaya settled down at important stations throughout the country. ' Thereafter the followers of the prince formed an

establishment, each for himself, all over Sihala 1 [Ceylon].

1 The Vaeddas still use this expression to designate the districts occupied by Sinhalese. The Sinhalese expression for the island is Lanka or Lankawa,THE LOST CITIES 243

On the bank of the Kadamba river [the Malwatta-oya], the celebrated village called after one of his followers Anuradha. To the north thereof, near that deep river, was the village of the brahmanical Upatissa, called Upatissa. Then the extensive settlements of Uruwela and Wijita, each subsequently a city. Thus these followers, having formed many settlements, giving to them their own names, thereafter having held a consultation, solicited their ruler to assume the office of sovereign/

Of these towns, neither Upatissa nor Uruwela has been identified. Upatissa is described in the Dipavansa (p. 162) in eulogistic terms :?' Upatissa founded Upatissa nagara, which had well-arranged markets, which was prosperous, opulent, large, charming, and lovely/ It ought to be discovered when the ruins along the course of the Malwatta-oya have been completely explored.

A highway formerly ran northward from Anuradhapura through what is now the Northern Province, the ancient Nagadipa. It crossed the Malwatta-oya by a bridge formed at the ends by stone posts fixed in rocks in the bed of the river, a few of them being still visible at the banks, according to information given to me by villagers. It passed immediately below the embankment of a large and very early reservoir, now calledPavat-kulam, the original name of which is unknown. Across the water which escaped over the waste weir or flood escape, the road was carried by means of another bridge consisting of stone beams, laid on stone posts, part of it still remaining at the spot (Fig. No. 125). It is extremely probable that this great highway, a continuation of that from Tissa, was carried through or close past Upatissa-nuwara, the only large town which is described as being north of Anuradhapura. I believe, however, that no ruins likely to be the remains of such a city have been found as yet.

After a lapse of more than two thousand years the ground occupied by the early houses will doubtless be covered by an accumulation of soil. At Anuradhapura, the floors of many buildings the majority of which must have belonged to post-Christian times have been buried under two or three feet of soil. Only slight mounds, or the ends of a few broken stone244 ANCIENT CEYLON

posts may be visible at the surface as an indication of the site of what may once have been an extensive town. Even at Tissa, which was an important town down to the twelfth century A.D., nothing but the excavation of an irrigation channel revealed the portion of the city which once must have been thickly covered by the ordinary houses of the populace, now traceable merely by a layer of ashes, and bits of charcoal, and fragments of pottery some three feet in depth, which was entirely hidden under a coating of soil over which a dense growth of thorny jungle had spread. There was no mound of any kind to show that houses had formerly existed at the spot.

When it is considered that all the dwellings, with the exception of those devoted to the Buddhist monks, and perhaps also to royalty, would be made of mere sticks and mud, or, at the best, of wood alone, it is easy to comprehend that all trace of a great city may totally disappear from view in a few centuries, unless some prominent Buddhist ruins attract attention to the site. Still, it is always somewhat surprising to discover how completely these early cities disappear from view, while many insignificant hamlets, with their little mud-walled huts under the shelter of their ancestral trees, are found still occupying the spot on which they were established, in some cases more than two thousand years ago, with the inhabitants doubtless leading nearly the same simple life as their distant forefathers.

At a few miles to the north-west of the great north road^ and three miles south of the Malwatta-oya, there was an extensive and very early monastery at a place now known as Tantiri-malei, a wilderness of rocks about a quarter of a mile across.1 The bricks in the dagaba are 3*23 inches thick and 9*04 inches wide, Bt. being 29-2, and the length either 18 inches, making the contents 525 cubic inches; or, if it was,, six times the thickness, 19*38 inches, which would make the

1 Reference is made to it in Mr. Bell's Annual Report for 1896, pp. 7 and 8. He met with, an inscription which is cut in the rock there, for which I made unsuccessful search on my visit ten years previously ; but he does not state its contents.THE LOST CITIES 245

contents 565 cubic inches. In either case the size points to a time late in the second century or early in the first century B.C., as the date when the bricks were burnt, that being probably the only period when these excessively large ones were made.

Colossal sitting and reposing statues of Buddha cut out of the solid rock, at which the bricks are of pre-Christian date or the first century A.D., and other works, prove its importance ; while fragments of rough pottery which cover the beds of small water-courses near it show that many people lived there for a long period. Yet it seems to be too far from a dry-season water-supply sufficient for the inhabitants of a town, to be the site of a city that was the capital of the country for half a century or more. I incline to the opinion that the city called Upatissa was more likely to be close to the place where the northern road crossed the Malwatta-oya, at a point probably some miles to the north-west of Anura-dhapura, a neighbourhood that I had no opportunity of exploring, though I heard of a low hill near it on which some carvings (statues or reliefs) are to be seen.

The references to the town in the Mahavansa are very meagre. I have already mentioned that it became the station of one of Wijaya's chiefs, probably in about 400 B.C. When that king died, the ministers who carried on the government are said to have made it their headquarters, notwithstanding its great distance from the former capital at Tissa. Panduwasa Deva seems first to have made Wijitapura his capital (Mah., i, p. 37), but in the latter part of his reign he is described as living at Upatissa, where he is represented as having an extensive establishment.

The next king, Abhaya, his eldest son, is stated to have reigned at Upatissa for twenty years (Mah., i, p. 41). He was deposed, and was succeeded by his brother Tissa, who also resided at Upatissa for seventeen years (Mah., i, pp. 42, 44). The following king, Pandukabhaya, transferred the seat of government to Anuradhapura, and little more is known of the old capital; but during the reign of Devanampiya Tissa in the tMrd century B.C. it is mentioned that five hundred youths246 ANCIENT CEYLON

of Upatissa became monks. Thus it is seen that Upatissa continued to be the capital for fifty or sixty years in the fourth century B.C., during which period it may have been the largest city in the country.

URUWELA

The other early city, Uruwela, never became the capital of Ceylon. After the reference to the settlement of Wijaya's chieftain at it, it is next mentioned in connection with Duttha-Gamini, in the second century B.C., as follows :?* To the westward of the capital [Anuradhapura], at the distance of five yojanas, at the Uruwela town, pearls of the size of the Amalaka fruit [Myrobalan], interspersed with coral rose to the shores of the ocean. Some fishermen seeing these, gathering them into one heap [lucky fishermen !], and taking some of the pearls and coral in a dish, and repairing to the king, reported the event to him' (Mah., i, p. 107). Thus the writer of the original annals was aware that Uruwela was on the western coast, close to some pearl-banks, where coral also was found.

The town is mentioned again as a place where King Subha (60-66 A.D.) built a wihara (Mah., i, p. 140). There is no further information respecting Uruwela until the reign of King Para-krama-Bahu I (1164-1197 A.D.), who is stated to have (re)-built two hundred and sixteen tanks that belonged to the Buddhist monks, among which is specially included ' the great tank Uruwela' (Mah., ii, p. 265).

Thus there is not much information in the histories to enable even the approximate position of this town and its ' great tank' to be ascertained, yet when utilised with a knowledge of the country these indications are not quite so vague as they appear at the first glance.

It should be noted that the early writers are rather indefinite in their accounts of the direction in which places lay from Anurldliapura. They generally refer to them as being to the east, west, north or south, and often omit the intermediate points of the compass. When, therefore, they describe Ura-wela as being * to the westward' this may have a rather wide application, and does not necessarily mean due west of theTHE LOST CITIES 247

capital. If this assumption be permissible, we may at once proceed to search for Uruwela near the site of the pearl-banks of the Gulf of Mannar.

In that part of the country there were only two tanks of importance. One is now called Periya-kattu-kulam; its embankment, two miles from Marisi-kattu, a village not far from the ancient Kutirei-malei, or * Horse-hill/ promontory, is extremely low ; and it cannot have held more than a depth of three or four feet of water. A low masonry dam, forty feet thick, was built across the M5daragama-oya, a stream which is usually dry throughout the summer months, in order to divert water into this tank. This appears to be a construction of much later date than the embankment of the reservoir, which would perhaps have been raised had these improvements been completed. No stonework has been found at the embankment, which, however, may have retained enough water to ensure a crop of rice off suitable lands lying near it. Although the tank had a long bank it could hardly be described as a ' great * reservoir.

The other reservoir is the work now known as the Giant's Tank;1 its original name has been lost. This also was an unfinished work until its recent restoration, but a lower embankment may have existed from ancient times, sufficiently high to impound a shallow sheet of water which would cover a great extent of ground, the bed being extremely flat. Even a depth of five feet of water would have spread over 1930 acres. Thus, although the bank was low, the expression ' great J was a suitable one to apply to this work.

A stone dam, 90 feet thick, called in Tamil the * Tekkam/ was built at a later date across the Malwatta-oya, at a point twelve miles away, in order to turn water into this Giant's Tank. At a much earlier period a line of square socket-holes was cut in the rock on which the dam is founded, in the bed of the river, evidently in order to permit strong wooden posts to be inserted into them. These would then form the main supports of a temporary dam which must have crossed the

1 A translation Of Its Tamil name, Sodayan kattu karei, Giant-built Embankment.248

ANCIENT CEYLON

river at the site of the present stone dam. At the distance of a few feet on the down-stream side of each post-socket, a sloping socket-hole was also cut in the rock, to hold the lower end of a sloping strut that would support the post near it.

FIG. 63. The Giant's Tank (before restoration).

All the sockets, excepting a few at the northern end of the structure, have been covered up by the later stone dam, but sufficient remain visible to prove clearly and unmistakably for what purpose they were made.

It is manifest that the sockets were cut long prior to the building of the stone dam, in order to enable a dam of sticks and earth that could be repaired easily, to be made across the river for the purpose of diverting water down the only channel cut from it, which runs directly into the Giant's Tank. We see, therefore, that these first works are of early date, for increasing the water-supply of the reservoir. They may be considered proofs that a shallow tank existed there long before the stone dam was built, and that the later work both at the dam and the reservoir consisted only of an improvement and enlargement of the original scheme.

Some kind of regulator was buUt of brickwork, at the inlet/THE LOST CITIES 249

of the channel, in order to check too great a flow of water down it; and the bricks which still remain at the spot, being of two sizes, may indicate the age of the first wooden dam and of the later stone one. The average size of a good series of the larger bricks is a thickness of 2-51 inches and a breadth of 8-57 inches ; Bt. is thus 21*5 inches. If the length was ?six times the thickness it would be 15-06 inches ; if five times, 12-55 inches, the proportion being almost invariably between these figures in the case of such bricks. The contents would be 324 cubic inches, or 270 cubic inches ; and the dimensions point to some time from the second to the fourth century A.D. as the date when the bricks were burnt. It may be assumed that the Giant's Tank was already in existence before this period; such a long channel would not be opened until It had been found that a better supply of water was necessary. The later and smaller bricks resemble those found at Polan-naruwa in buildings of the twelfth and thirteenth centuries, and this would appear to be the time when the stone dam was built.

We have seen that the Mahavansa records the repair or re-construction of * the great Uruwela tank * by Parakrama-Bahu I. This monarch, or Nissanka-Malla, who reigned a few years later, had a most laudable habit of leaving a record cut on a stone pillar at the larger reservoirs restored by him. There is one at Padawiya tank, another at Panda-waewa in the North-western Province, and there are several at the chain of tanks ^adjoining Topa-waewa, the reservoir at Polan-naruwa. If the so-called Giant's Tank is the ancient Uruwela tank enlarged, we might accordingly hope to find at it a similar record of its repair.

In clearing out the bed of one of the breaches at the tank, the late Mr, N. M. Walker, the engineer who completed the recent restoration of the work, discovered a considerable number of cut stones, evidently brought to the spot for the purpose of building a sluice for passing water out of the reservoir. These had been previously covered up by soil washed over them by floods, and their presence was unsuspected. They were ail stones such as might be found at abandoned250 ANCIENT CEYLON

temples, and some of them had parts of inscriptions on them, two of which recorded grants to wiharas. Another was an octagonal pillar; and on its lower part, which was square and had been fixed in the ground at its former site, was cut the following inscription (see Fig. No. 153), ready to be set up on the embankment when the work was completed, the pillar being then reversed :?

Srimat Sihapure jata Sri Parakrama Bahu nakaritan wiswa lokattha karyyavya paritat mana.

Made for the benefit of the whole world by the prosperous Sri Parakrama-Bahu, born at Sinhapura, minded of what was fit to be done.1

The record is almost a copy of that which was left at Pada-wiya, in which the king gives himself the epithet Srimat, prosperous, which therefore is to be applied to him and not to the town. It is merely placed first in order that, according to an old custom, the record may begin with an auspicious word.2 It is surprising to find that the king records his birth at Sinhapura. It appears to be clear from the statements in the Mahavansa (ii, p. 118) that Parakrama-Bahu I was born at Punkha-gama in southern Ceylon, whereas we find Nissanka-Malla stating in more than one inscription that he himself was born at Sinhapura, in India. As he also in his long Dambulla inscription gives himself only the name * Parakrama-Bahu/ it would appear either that all these records at the great tanks in reality belong to him, or otherwise, as is more probable, that he carried on and completed some of the works begun by his great predecessor, and copied his records when writing this and perhaps other inscriptions. Parakrama-Bahu and not Nissanka-Malla receives all the credit of the works in the histories.3

1 I have followed the words of Mr. Bell's translation of the Padawiya,

inscription so far as the two inscriptions are identical.

a Cicero says in his work on Divination * Our ancestors were persuaded that much virtue resides in certain words, and therefore prefaced. their various enterprises with certain auspicious phrases.5

3 Dr. E. Muller thought * that some of Kissanka-Malla's deeds may have been put on Parakrama-Bahii*s account in the Mahavansa.* (Ancient Inscriptions in Ceylon, p. 19.)THE LOST CITIES 251

It may be concluded that the Giant's Tank was one of the more important irrigation works the improvement of which was at least begun by Parakrama-Bahu I, and that it was in existence for centuries before his time. This alone does not amount to proof that it is ' the great tank Uruwela' ; but as there is no other reservoir in the neighbourhood of the pearl banks which can be accepted as such, it would seem that in the present state of our information the identity must be granted.

If so, the Uruwela city must have been somewhere near the coast in that part of the country, where remains of ancient Buddhist edifices have been found in many places, as well as statues of Buddha. There are some ancient remains also at Mantota (called in Tamil Maka-tottal, and Maka-totam) opposite the southern end of the island of Mannar, including those of a celebrated Tamil temple dedicated to Tirukesvaram, that is, Vishnu ; but this place is generally believed to be the Mahatittha of the historians, {the great landing-place ' of travellers from southern India, although I am not aware that there is anything but the Tamil name to confirm the identification.

I should be inclined, 'however, to look for Uruwela nearer the mouth of the Malwatta-oya, or Aruvi-aru as it is called in that district, where a permanent supply of fresh water would be obtainable easily by means of shallow wells, and where the attraction of the pearl fishery would induce a considerable population to reside. In all probability this was the original reason of the establishment of a town or trading settlement at the place, long before Wijaya's time. Beyond this general idea of the position of Uruwela city we cannot go until the discovery of some suitable remains produces evidence of its actual site.

PARANA NUWARA.

I next come to another city, regarding the early history of which the annals are silent. Unfortunately its original name has been lost; for many centuries it has been -called merely Parana Nuwara, 'the Old City/ Its site is well known ia252 ANCIENT CEYLON

the district around it, but elsewhere even its modern name Is not recognised. It is on the bank of the Daeduru-oya, and about a mile from an ancient reservoir at Batalagoda, near Kurunaegala, which was restored by me in the last decade (see Fig. 134).

At one time it was a very important post for the protection of the frontier districts of the kingdom of Kaelani, or southwestern Ceylon, and perhaps of Ruhuna, or southern Ceylon. The fort established at it agrees more closely with the account of that at Wijita-pura than any other I have seen, being surrounded on three sides by three high earthen banks separated by wide ditches; on the fourth side the steep bank of the river acted as a protection, and only one earthen embankment was raised there.

The extent of the town itself is unknown ; it stretched along the side of the river and over some adjoining ground on the opposite side of a narrow rice-field. It had also several subordinate villages near it in which the various classes of artizans and workpeople whose services were necessary in the city were quartered. In one the smiths and tom-tom beaters lived, in another the washermen, and the same castes still occupy them. At a third a caste of hunters kept the king's hounds. A small wihara and dagaba were on the bank of the river to the south of the fort.

For the water-supply of this town the Batalagoda tank was made in pre-Christian times, according to the evidence of the bricks found at it. It now covers 635 acres, and is about twenty feet deep. Bricks at one of the sluices were 2-83 inches thick and 9-9 inches wide; Bt. is 28, and the contents 476 cubic inches if the length was six times the thickness. The width and thickness closely resemble those of the inner part of the Ruwanwaeli dagaba at Anuradhapura, and therefore the bricks may belong to the second half of the second century BX,

From a high-level sluice at the reservoir water was carried by a channel into the city.

Of the great antiquity of the town there can be no doubt. The bricks found at what is traditionally reported to be theTHE LOST CITIES 253

remains of a wihara are of a size which indicates thatlthey were burnt in the second century B.C. or early part of the first century B.C. Their width is 9-5 inches and thickness 3 inches; Bt. is 28*5, and the contents may be 513 cubic inches ; they resemble those in the dagaba at Ottappuwa in the North-central Province, which an inscription 1 proves to have been in existence before 30 A.D., and which tradition attributes to Devanam-piya Tissa.

Even in the third century A.D. it had lost its first name, and was already ' the Ancient City/ An inscription of this period (see Fig. 153 for facsimile) cut over the entrance to a cave-shelter under a rock at Peddawa, a village six miles away, is as follows : Siddham, Pubaga nakaraka wasike bhojike CulM-taha lene. { Hail! The cave of Culuttha, a headman dwelling at the Ancient City/ Incidentally, we may infer from this inscription that there was already in existence another town termed ' the New City/ that is Alut-Nuwara, at Mahiyangana.

The town is believed locally to have been the seat or capital of the ' Great Scholar* king, Kumara Dhatu-Sena (515-524 A.D.), and it is said to have been here that the incident occurred which led to his self-immolation on the funeral pyre of Ms friend the Sinhalese poet Kalidasa. Another and better-known, but perhaps not better-founded, "tradition places the tragic event at Matara, in the extreme south of Ceylon, an unlikely spot to have been selected in those days for the residence of the king.

The place is first mentioned in the historical works in about 1081 A.D., when the Mahavansa (ii, p. 100), includes it with others of the district, in a list of towns captured from the Solians by a general of King Wijaya-BShu (1065-1120 A.D.), It was then called Badalat-tala. It was here that the ceremony of the investiture of Parakrama-Bahu with the sacred thread was held with great pomp and rejoicing (Mah.» ii, p. 125).

At a little later date the importance of the fort is shown by the story regarding it in the Mahavansa (ii* p. 128 ff.) which relates how Prince Parakrama-Bah.il, who afterwards became 1 See Fig. No. 152.254 ANCIENT CEYLON

the first king of that name, and the most energetic ruler whom the country ever had, first proceeded to this place on his way to attack his cousin King Gaja-Bahu, who reigned at Polannaruwa, from whom he hoped to acquire the sovereignty.

At that time a general, Sankha Senapati, ' a man of great weight and valour, the most powerful general in the kingdom/ was stationed at it by the king of south-western Ceylon, with a body of troops, in order to guard the frontier districts, which then extended up to the Kala-oya. The general received the prince well, but on various pretexts continued to detain him pending the receipt of instructions from his master as to the course to be pursued regarding him. In the end, Parakrama and his men, losing patience, killed him at this fort. We find it mentioned several times afterwards during the desultory fighting of that period.

The last reference to the place is contained in an inscription which was left on a large slab on the embankment of the reservoir, by Queen Kalyanawati (1202-1208 A.D.), the widow of King Nissanka-Malla, in the third year of her reign, that is, 1204 or 1205. In it she recorded her restoration of the tank at ' Badalagoda at Mahala-pura/ the old town, and her (re-) construction of a wihara?now termed Kota-weriya, from its ' short' dagaba, the Kota Waehaera?at an adjoining village, Pannala, as related in the Mahavansa (ii, p. 268),

After this, the history of the old town relapsed into the fatal silence of all the other forgotten sites in the island, the fort was abandoned, and the inhabitants disappeared.

SlRIWABDHANA-NtlWARA

Another city of some interest on account of the prominence given to it by the Right Rev. Dr. Copleston in his work Buddhism (Appendix, p. 487 £L), is Siriwaddhana-nuwara, as to the position of which considerable doubt has existed owing

to the vague statements regarding it in the histories. Dr. Copleston has given a summary of the history of its identification with a village called Nanbambaraya, said to be eight miles from Dambadeniya, in the North-western Province, which was the capital of the kingdom in the thirteenth centuryTHE LOST CITIES 255

A.D. He has expressed his approval of this identification, and has held it up as an example of the critical acumen of the modern Sinhalese students of their country's history. I may add that if their judgment is correct in this case it is almost the solitary instance in which they have cleared up a single doubtful point in the history of Ceylon.

Dr. Copleston has explained how, by a mistaken reading of the manuscript of the Mahavansa, or through defective copies, the learned editors of the Sinhalese edition?not, I think, an independent translation from the Pali language, but an amended edition of an early manuscript?made the distance of Siriwaddhana-nuwara from Dambadeniya attha, "eight/ yojanas, instead of addka, ' half/ a yojana. The translator of the English edition followed the same reading, and made the distance eight yojanas. The author of the Pujavaliya also adhered to this distance.

The statements in the Bishop's summary tend to show that Siriwaddhana-nuwara had been wrongly supposed to be much further from Dambadeniya than was really the case, and that the highway the lavish decorations of which are fully described in the history of the Festival of the Tooth-relic of Buddha (Mah., ii, p. 286), instead of being many rallies in length was in reality a very short one.

In connection with this identification he remarked that the length of the ydjana is twelve miles (p. 488), but this is not in accordance with the latest researches. Several estimates have been made of this distance. At first it was supposed to be sixteen miles ; this was afterwards reduced to twelve miles, as given by Mr. Childers in his Pali Dictionary ; and an estimate by Professor Rhys Davids in his work On the Ancient Coins and Measures of Ceylon, p. 17, made it between seven and eight miles. This, however, depends chiefly on Indian distances. The Mahavansa contains several references to it, some of which may assist in showing what this measure of length was in Ceylon.

When King Duttha-Gamini was about to build the Ruwan-waeli dagaba at Anuradhapura, we are told (p. 106} that some silver was discovered at Ambattha-kola?now called on account256 ANCIENT CEYLON

of it, Ridi-gama, ' the Silver Village/ in the North-western Province?which is stated to be eight ydjanas from Anura-dhapura. The actual distance in a straight line is 55-3 miles. If we add one-tenth on account of the windings of the road we get 60-8 miles, or a length of 7^- miles for a ydjana. ' Uruwela city is also said (p. 107) to be five y5janas from Anuradhapura. The distance in a ^direct line to the mouth of the Malwatta-oya, near which it may have been built, is 45-6 miles. Adding one-tenth again the distance becomes 50-1 miles, which makes the ydjana 10 miles. The sea due west of Anuradhapura, at Ponparippu, is 39-2 miles away; this, with the same addition, would make the ydjana about 81 miles if Uruwela were there, and on the shore, at the point on the coast which lies nearest to the capital. Thus, in this instance we have a maximum of 10 miles and a minimum of 8| miles, as the possible length.

* Pelivapi is stated (p. 107) to be seven yojanas north of Anuradhapura. This tank is the reservoir now called Vavu-nik-kulam, formed by raising an embankment across the valley of the Pali-aru, on which no other tank is known. The river at the breach in the embankment is 51*2 miles from the capital, aM the addition of one-tenth makes the yojana 8 miles. ,. We also learn (p. 106) that going seven yojanas eastward from Anuradhapura takes us into the district across the lower plrt of the Mahawaeli-ganga. Measuring up to any part of the river there the general distance is about the same, that is, 56 miles ; so that when one-tenth is added the yojana becomes in this case a little over 8J miles. Professor Davids adds a little more to allow for the winding of the path ; it would of course increase the length of the yojana slightly if this were done. Although these are only approximate estimates in the Maha-

j-vansa, they agree so closely that the mean length of the yojana found by them may be accepted as being nearly correct when

.applied to similar records of distances in Ceylon. If, as I believe, Uruwela was near the mouth of the Malwatta-oya, the mean length of the yojana becomes 8| miles. This is not

necessarily the actual length of a measured yojana; it is prob-THE LOST CITIES 257

ably the length ascertained by the time occupied in walking" from one place to another.

The identification of Nanbambaraya village as the site of Siriwaddhana-nuwara depends on three statements in the history :?firstly, the distance of the place from Dambadeniya, variously given as half a yojana and eight ydjanas, neither of which agrees with the actual distance of the village from the capital; secondly, the statement that before his accession to the throne Parakrama-Rahu II lived at Nanbambaraya; and thirdly, another statement that his wife, who of course lived there with him, was termed the Siriwaddhana Bisawa (queen).

According to the Mahavansa, he himself was born at Siriwaddhana ; thus the third piece of evidence merely .shows that he married a lady whose native place was the same as-his own. The second statement would be of value only if the traces of some early city, and of the temple to which the Tooth-relic was taken, had been discovered at Nanbambaraya; but regarding this point the Sinhalese scholars furnish no information, although it is one that they could easily investigate. Without this support the whole argument hangs in the air, awaiting the construction of some solid foundation on which it may rest. All is paper evidence of an unconvincing type. Although much has been written to show that in the opinion of the writers Siriwaddhana-nuwara ought to be at Nanbambaraya, there is not a line to prove that it really was there.

The evidence seemed to me so unsatisfactory that I made careful enquiry into the matter from the Kdrala, or chief of that district, who knew the country well, and lived in the neighbourhood. He informed me that there is no local tradition that Siriwa(J Jhana-nuwara was in that part of the country,, or that the Tooth-relic was ever deposited at any place in the district excepting Dambadeniya. He knew of no -traces of any ancient city anywhere round that town. As a matter of fact, he and all others whom I interrogated on the -subject stated that the people of the district had always understood that Siriwa44hana-nuwara was not there, but in the Wanni Hat Pattu, which extends between the Daedura-

s258 ANCIENT CEYLON

oya and the Kala-oya. This might merely point to the ancient town at Yapahu, which was also sometimes termed Siriwad-dhana-miwara, and was the capital for a short period in the thirteenth century A.D.

As a result of other enquiries, I learnt that there is a place at Katuwannawa, a village two miles north of the junction of the Kimbulwana-oya with the Daeduru-oya, which still bears the name of Siriwaddhana-nuwara; and I* took advantage of the first opportunity to visit it.

There is an early wihara at the spot, with a small brick dagaba, called the Sigiriya Waehaera, a raised platform round a Bo-tree, and two small rock-caves prepared for the monks. The only inscription known consists of four letters, of the second or third century A.D., on a flat rock near the dagaba, reading mi simita, with a probable meaning, (this (is) for the boundary/ The bricks of the dagaba are of two sizes, of which those of the earliest type average 2-93 inches in thickness, 9-07 inches in breadth, and are nearly 18 inches in length, a fragment being broken off the most perfect one I could find. Bt. is 26-5, and as the length is evidently, as usual, six times the thickness, or 17-58 inches, the contents becomes 466 cubic inches. These dimensions indicate the third century B.C. as the probable time when the bricks were burnt.

Water was supplied to the place by a cut channel with a bed from 15 to 18 feet wide, which branched off from a main channel that was opened from a stone dam, now breached, built across the Kimbulwana-oya. This main channel was carried on to Talagalla tank, a large reservoir about four miles away, The restoration of these works by Parakrama-Bahu I is mentioned in the Mahavansa (ii, pp. 148 and 265), the site of the dam being there termed Sukara Nijjhara. In the ground all around, the villagers informed me that when digging for cultivation purposes they met with large-sized ancient bricks, the presence of which proves the existence of numerous monastic buildings there at an early date.

There is, however, a general absence of ruins above the surface of the ground, with one notable exception. This is a ruin known as the Dalada Maligawa, the Palace of the Tooth-THE LOST CITIES 259

relic. It was a circular building of a special type, perhaps unique in Ceylon, 40 feet in diameter to the outer sides of the 20 octagonal pillars that supported the roof, each being about I4f inches thick, and standing now j\ feet out of the ground. Four larger square pillars, with sides of 18 inches, are arranged in a square 10 feet 6 inches across, in the centre of the circle. Inside this central chamber there is a stone flower-altar formed of a single well-cut slab 8 feet 7 inches long and 3 feet 7J inches wide, close to which, on the west side, in the middle of the room, is the spot now pointed out as the site occupied by the case or * karanduwa' of the Tooth-relic. A second stone flower-altar 4 feet 101 inches wide, is fixed to the eastward of the inner room, in the outer circular chamber which surrounds it.

According to the local tradition, the building had three stories; all the upper part must have been built of wood, as in practically all other instances in Ceylon, and it has of course disappeared. The whole place was overgrown with jungle, which was partly cleared away to enable me to examine it.

At the time when the great Festival of the Tooth-relic took place the king is said to have restored the present wihara and the dagaba. The villagers expressed surprise that doubts had been cast upon the identity of the town.

According to the Wanni Kadayin Pota, * the book of the Wanni (district) Boundaries/ the limits of that part of the Visideka Kdrale of the Wanni Hat Pattu, in which the town lay, was defined as follows in the fifteenth century :?f Having first taken the Daeduru-oya up to SrI-war4 Jhana-nuwara, the boundary was made as follows: On this side of the rocky ridge at Ratmala ; the Degadatura mountain; Potuwe-pitiya; Moragoda hill; Gurugocja wihara; the wihara of Niyandawana were made the boundaries. This additional country is the end of the boundaries for the Visideka/ *

This extract proves that the city was close to the Daeduru-oya, and at the edge of the district; that is, at the site just described at Katuwannawa. It is evident that it does not

1 See also Upham's Tracts, p. 215, where the translation

is defective.260 ANCIENT CEYLON

apply to Yapahu-nuwara, which is neither near the river nor is the boundary of the district.

The road running from the city in the direction of Damba-deniya is said to have crossed the Daedura-oya by means of a bridge on wooden posts set in sockets cut in the rock in the bed of the river. That capital was 24^ miles away in a direct line; and if, as before, one-tenth be added, we get about 27 miles as the probable length of the path to it?or a little over three yqjanas.

In any case, it is difficult to see how the distance given in the history is to be reconciled with the facts ; but some of the other measurements supplied by later writers are also widely wrong. For instance, in the Mahavansa (ii, p. 309) it is stated that Polannaruwa is five yojanas from Dambadeniya, while the distance in a straight line is about 71 miles, or by the present road, which is very devious, 86 miles. The relics wer^ taken there in a procession like that to Siriwaddhana-nuwara, during the reign of the same king.

Possibly the word yojana was written instead of gawuwa, which commonly means about four miles. If the road was more devious than usual its length would be a little less than eight gawuwas. It is to be noted that in the translation of the Mahavansa published by Upham in 1833, the distance of Siri-waddhana-nuwara from Dambadeniya is not mentioned; apparently it was not in his manuscript.

The Mahavansa relates (ii, p. 288) how, after causing the road to be levelled ' like the face of a drum J and covered with sand, * the king, followed by the sound of the five instruments of music, and forming a procession of great magnificence, carried the relics [the Tooth-relic and the Alms* Bowl of Buddha] by stages along the decorated highway into the city of Siri-wacjdhana, and placed them on the seat that was prepared for Buddha in the spacious ornamented hall that was built in the middle of the wihara.' The chief quality of the music was. its louclness; it is described as being ' like a blast proceeding from the sea of his merits, which sufficed to drown the roar of the ocean and put to shame the thunder of the clouds/IX THE EARLIEST DAGABAS

WHEN the ancient Egyptian desired to give the earthen mound or tumulus that was raised over the dead a form that would permanently guard his remains he designed a four-sided pyramid of stone or brick. In the East the structure took the form of a solid dome of stone or brick, called in Ceylon a Waehaem, Saeya, Dagaba (relic-chamber), Thupa, or Cetiya, and in India Caitya or Sthupa (tope). In Ceylon two of the intermediate stages between the plain earthen mound and the solid stone or brick structure have survived, one being in the form of an earthen mound enclosed in a hemispherical shell of brickwork, and the other being a wide cone of brick. Both these forms are comparatively rare.

Whether the people of the East borrowed the idea of the dome-shaped building from the Phoenicians it is impossible to say; there is at least a great probability that they did so, since before such dagabas or sthupas were constructed in India and Ceylon Phoenician tombs were already in existence of a nearly similar design, consisting of a segment of a hemisphere resting on vertical-sided cylinders of larger diameter- As they borrowed the alphabet from the Semites they might equally adopt this form of durable tomb, seeing that many other * motives ' in the art of the East are derived from those of the Euphrates valley and Phoenicia. In Ceylon, at all events, the majority of the details used in early decorative art can be traced to those countries. That such copying of the shape of the tomb took place is rendered the more probable by the fact

that in Ceylon the dome, in all the types of the dagaba, was

aei ?262 ANCIENT CEYLON

almost invariably raised from the ground on one or more basal cylinders, as in Phoenicia. It was from India, in the third century B.C., that the idea of the dagaba was first directly borrowed in Ceylon, and the earliest ones of which we have any record were constructed during the reign of the famous Indian Emperor Asoka.

Having once adopted this type of relic-tomb the constructive and artistic genius of the Sinhalese race proceeded in the following century to develop the design to an extent not found elsewhere. The most important examples erected in Ceylon are comparable with the greatest pyramids of Egypt. By some persons this comparison is looked upon as inappropriate, but as a matter of fact the two largest dagabas at Anura-dhapura surpass in contents, and three dagabas exceeded in height, all but the two enormous pyramids of Khufu and Khafra, at. Gizeh.

The minor structures of this class are found throughout the whole country, and must have eventually amounted to thousands. The present account deals only with the earliest works which can be identified, regarding some of which no measurements are yet available.

THE ANURADHAPURA DAGABAS.

Putting aside the mythical story of the building of a small dagaba at Mahiyangana, in Eastern Ceylon, during the lifetime of the last Buddha, in order to enshrine a handful of his hair, the first historical notice of the erection of this kind of relic-tomb in Ceylon belongs to the reign of Devanam-piya Tissa (245 ? B.C.), who is recorded to have built two, the ThupSrama Dagaba and the Pathama Cetiya, at his capital, Anuradhapura, and apparently one at Mihintale, a rocky Mil eight miles away, besides other unnamed small ones elsewhere. The first and last of these three are still in existence, but the Pathama Cetiya has not been found, and therefore it cannot have been a large building. Of the two which are known,THE EARLIEST DAGABAS

263

the first to be erected was the Thuparama dagaba, in about 244 B.C. The others must have been built within the next ten or fifteen years.

FIG. 64. The Thuparama Dagaba, 1873.

THE THUPARAMA DAGABA

The Thuparama dagaba was formed in order to enshrine two relics of Buddha, his right collar-bone (dakkhinakkhaka) and the plate off which he was accustomed to eat Ms food. Its original shape is not recorded; but at the early date at which it was constructed it is unlikely to have differed from that of the dagaba built in the same reign at Mihintale, which is a hemisphere resting upon three very short wider cylinders that form basal ledges round it. Like it, the Thuparama dagaba would have 'a square block of brickwork, now termed a * tee/ an expression borrowed from the Burmese, on the top of the dome, and a spire rising out of a short cylinder set on this. Unlike other works of the same character, it is not stated to have been provided with a terminal member in264 ANCIENT CEYLON

the shape of a ' chatta/ or solid umbrella, on the summit of the spire.

Around its base was formed a circular paved court-yard 164 feet 6 inches in diameter,1 raised n feet 4 inches above the adjoining ground, the ascent to this being made by two sets of stone steps on the east and west sides, each consisting of two flights. This enclosure is supported by a brick retaining wall, which has evidently been reconstructed since its erection, and in which bricks of the earliest type are not found.2 Extremely graceful slender stone pillars with ornamental capitals, but no bases, were fixed in the court-yard in four concentric circles round the dagaba.

It is recorded that various later kings, by way of showing their piety, caused costly decorated network coverings to be placed on the dome. It is uncertain if a roof was ever built over the dagaba, nor is there any actual record of such a construction, although artists of the eighteenth century, if not earlier ones, have represented one in their wall-paintings in various wiharas. This must remain a doubtful point, as it is mentioned in the histories that two other dagabas at Anuradhapura, of nearly the same size, were sheltered by roofs erected over them, as well as a few dagabas in other parts of the island. A work containing relics of sttch importance as those deposited in the Thuparama* dagaba would be likely to receive the same protection.

The chamber in which the relics of Buddha were placed was formed in the upper part of the dome, and according to the account of it appears to have been a small one. No description of its original internal arrangement or decorations has been preserved.

1 For almost all the dimensions of the Anuradhapura dagabas I am indebted to Mr. J. G. Smither's valuable work on them entitled Architectural Remains, Anuradhapura* It was prepared by order of the

Ceylon Government, Mr. Smither being then the Government Architect.

2 The size of the larger bricks appears to belong to a late date in the first century B.C, The wall must have been completely rebuilt by one of the first two Parakrama-Bahus, as there are 2-inch bricks in the mouldings at its base. It has half octagonal pilasters, yj Inches wide and 8 feet 9 inches apart.THE EARLIEST DAGABAS 265

A small room for containing other relics was also built on the southern side of the dagaba in the paved court-yard. It was looked upon as a building of extreme importance, and in the reign of Dappula III (827-843 A.D.) we are told that his ' General named Vajira who was a man large at heart , . . covered the Thupa house at the Thuparama with tiles of gold as became it, and fixed doors also of gold in the house * (Mah., ii, p. 61). Mahinda IV (975-991 A.D.) made a door of gold for it' like the Mount Sineru shining with the rays of the sun' (Mah., ii, p. 87).

King Lajji-Tissa (119-109 B.C.) is stated to have c enclosed the cetiya in a superb case of stone' (Mah., i, p. 128). If this was a course of cut stone which covered the whole dome no trace of it remains. A golden pinnacle was fixed on the spire by King Upatissa II (370-412 A.D.), the dagaba being despoiled of it by Dathopa-Tissa I (640-652 A.D.).

In the time of Aggabodhi II (598-608 A.D.) a large section -of the structure slipped down, exposing the relic-chamber, in which the relics were found lying undisturbed. They were replaced when the repairs were made by this king. Of the relic-chamber it is said (Mah., ii, p. 21), * he arranged four images throughout the relic-room, also a throne made of solid stone, and a golden canopy, and other works of art inlaid with stone and ivory/ The room as rebuilt appears to have been one of considerable size.

During the reign of Aggabddhi III (624-640 A.D.) it is related (Mah., ii., p. 31) that this dagaba was rifled by the sub-king Kassapa of the invaluable relics and gems placed in it in the time of Devanampiya Tissa, and was completely demolished ; but doubtless this refers only to the upper part of the dome, where the relic-room was made. It was restored probably to its original form during the same king's reign, at a cost of only 1,000 pieces of money, an amount which shows that the damage was partial only (Mah., ii, p. 31) ; and a pinnacle studded with gems was fixed on the top of the spire by' Kassapa after he succeeded to the throne, and found it advisable to conciliate the influential Community of Monks.266 ANCIENT CEYLON

Mahinda III (787-807 A.D.) made for this dagaba a cover of gold and ornamented it with bands of silver. These were carried off by Pandiyan invaders from Madura, in the reign of Sena I (846-866 A.D.). His nephew Sena II (866-901 A.D.) invaded Southern India, and sacked Madura in revenge for this and other spoliations (Mah., ii, p. 69).

Udaya I (901-912 A.D.) 'covered the Thupa at the Thuparama with a band of gold/ and Mahinda IV (975-991 A.D.) also fixed bands of gold and silver on the dome.

It was broken into during the domination of the Tamil invaders in the eleventh century, and was surrounded with jungle when Parakrama-Bahu I (1164-1197 A.D.) undertook its repair.

During the reign of the Kalinga conqueror Magha (1215-1236 A.D.), the dagabas throughout the whole country were ransacked for treasure, and that at the Thuparama was certainly one of the first to suffer, but it was restored again in the reign of King Parakrama-Bahu II (1240-

1275 A.D.).

In the first half of last century the illustration given by Major Forbes1 shows it as nearly flat on the top, which was covered with brushwood.; it was considerably narrower below. An earlier drawing belonging to the time of Kirtti-Sri (1747-1780) on the wall of the Dambulla cave wihara represents it as being of the ordinary bell-shape, and without a ' chatta/ or umbrella, on the top of the spire, the general idea being perhaps copied, as the monks at the temple state, from an earlier illustration of it there, done in the reign of Nissanka-Malla (1198-1207 A.D.).

It was finally restored in the form of a bell-shaped structure of very graceful proportions. The diameter at the springing of the dome of the bell is 31 feet, and at the base 40 feet 6 inches, the latter being probably nearly its original measurement. The height to the top of the spire is 55 feet 6 inches.2

1 Eleven Years in Ceylon, Vol. I, p. 226. It was In the same state when Sir Emerson Tennent visited the town in 1848; It must have been restored soon afterwards.

2 Smither. Architectural Remains, p. 3.THE EARLIEST DAGABAS

267

The arrangement of the pillars,* all being of gneiss, which surround the dagaba is stated by Mr. Smither to have been as follows. In the inner circle there were 52 pillars, each, like those in the next two circles, being 12 inches square in the lower part and octagonal in the upper part; they are 22 feet 10 inches high to the tops of the capitals, which have long tenons projecting. In the second circle there were 36 pillars, 21 feet 3 inches high, also with tenons on the capitals ; in the third circle 40 pillars, 19 feet 9 inches high, with a round boss

3OOOOOOOOOOC

FIGS. 65-69. The Thupaxama Dagaba. ^

in place of a tenon ; and in the outer circle 48 octagonal pillars, 14 feet high, with a similar boss. The shafts are all monoliths, and they and the capitals are admirably cut. The

histories do not record their erection; doubtless they are of considerably later date than the body of the dagaba, and their general resemblance to those fixed round the Ambatthala dagaba, described below, although some of the details are of an older type, may indicate that they belong more nearly to the period when the latter were cut (which was possibly early in268 ANCIENT CEYLON

the first century A.D.), or* to some time approaching that date, say, the first century B.C.

The illustrations (Figs. 65-69) show the outline of the dagaba as now restored,1 as well as the form of the capitals and the decorations of these beautiful pillars. The dwarfs carved on them are repeated on the outermost pillars ; on the others their place is taken by horned lions, sitting upright on their haunches and facing front, with their fore-paws raised to the level of their faces, as though about to spring forward, and by standing crested birds with elevated wings, also facing outwards.

In his History of Indian and Eastern Architecture, p. 194, Fergusson stated his confident opinion that ' it can hardly be doubted that these [pillars] represent, and take the place of, the rail of the northern [that is, Indian] topes, and subserve the same purpose, but in what manner is not at first sight very apparent. Referring, however, to what was said above, about the Ceylonese preferring painting to sculpture, it does not seem difficult to explain the anomaly. These pillars were originally, I fancy, connected with one another by beams of wood on their capitals, and from these, frames or curtains may have been suspended covered with the paintings which are so indispensable a part of Buddhist decoration/ In this view Mr. Smither concurred.2

Notwithstanding the high authority in favour of this explanation, I venture to express my inability to accept this theory. It does not account for the absence of tenons from the tops of the pillars of the two outer circles. Mr. Smither believed that the frames were hung only at the two inner circles of pillars ; this still leaves the outer circles without any apparent function, and the tenons of the inner pillars, some of which are 8| inches long, are much larger than such a purpose would require.

It is evident, also, that the meaning of the * Buddhist

1 Reduced from Mr. Smither* s drawing, by the kind permission ?of the Secretary of State for the Colonies,

2 Architectural p, 5,THE EARLIEST DAGABAS 269

railing' has been completely misapprehended. The railing forms a magical protection against evil spirits?the magic circle or square?for the relics enclosed within it; and the three rails usually found in it most probably typify the three protecting * Refuges ' of Buddhism?the Buddha, the Law, and the Community of Monks. That this is the chief if not the only function of the railing is proved by the stationing Nagas and Yakshas as guards at the entrances in it at Bharhut, in India ; they were not there to keep away the human beings for whose use the openings were made, but to forbid the approach of evil spirits, whom they alone could detect and stop,1 just as Nagas (Fig. 8) guard the great dagabas of Anura-dhapura, and Rakshasas act as protectors at the Gopuras of Southern India (see Fig. 4).

Thus the principal member is the railing itself; the uprights, however much they may be decorated, are merely secondary, as its supporters. It is therefore impossible that a series of slender pillars can fulfil its function and take its place.

At Maederigiriya, five miles south-east of Kawdulu tank, the late Mr. levers, when Government Agent of the North-central Province, found a dagaba, ' a copy in miniature of the Thuparama [dagaba]/ at which, between the outer pillars there was * a wall about three feet high, generally formed of a single slab of stone deeply carved in the post-and-rail pattern/2 There is nothing to indicate that any detached fence of this kind existed at the Thuparama or any other dagaba at Anuradhapura.

No example of the hanging of paintings round dagabas, either in Ceylon or elsewhere, has' been quoted by Fergusson? nor is it necessary. The purpose for which the circles of pillars were erected round them is explained quite clearly in the histories, and will be found stated in my account of the Ruwan-waeli dagaba. They were employed for supporting festoons of lamps, and two instances are mentioned in which such pillars

1 On the principle of setting a thief to catch a thief.

2 Manual of the North-Central Province, p, 240.270 ANCIENT CEYLON

were so utilised. This was not merely a subsidiary function ; it was their chief purpose. It is recorded that wooden pillars were fixed round two of the dagabas, at one of which this is explicitly stated to have been the special reason for their erection. Such festoons of lamps were not hung simply as decorations; they were well-known demon-scarers. Even at the present day large numbers of small lamps are lighted round some of the dagabas at festivals, and I know that one procession of pilgrims from the North-western Province presented one thousand lamps, as well as oil for them, on one of these occasions.

With respect to the tenons on the capitals of the two inner circles, the facts that the pillars of the innermost circle are only two feet distant from the base of the dagaba, and that their centres are only about four feet apart, afford strong indications in favour of their being originally intended, as one of their duties, to support a light roof over the dagaba ; and in my opinion the chief evidence which tells against the existence of such a covering is the record of the fixing of a golden pinnacle on the spire in the fourth century A.D. It is possible, however, that such a roof may have been erected, and may have been removed by that date. It would not be a very difficult matter to construct a conical roof resting on the two inner rows of pillars, that would exert no outward thrust. The weight being well distributed over a large area its stability would depend on the character of the foundations, If these pillars did not uphold such a roof, the tenons show that the two inner rows must have sustained a covering over a circular procession-path round the dagaba, in addition to the special duty of all the pillars as supporters of festoons of lamps,

THE TOPOGBAPHY OF ANUEADHAPURA

It is necessary to gain a clear idea of the general outline of

Anuradhapura in. early times in order to understand any arguments regarding the positions of the chief dagabas in it, and

the reader is referred to the annexed plan in connection with the following remarks. vwl

I272 ANCIENT CEYLON

A low flat-topped ridge runs north and south on the western side of the Kadamba river, now called the Malwatta-oya,1 parallel to it and nearly a mile distant from it. The fortified part of the town was built along the top of this ridge, with at least one gate2 on each side facing the cardinal points. The principal gate at the southern end of the city led into two ornamental gardens of the king, called the Nandana and Mahamegha Gardens. From the eastern gate a road passed nearly due east to the Mihintale hill, eight miles away, crossing the Kadamba river by a bridge carried by upright posts, like all other ancient bridges built across the rivers of Ceylon.

The Nandana garden was also known as the Jdtivana (Mah., i, p. 64), and was evidently a narrow enclosure, ' in a delightful forest, cool from its deep shade and soft green turf/ It was immediately outside the southern gate of the city (Mah., i, p. 54) ; and to the south of it, and extending to the bank of the river, lay the Mahamegha garden, a much larger tract of ground planted with flowering bushes and fruit trees, which was enclosed by King Muta-Siva, the father of Devanam-piya Tissa, in the first half of the third century B.C. In these two gardens, which were both made over to the first Buddhist monks, the Maha Wihara, * The Great Monastery/ and the Tissarama and Thiiparama monasteries were established, these latter being parts of the former, which probably included other subordinate wiharas.

The Thiiparama wihara and dagaba were constructed in the Nandana garden, the position of which is thus fixed by them. The Bo-tree, a cutting from the tree at Gaya in India, under which Gotama attained the position of Buddha, * the Enlightened One/ was planted in the Mahamegha garden, in which the great Ruwanwaeli dagaba was also erected in the

1 * The Flower-garden river/ perhaps so called because it ran along one side of the Mahame'gha flower garden. Mr, Bell, the Archaeological

Commissioner, terms it Malwata-oya; I give the name as I heard it in 1873, "r "~

25 ' The four gates of the capital* are mentioned (Mali., i, pp. 119, 136 and 141). These would be the four principal gates, one being near the middle of each side.THE EARLIEST DAGABAS 273

second century B.C. As the Abhaya tank, now called Basa-wak-kulam, was in existence before this garden was enclosed, it is clear, from the references to it in the Mahavansa, that the latter included all the land from the embankment of the tank, which is to the west of the garden, up to the river. The Maha-megha garden was bounded on the north by the Nandana garden, and on the south by the low ground which forms a rice field.

The limits of the Nandana garden, or Jotivana, on the east and west are not stated by the old writers. We may safely assume that on the west it included the narrow strip of ground extending up to the Abhaya tank ; but on the eastern side it is uncertain if it reached quite up to the river. On the northern side there can be no doubt that it was separated from the city by the ditch of the fortifications, the position, of the southern gate of the town being definitely indicated by the story given in the Pali Thupavansa regarding the transport of the cutting of the Bo-tree from the port at which it was landed to Anura-dhapura, by King Devanam-piya Tissa, in 244 B.C.

After describing its arrival at the port called Jambukola, and the proceedings there, the account is as follows :?' Then, on the fourth day he took the Great Bddhi (tree), and making superb offerings in due course reached Anuradhapura. Having given it a great reception at Anuradhapura, too, on the fourteenth day of the month, with the growing shadows, he brought in the Great Bddhi by the northern gate, and having conveyed it through the middle of the city, and taken it out by the southern gate to the site, five hundred bow-lengths from the southern gate, where our Supreme Buddha seated himself and entered into the Nirodha meditation, and the three former Supreme Buddhas indulged in meditation and sat, and where the Sirisa Bodhi of Kakusandha the Blessed One, the Udumbara Bodhi of Kdnagamana the Blessed One, the Nigrddha Bodhi of Kassapa the Blessed One were established?in that place, cleared for the occasion, which was like the forehead mark ..(tilaka) of the Maha-megha garden, at the portico of the palace he caused the Great Bodhi to be fixed/

T274 ANCIENT CEYLON

My friend Mr. J. A. Balfour, of the Irrigation Department, was good enough to get the distance carefully chained from the B6-tree along the road which passes the Ruwanwaeli dagaba, to the middle of a trench which runs east and west at a short distance to the north-east of the Thuparama dagaba, and which appears to be the ancient ditch outside the southern wall of the city. The actual length is 3986 feet, and it is 33 feet further to a raised bank on the northern side of the trench ; so that if the southern gate was on that road, and at the line of this bank, it would be 4020 feet from the Bo-tree. This would give a measure of eight feet for a bow-length, a size in excess of the length of most modern bows, which are usually six or seven feet long, but not greater than one in the British Museum. In some manuscripts there is mentioned a measure which is termed a * Great Bow '-length (Maha Dunna) ; this may be the measurement referred to by the author of the Thiipavansa. Dr. Davy, writing in 1816-1820, stated that the length of a bow was then usually nine feet.1

The distance from the Bo-tree to the city gate cannot be reduced, or it would fail to meet with any trench or bank such as would mark the boundary of the *city; and in fact were the city wall more than a trifling distance nearer the Bo-tree it would run into the buildings that were erected round the Thuparama dagaba, which are known to be outside the wall of the city. Although the number of bow-lengths mentioned in the Thiipavansa must be merely an approximate round number it thus sufficiently confirms the position of the southern gate of the city. As the line along which the measurement was taken is that of an ancient road leading directly from the Bo-tree into the old city it is thus practically certain that the southern gate was at the point where it crosses the bank at the side of the trench, which is now marked by an irrigation channel from Basawak-kulam, laid out by me along the old ditch in 1873. The two royal gardens included all the ground from this channel up to the ricefield to the south of the Bo-tree.

The position of the northern boundary of the fortified part

1 An Account of the Interior of Ceylon, p. 244, footnote.THE EARLIEST DAGABAS 275

of the city is more doubtful. In all probability it was fixed at the point where the ridge ends at that side, at a distance of about three-quarters of a mile from the southern gate. If so, the shape of the fortified part would be a narrow oblong, extending only along the top of the ridge, and not into the low ground on the east and west sides. The sites of the various suburbs of the city are not now distinguishable, but one or two of them will be considered in dealing with the identifications of the edifices mentioned below,

THE PATHAMA CETIYA.

The second dagaba erected at Anuradhapura was the Pathama Cetiya, which was raised to commemorate the spot where the Buddhist apostle Mahinda and his companions were supposed to have alighted when they proceeded from Mihintale on the occasion of their first visit to the city. From the account of their coming which is given in the Mahavansa (i, p. 53), it is clear that this place was on the side of the public highway leading out of the town to Mihintale, and we are expressly told that it was 'in the eastern quarter of the city/

In the description of the consecrated boundaries fixed by Devanam-piya Tissa, which included the city, this dagaba is mentioned as lying north-west from two special trees that were on the bank of the Malwatta-oya. Thus it appears to have been at some moderate distance from the river, but not very far away. It was also distinguished by being selected as one of the places where the eight first shoots of the great Bo-tree were planted.

The dagaba is mentioned only once more in the Mahavansa, in the description of a royal procession through the city, on which occasion King Mitta-Sena (435-436 A.D.) rode on the white elephant that was kept for the temple services. The words are, * And he mounted Mm, and rode through the city in procession, and commanded that he should be stationed at the Pathama Cetiya, outside the eastern gate/

THE MIHINTALE MAHA SAEYA The third dagaba, built on the hill at Mihintale, is stated,276

ANCIENT CEYLON

but not in the historical works, to contain a single hair of Buddha. It seems to have been a structure in which the old annalists took little interest, and as a result there are almost no records respecting it. I have already mentioned that it was one of the works of Devanam-piya Tissa, who built it probably about 243 B.C.

Its shape is a hemisphere resting on three low circular basal platforms, and it had the usual square tee, faced with post-and-rail work in false relief, and doubtless also a spire, probably surmounted by a chatta or umbrella, like all the other large dagabas.

-"7 f' ii'v^Sg?^:?-5*r

'C^~=£?==X:_.

"' . . FIG. 71. The Maha Saeya, Mihintale. '

It is much larger than the Thuparama dagaba. The dome is about 84 feet in diameter and some 44 feet high. The tee was about 20 feet wide and 10 feet high. The total height

of the present ruin is 65 feet. The basal platforms form steps

each about 4 feet wide and rather less in height; there is a quadrantal moulding round them. The ' wahalkadas' found at the other great dagabas of Anuradhapura are absent, and there are no encircling stone pillars, but wooden pillars wereTHE EARLIEST DAGABAS 277'

erected in their place between 9 B.C and 21 A.ix1 The above noted dimensions are taken from a photograph by Messrs. Skeen and Co., of Colombo, rv^v

From the Mahavansa (i, p. 128) we learn that King Laj ji-Tissa ^ ? ' encased with stone ' this d|gaba (as well as the Thuparama dagaba) at a cost of one hundred thousand pieces of money; but like the similar covering at the Thiiparama all traces of such work have disappeared, impossible as it would seem at such a site. Considering the size of the dagaba, I should think it not improbable that there has been some misunderstanding regarding some expression of the pre-Christian annalist; and that it is most likely that the laying of the flooring of the platform round the dagaba was the work done at both structures.

It must have suffered like the Anuradhapura dagabas during

* the periods when South-Indian invaders ruled the country in

the eleventh and thirteenth centuries, and we may assume

that it was included among the sixty-four dagabas whicti

Parakrama-Bahu I repaired at Mihintale in the twelfth century*

In the latter part of last century it was in little better state than some of the other early works, and the spire had fallen, as Well as large sections of the face work of the dome, and the structure was nearly surrounded by a talus of fallen brickwork covered with bushes. Its repair made considerable progress under the direction of the late Mr. R. W. levers, when he was the Government Agent of the Province, and its further destruction was thus arrested.

The dimensions of the bricks used in this structure have been given in a former chapter. It is of archaeological interest to note that when vainly searching for letters or marks that might have been left on them by their makers, I found on the side of one of them, which I handed over to the Archaeological Commissioner, a representation of a plain 'Buddhist railing/ consisting, I think, of three uprights and three cross bars, a post and rail fence like those built in stone in India. As the brick was one that had fallen out of the body of the

*: "? I Pujavaliya, p. 20.278 ANCIENT CEYLON

dagaba with others, and is also of the size of the earliest ones used in the dagaba, which must belong to the original work, the discovery of this design on it proves that the knowledge of this form of construction dates in Ceylon from the middle of the third century B.C.

SMALL DAGABAS

During the reign of Uttiya, brother and successor of Devanam-piya Tissa, it is recorded that two dagabas were built over the ashes of the introducer of Buddhisid, the great apostle Mahinda, and his sister Sanghamitta, the first Superior of the Nuns. Evidently they were comparatively small structures.

The remains of a dagaba 21 feet in diameter, which now bears the name ' Sanghamitta Thupa' and lies northeast of the Thuparama dagaba, were excavated by Mr. Bell; although he found nothing to prove that the modern name is correct he thought it possible thate some of the ashes of the princess may have been deposited at this site/ l The dagaba in which the ashes of Mahinda were laid was in the eastern part of the grounds of the Maha Wihara ; it has not been traced.

The same king is also stated to have built a dagaba, also doubtless a small one, to mark a spot where two previous Buddhas, Konagamana and Kassapa, were supposed to have preached at the Sudassana or Somana Malaka, ' the Beautiful Enclosure/ This place also has not been identified ; the context seems to show that it was not far from the site of the Ruwanwaeli dagaba, and probably to the southward of it.

To the south of the Thuparama dagaba another small structure of this Mnd was also erected by a younger brother of King Uttiya, called Asdka (Mah., i, p. 61), who is perhaps the same as the Asela who subsequently succeeded to the throne

1 Annual Report, 1895, P- 2- He found a 'small broken cetta or relic

chamber, in the form of * an even cross/ in it, at about the level of the top of the basal platform or step. The dagaba was built on a

circular platform, 31 feet in diameter, paved with brick.tsui\i T/ -0

^3^'-^

THE EARLIEST DAGABAS 279

near the end of the third century B.C. Its object was to commemorate a site at which the preceding Buddha, Kassapa, was said to have preached when he visited the fabulous Visala Nagara, which was supposed to have been at that time the capital of Ceylon. ;

Another similar dagaba to mark a place called the Naga Malaka,' the Cobra Enclosure/ where the Buddha Konagamana was believed to have preached, was erected in his father's life-time by Thulathanaka, who was king in 119 B.C. This structure was to the southward of the last-mentioned one. Thus there were three that were roughly in a north and south line, that erected by Asdka being in the middle, and the Thuparama dagaba at the northern end of the line. Both these small works seem to have completely disappeared, unless a mound that is now surmounted by some well-cut pillars of a later ruin is one of them.

THE RUWANWAELI DAGABA

We next come to the period of KingDuttha-Gamini (161-137 B.C.), who built two large dagabas at Anuradhapura. As one of these was north of the Bo-tree, and south-east from the Thuparama dagaba, there can be no doubt as to the identification of the building now known as the Ruwanwaeli or * Gem-Sand * dagaba, and formerly called also Hemamali, Sonnamali, Ratanavali, and the Maha-Thupa, * Great Dagaba/ even after larger ones had been erected. Owing to the interest with which the work was invested on account of its originator, and through its being the earifest of the greater, dagabas at Anuradhapura, we possess a much more complete history of it and its construction than of any other early building, / either in Ceylon or India. , *

Duttha-GSmini is described as dying in 1376.0., before this work was finished^ and his brother and successor, Saddha-Tissa (137-119 B.C.), is said to have completed it after his death. According to the narrative the dome itself was built during the life-time of Diittha-Gamini to hold some undescribed of Buddha; and his brother constructed the spire,28o . . ANCIENT CEYLON

its base, and an enclosing wall ' decorated with the figures of elephants/ We learn from the Mahavansa (i, p. 114) that the original dagaba had the usual three basal ledges.

King Lajji-Tissa (119-109 B.C.), the son of Saddha-Tissa, then erected three stone ' altars J at the dagaba, each costing one hundred thousand coins of some kind. The amount expended on them shows that the 'frontispieces* orwahalkadas^ must be referred to or included, and not merely the ordinary flower-altars. Up to this time the square round the dagaba had not been paved with stone slabs, since it is stated of the next king Khallata-Naga that 'enclosing the beautiful Great Thupa Hemamali, he formed a square strewed with sand with a wall built round it" (Mah., i, p. 129).

In the reign of Bhatikabhaya (20 B.c.~9 A.D.) ' two basement cornice ledges' were built at the dagaba. What these were is not quite clear; the remark does not seem to be applicable to the stone-work on the basal platforms which surround the dome, as these have no cornices. Some additional stone cornices on the wahalkadas perhaps may be referred to.

The next king, Maha-Naga (9-21 A.D.), laid the flooring on the square round the dagaba, and appears also to have made the lower outer square, which was * strewed with sand, (Mah., i, p. 136), His son Amanda-Gamim / fixed a chatta [or umbrella-shaped top] over the chatta of the Maha Thupa, as well as cornices on the base and crown [tee] of that edifice ' (Mah., i, p. 137). The first chatta may have been part of the original work of Saddha-Tissa. Evidently the spire had now two chattas, one superimposed over the other.

In the reign of Siri-Naga I (196-215 A.B.) we read (Mah., i, p. 144) of the construction of a gilt chatta at this dagaba; this apparently was a third one fixed above the other two. The Dlpavansa attributes to his son Voharaka-Tissa (215-237 A.0). the construction of another also. Sangha-Tissa I (248-252 A.D.) caused the chatta to be re-gilt, and we learn that on each of the four faces of the base of the spire [in reality the tee] there was a representation of the sun, in the centre of each of which the king placed a gem which cost one hundred thousand coins. A glass pinnacle was also placed onTHE EARLIEST DAGABAS

281

the summit of the spire (which thus appears to have passed through the upper chatta), from a mistaken idea that it would prove a protection against lightning. Evidently the spires of some dagabas had been damaged by thunderstorms before this date, as might naturally be anticipated ; they could not fail to be struck sometimes. A golden chatta was again constructed at this dagaba by Dhatu-Sena (463-497 A.D.) ; this may have been merely a restoration of the former upper one. King Moggallana (608-614 A.D.) presented a new cloth covering to the dagaba ; and Kassapa II (652-661 A.D.) fixed

FIG. 73. The R-uwanwaeli Dagaba.

a jewelled pinnacle on the spire, which again indicates that it passed through the uppermost chatta.

In the time of Kassapa V (929-937 A.D.) the second queen,

Rajini, * made an offering of a silken covering for the Hemamala

cetiya' (Hah., ii, p. 80). This offering was repeated in the reign of Mahinda IV '(975-991 A.D.).

The dagaba appears to have been damaged by invaders from

southern India in the eleventh century, and with the other chief structures at Anuradhapura was repaired by Tamil prisoners of war during the reign of Parakrama-Bahu I (1164? 1197 A.D.). A relic of this work is to be seen in an inscription on one of the stones of the flooring of the enclosure :282 ANCIENT CEYLON

Baku sabha pahanayak, ' a stone (presented by) the Gaja-Bahu Assembly.'

It was again broken into by the invaders from Kalinga in the time of Magha (1215-1236 A.D.), and was restored for the last time in the reign of Parakrama-Bahu II (1240-1275 A.D.). This work had been commenced by his father, who was unable to finish it (Mah., ii, p. 306).

The dagaba was then left to fall into ruin once more, by the neglect of centuries, and the spire, the greater part of the tee, and the upper part of the side of the dome slipped down in a high talus that covered all the base of the structure, which then becamfe once more overrun with bushes and trees. At the beginning of 1873, its restoration was again undertaken by the energetic young Buddhist monk who was in charge of it, and it is still making slow progress, dependent on the subscriptions furnished by the large numbers of pilgrims who visit the old city at the annual and other festivals. The re-facing of the dome is not yet completed.

After the fallen debris had been dug away, and the support which it had given to the lower part of the cupola had been thus removed, a slip occurred of a section of the brickwork on the southern side of the dome; and on the occasion of a visit that I paid to the town at Christmas, 1886, I was surprised to find that this slip, which had taken place in 1885, had exposed the finished but unplastered surface of an inner dagaba, round which a shell of brickwork, twenty feet thick, had been built.1 The mass of brickwork that had fallen consisted merely of this outer shell; the inner work was intact, and disclosed throughout all the exposed surface the original face-work of unbroken and evidently undisturbed bricks, all laid as ' headers/ with very fine joints. I was informed that there is a tradition that this also is only a shell, and that inside it there is a still smaller dagaba; but no reliance can be placed on such tales when they are unsupported by the authority of any of the historical works.

In any attempt to explain this method of building the dagaba

1 The actual thickness as measured by me was 19 feet uj Inches.THE EARLIEST DAGABAS 283

it is obvious that the evidence afforded by the sizes of the bricks employed in the two portions of the work must be all-important. Those in the outer shell average 8*99 inches in width, 2-90 inches in thickness, and only 14-06 inches in length; Bt. is 26-1, and the contents 366 cubic inches. Those in the inner work average 9-67 inches wide and 279 inches thick ; * Bt. is 27. The length could not be measured as all are' headers/ The difference in the average widths proves that entirely different moulds were used for the outer bricks ; the manner in which the outer shell is built is also much rougher than in the inner work. It is therefore certain that the outer work was not carried on without a break or stoppage in the brick-moulding, and probably also in the building work ; and thus there is every probability that the outer shell was built by another king than Duttha-Gamini.

On making a careful examination of large numbers of bricks that had fallen out of this outer shell, I discovered on several of them a small series of letters that must have been inscribed on them before they were burnt, by the persons who made them. They are of the early angular types which date from prior to the time of the GaUena inscriptions,1 or say 85 B.C. ; and thus we must ascribe the building of the outer shell to some period between that date and the death of Duttha-Gamini in 137 B.C.

We are therefore reduced to five kings who reigned during this period, to one of whom the work, which would occupy several years, must be attributed. Of these, Watta-Gamini was fully engaged with the construction of two other dagabas, one of them being much larger than the RuwanwaelL His brother Thullathana reigned only forty days. Of the other three kings, Saddha-Tissa and his sons Laj ji-Tissa and Khallata-Naga, by far the most likely person to undertake the work was Saddha-Tissa, who reigned for the longest period (137-115 B.C.), and was the brother and successor of Duttha-Gamini. The completion of the dagaba is attributed to Mm by the Maha-vansa, which says :?* This monarch, whose name implies

1 See below, The Earliest Inscriptions, Nos, 68-72 (Fig. 153).284 ANCIENT CEYLON

the sincerity of his faith, completed the pinnacle and plastering" of the dome, and the enclosing parapet wall decorated with figures of elephants, of the Maha Thupa' (i, p. 128).

There is still a brick wall round the upper enclosure at the dagaba, which has the fore-parts (heads and fore-legs) of elephants built in relief, four feet seven inches apart, in the outer face. The bricks used in these figures are of varying" sizes, some of the lower ones being only two inches thick ; these belong to some time in the ninth to twelfth centuries, and they clearly prove a complete reconstruction of the wall at about that period, perhaps by Parakrama-Bahu I. The larger bricks are 17 inches long and about 2-85 inches thick, dimensions which are certainly pre-Christian. The difference between their length and that of the bricks in the outer shell of the dagaba proves that they were not moulded at the same time as the latter ; they may belong to the time of Khallata-Naga, who is recorded to have enclosed the square round the dagaba.

In that case, the figures of elephants attributed to Saddha-Tissa may perhaps be those in the uppermost platform or basal ledge, which is ornamented with the heads of elephants carved in limestone and set in the face of the top course. If so, he must have been the king who enlarged the dagaba. The similarity of the dimensions of the bricks in the inner and outer work points to the lapse of a very short interval between the building of the inner dagaba and the resumption of brick-making for its enlargement.

On the available evidence, it may be decided as a practical certainty that Saddha-Tissa added the outer shell of the dagaba. Evidently the early annalists or later historians, in their desire to exalt the fame of their favourite hero, Duttha-Gamini, omitted to give his brother the credit due to Mm for the greater part of the work done by him.

In other respects, the fact that there is no trace of plaster on the surface of the inner dagaba is another proof of the extreme accuracy of the account in the Mahavansa, which states (i, p. 123), * When the construction of the spire and the plastering of the cetiya alone remained to be completed, the king was afflictedTHE EARLIEST DAGABAS 285

?with the disease which terminated his existence/ We now see that his brother decided to enlarge the whole structure, and then, only, to add the necessary protection of the plaster. It was no slight work to undertake ; if the outer shell is of the same thickness throughout, the amount of building done by Mm forms considerably more than one-third of the whole volume of the dagaba.

The description of the relic-chamber leaves no doubt that It was in the upper part of the dome, like that of the Thuparama dagaba, and was a comparatively large room. It is described as having the walls covered with paintings, and containing in the centre a Bo-tree with a silver stem and golden leaves, above which was suspended a priceless canopy hung with pearls, while at each of the four sides of the chamber there was a small golden statue of Buddha sitting on a golden throne. On another throne, relics believed to be those of Buddha, enclosed in a golden casket, were placed by .king Duttha-Gamini in person. Above this relic-apartment another room was formed in which the chiefs and people deposited large quantities of jewellery. (Mah., i, p. 123.)

A very much smaller dagaba at Hettipola in the Northwestern Province, broken into by treasure-seekers in 1877, has similarly two large (but undecorated) relic-chambers one over the other, the intervening floor being formed by slabs of stone that passed across from wall to wall, partly supported by two stone beams fixed under them transversely, an arrangement evidently like that of the Ruwanwaeli dagaba. In the Hettipola dagaba eight small sedent figures of Buddha, made of some kind of cement and covered with gold, were placed on thrones similarly made and covered with silver, which were set in four rectangular niches formed in the four walls of the lower compartment. In front of each throne there was a small relic-casket or karanduwa of clear quartz, enclosed in a golden dagaba-shaped case, with a spire and tee. Unfortunately I have no measurements of the bricks of which this structure is built, and therefore I can express no opinion regarding the age of the work. It may be an ^early one.286 ANCIENT CEYLON

Inside the relic-chamber of a smaller dagaba near it there is part of a pillar on which a royal grant in letters of the tenth century had been cut. This dagaba can hardly be of earlier date than the twelfth century. As far as I remember, the chambers in the larger one were six or eight feet across.

With this example before us it is easy to believe that with the exception of the enormous dimensions attributed to it (80 cubits square, which may be safely divided by ten),?the detailed account of the room in the Ruwanwaeli dagaba, as preserved in the Mahavansa, is a true description of the apartment and its decorations, written by a contemporary annalist who either actually saw it, or heard it described by others who had seen it.

The shape of the Ruwanwaeli dagaba is thus explained in the Mahavansa (i, p. 112). Duttha-Gamini enquired of the architect in what form he proposed to build it. It is therefore clear that various shapes of dagabas were even then known. 'The bricklayer, filling a golden dish with water, and taking some water in the palm of his hand dashed it against the water in the dish; a great globule like a ball of crystal rose to the surface ; and he said, " I will construct it in this form." The monarch, delighted, bestowed on him a suit of clothes worth a thousand, a pair of slippers, and twelve thousand kahapanas/ It is refreshing to read of a king who gave such desirable marks of his appreciation of an architect's intelligence; he resembled in this respect some of the worthy Egyptian monarchs. At the present day even the slippers are not given to successful architects in Ceylon.

When drawings of the chief dagabas at Anuradhapura were made in 1877 by Mr. Smither, the Government Architect, he was of opinion that the dome of this dagaba was a hemisphere, as described by the old writer. It is 254 feet in diameter. It rests on three short cylinders, the upper one having a diameter 12 feet greater than that of the dome and being 5 feet 6 inches high ; the middle one is 14 feet wider still and 4 feet 9 inches high ; and the bottom one is 14 feet wider than the middle one and 5 feet 9 inches high. Thus the height of the three cylinders is 16 feet, and they form three basal ledges or nar~s. Plan ana ElcvatJort of S. WShalkadai

FIGS. 74~79« The Ruwanwaeli Bagaba.288 ANCIENT CEYLON

row platforms round the dome. All are paved on the top with small blocks of limestone each 3 inches thick, 10 inches long, and 5 inches wide. There is a quadrantal moulding of limestone 15 inches high round the base of the lowest one. Out of the limestone coping of the retaining wall of the upper platform projected 133 elephants' heads, also cut in limestone. The face of the retaining walls of the platforms is built of small limestone blocks, the top course of the lowest one being carved with a * Buddhist railing ' of two bars in false relief, evidently in imitation of the detached railing of early Indian works. Round the upper platform of a broken stone relic-case, apparently taken out of the chamber behind one of the ' wahalkadas/ a similar rail of two bars is carved; it probably belongs to nearly the same period.

On the top of the dome there would doubtless be a square * tee ' of brickwork, ornamented, as in the other great dagabas, with post-and-rail work in false relief, and having a circular disk of the sun, the great demon-scarer, in the centre of each face/ Above this must have risen the spire, tapering slightly, and probably, like those of similar buildings, springing from a cylindrical base. At its top, or immediately below it, there dppears to have been from the first a solid mushroom-shaped or lens-shaped ' chatta/ as a symbol of the royal honours paid to the relics, and perhaps considered to be quite as important as a magical protection from evil. The whole height is recorded to have been 120 cubits (Mah., i, p. 62), and the same figure is given by a later historian as the height when Parakrama-Bahu I restored it. If, as is likely, the early cubit was two feet in length, this would be 14 feet less than the diameter of the dome. The top of the present mound is 178 feet 8 inches above the pavement at its base. The paved platform on which it rests measures 475 feet by 473 feet.1

On three, if not four sides, facing the cardinal points, a

1 By the kind permission of the Secretary of State for the Colonies, 1 have reduced from Mr. Smither's drawings a plan of this dagaba, and a plan and elevation of the Southern Wahalkada. (See Figs. Kos. 74 and 75.) /THE EARLIEST DAGABAS 289

rectangular' frontispiece/ as Mr. Smither termed it, in Sinhalese-commonly called a Wahalkada, was built, projecting outwards from the dagaba. This consisted, in the face, of a series of tiers of horizontal stone cornices or projecting moulded-bands, separated by plain smooth-dressed stone-work. It was flanked at each end by two pillars, a high inner one on which sat a lion, looking outwards, with open mouth, and a short outer one, on each exposed side of which were conventional decorations in sunk relief. In 1886, I observed fragments of gilding on one pillar, and of painting on another. Slabs with roughly carved five-headed cobras, and in other respects like those at the Jetavana dagaba described below, were fixed outside these pillars. Twenty-six elephants' heads x carved in stone project from the plain course above the lowest cornice of the wahalkada.2

The object of these highly decorated offsets, each 34 feet 2 inches long, appears merely to have been to form ornamental and also protective backgrounds for disengaged stone flower-altars placed on the pavement before them. Steps were also built behind these wahalkadas, leading to the two upper basal platforms, and a room for relics, measuring about 13 feet by 6J- feet, was constructed behind each of them.

The Dipavansa states that King Bhatikabhaya (20 B.c.-g A.D.) ' made strong pillars for placing lamps round the foot of the Thupa ' (p. 213) ; and the Pujavaliya (p. 20) also records the erection of wooden posts round it and the large dagaba at Mihintale, by King Maha-dathiya Maha-Naga (9-21 A.B.),

1 The elephants* heads at this and the dagabas next described were probably inserted for protective purposes. Those in the wall round the court-yard would have a similar function; others were also built in the surrounding walls of two other dagabas (Mah., i, p. 163). See my remarks on the protective power of all auspicious objects, in the final chapter. The elephant, as the Vahana or riding-animal of Indra, was a demon-scarer.

2 Although it is not a suitable name for these structures, since it commonly means the gate-way of a palace, I employ their usual colloquial Sinhalese title, in preference to Reredos. * Frontispiece,* the term applied to them" by Mr. Smither, is inapplicable, as they are in reality projecting backgrounds, and ' Altar-background * is cumbersome* ' Offset * and * Screen' fail to indicate their chief function.

0? ANCIENT CEYLON'

apparently in order that festoons of lamps might be hung on them. Similar strings of lamps were hung round the. Ambatthala dagaba at Mihintale, which is surrounded by two rows of stone pillars (Mah., i, p. 136). This certainly indicates the chief purpose of such pillars round the dagabas.

After the original structure was erected, a small attached relic-house, of a rectangular shape, was built on the eastern side, standing out onto the pavement like that at the TMparama dagaba. The bricks behind it are 1678 inches long, 8-26 inches wide, and 2-36 inches thick; Bt. is 19*5 and their contents 327 cubic inches. These dimensions point to about the second century A.D. as the time when they were made. It is possible that a room of this kind existed on that site at a much earlier date, and was replaced by a new building in the second century.

On each side at the bases of the steps at this dagaba, as well as at all the more important buildings of Anuradhapura, a thin upright slab with an arched top is erected across the end of the balustrade. It resembles the stelae of Assyria on which the statues of the kings were carved. Many have plain tops like those of Assyria; others end in a blunt point at the centre of the arch. In the more elaborate examples the figure of a protecting deity is carved in high false relief on the face of the stone (Figs. 157 and 160); on others a vase is represented out of which spring lotus flowers, buds, and leaves; some are without any decoration of this kind.

An interesting feature at many of these stelae at Anuradhapura and a few other early sites, is a pilaster in relief on the outer side of each stele (Figs. 79 and 157), often surmounted by an animal facing outwards, which is always a lion, a humped bull, an elephant, or a horse. On the face of a guardian slab at Mihintale there is a bird?probably a hansa?standing on a pillar at the side of a figure of the Indian goddess Kali or Durga.

The function of these animals in these sites has not been explained ; it appears to be similar to those of the processions of the same animals that are sometimes carved on the semicircular slabs termed * Moon-stones * (in Sinhalese, Irahanda-THE EARLIEST DAGABAS 291

gala, ( Sun^and-Moon Stone '), which are often placed on the ground at the foot of the lowest steps at the entrances to religious edifices. Processions of this kind are carved in relief on the uppermost band of the wahalkadas at the Miriswaeti dagaba, which is next described. With these may be also compared the elephants1 heads that project from the face of these dagabas, and the fore-parts of elephants that were built of brick and plastered over, on the outer face of the wall of the inner enclosure at the Ruwanwaeli and two other dagabas. Processions of the sacred geese termed Hansa (which is identified with the Sun in the Rig Veda) are also carved on moon-stones ; and hansas, lions, horned lions, and elephants are carved round the capitals of pillars at dagabas and monastic buildings. The Mahavansa states (i, p. 114) that ' rows of animals and hansas ' were painted in the Ruwanwaeli relic-chamber, along with representations of c the eight auspicious objects/ This appears to show that all these animals were carved because they were believed to have protective powers against evil spirits. This was certainly a function of the hansa, the elephant, the horse, the humped bull, and also of the lion.2 The whole subject

1 On one of these stones at Anuradhapura a half sun is carved, the central part having lotus petal decoration, outside which are the rays.

2 At Saesseruwa, in the North-western Province, Mr. Bell found a dog and a ram-like animal carved with the others on a moon-stone. (Arch. Survey. Annual Report for 1895, p. 12.) The dog was a powerful demon-scarer ; see my remarks on it in The Earliest Coins.

The lotus which appears so often on moon-stones and elsewhere seems also to have had protective functions. In Egypt it was closely connected with Ra, the Sun-god, of one of whose forms it was considered to be a type. Nu is represented in The Book of the Dead (Translation by Dr. Budge, p. 141) as saying, " I am the pure lotus which springeth up from the divine splendour that belongeth to the nostrils of Ra." In a variant of the same Chapter 80 it is addressed, " Hail, thou lotus, thou type of the god Nefer-Temu ! I am the man that knoweth you, and I know your names among (those of) the gods, the lords of the underworld " (op. tit., p. 141).

In Fig. No. 78 and on the moon-stone above-mentioned it is represented as the central part of the Sun-emblem, I believe that where-ever the lotus appears at the entrances of buildings and at the dagabas, although its purpose" was partly decorative it had also a highly-protective function as a symbol of the Sun.292

ANCIENT CEYLON

is rather obscure; further information regarding it will be found in the chapters on the early coins and the Swastika* In Fig. No. 79 I have illustrated on the right* side one of these little pilasters, surmounted by a crouching lion, which is at the eastern entrance to the Ruwanwaeli dagaba. The shaft is only 2j inches wide. On the opposite side of the page is an elevation of a detached square pillar (Fig. 78) 8 feet 2 inches high, of an early type, which stands near the dagaba. At the top it has a procession of three hansas, each

Indian Dagaba

FIGS. 80-83. The Buwanwaeli Dagaba (restored),

carrying a lotus bud by its stem. The design below it is a Sun-disk, the rays being incised between the outer circles ; the next space in the figure contains lotus petals, and there is a flower in the centre. A row of pearls (which were very auspicious objects) follows, and below them and separated from them by a row of short bars are two plain loops separated by three bars. The pillar is highly symbolical and protective. In Figure No. 80 I have ventured to give a drawing which attempts to reproduce the outline of this interesting dagabaTHE EARLIEST DAGABAS 293

when in its complete state, according to the description supplied by the chroniclers. The winding line indicates the outline of the dagaba before its restoration, and follows the contour in a photograph taken more than thirty years ago by Messrs. Skeen and Co., of Colombo. The inner dotted semicircle shows the size of the dome built by Duttha-Gamini.

The restored dome is shown as a half sphere, agreeing in this respect both with the other great dagabas of Anura-dhapura and the account of it in the Mahavansa. The tee and base of the spire follow the proportions of those of the two other great dagabas which are described below. The height to the top of the lowest chatta is 240 feet, that is, 120 cubits at two feet per cubit, according to the scale adopted in a later part of this chapter.

The chattas follow the type of one carved in a crystal relic-case found at Tissa, and illustrated in Fig. No. 95. I have followed this unique example of an actual chatta in assuming that there was no pinnacle above the original one. There is also the authority of a later example in the Amaravati carvings at the British Museum (slab No. 34), of which I give a sketch (Fig. No. 81), as well as a beautiful stone flower-altar in the Ruwanwaeli enclosure which, as Mr. Bell has already pointed out, is of this form (Fig. No. 82). At the Mahanaga dagaba at Tissa, in the Southern Province, there is a much plainer flower-altar of this type.

I have represented, from a photograph for the loan of which I am indebted to Mrs. Waterfield of Malvern, the outline of the upper part of a highly decorated miniature stone dagaba with five chattas, which was dug up at a village thirty-six miles from Peshawar in India, by Mr. Stuart Waterfield of the Indian Civil Service, and which is now in the Calcutta Museum (Fig. No. 83). Although it is of much later date than the Ruwanwaeli dagaba, I have accepted the arrangement of the chattas on the upper part of the spire as a guide in drawing those of the latter work.

The total height to the top of the pinnacle thus becomes 305 feet; it may have been 25 feet lower if the spire passed througji the first chatta.294 ' ANCIENT CEYLON;

The appearance of the great white dome coverecfwith plaster and periodically white-washed, and of this high spire towering aloft in the blue sky, with its gilded upper chattas reflecting the bright rays of the tropical sun, must have been extremely effective and picturesque. It was a striking memorial of its great founder, and of the artistic genius of the Sinhalese race. But it was doubtless a far too prominent target for thunderstorms; and when Parakrama-Bahu undertook the restoration of the structure he found it advisable to leave the summit of the spire at its original altitude of 120 cubits.

THE MIRISWAETI DAGABA

Duttha-Gamini also constructed the Maricavati dagaba, now called the Miriswaeti or Mirisawaetiya dagaba, at Anura-dhapura, immediately after he gained the throne, completing it and the surrounding buildings in three years, that is, in 158 B.C. It was erected in order to enshrine the relic which had been his palladium through all his fighting with the South Indian invaders. It is related in the Mahavansa (i, p. 96) that when about to undertake the re-conquest of northern Ceylon, and while still at Magama, he fixed a relic of Buddha in the head of his sceptre ; and doubtless he would attribute to its magical power his constant series of victories over his enemies. The sceptre containing this relic is stated to be placed in the base of the dagaba.

. The Miriswaeti dagaba shared in the vicissitudes that befel the others at Anuradhapura.

King Voharaka-Tissa (215-237 A.D.) restored the chatta on the spire, which is not previously mentioned and thus appears to have been an original work of Duttha-Gamini Kassapa IV (912-929 A.D.) handed over the charge of the dagaba to the nuns of the Maha Wihara.

It ¥as included in the dagabas which the Tamils ransacked in the eleventh century; and it was restored, with the other large works, by Parakrama-Bahu I. In the time of King Magha of Kalinga, though allusion is not specially made to it by the historians, it was evidently broken into again;FIG. 84. S. AVahalkada, Miriswaeti Dagaba.THE EARLIEST DAGABAS 295

and like the other dagabas it was repaired for the last time during the reign of Parakrama-Bahu II.

Although Anuradhapura was visited by one or two later kings, there is no record of any further restorations of the dagabas there; and all alike were allowed to fall into ruin and become overgrown with jungle. When I first saw the Miriswaeti dagaba in 1873 it was little more than a conical mound covered with large trees and bushes, all the upper part having slipped down in a talus round its base.

Bricks of three sizes were used in the outer work. The largest, and certainly the original ones, were 10-41 inches wide and 3 inches thick. The intermediate bricks measured 14-2 inches in length, 8-2 inches in width, and 2-34 inches in thickness, Bt. being 19-19, and the contents 272 cubic inches. These dimensions indicate a restoration in about the fourth century A.D., which the historians have not recorded. The smallest bricks were those of the restoration effected by Parakrama-Bahu I.

The shape of this dagaba seems to have been a hemisphere, resting on three short cylinders which formed three basal platforms or ledges round it, like those at the Ruwanwaeli dagaba. On the top of the dome there would be the rectangular tee, ornamented as usual with posts and rails on each face in sunk relief, above which rose the spire. Apparently only one solid chatta surmounted the spire.

It had three high rectangular stone wahalkadas (Fig. No. 84), 25 feet long, facing the north, south, and west cardinal points, each formed of a series of cornices or deep mouldings separated by bands of plain stone-work. Twenty-one elephants' heads project from the band above the lowest cornice; and on the uppermost band are carved in relief four processions of animals in one line, all marching to the left, and consisting of horses, humped bulls, lions, homed lions, and elephants. At the left of the six animals in each wing of the wahalkada, a man or deity stands facing them, and holding up his left hand; and a similar figure stands facing each group of five animals in the central part.

The wahalkadas are flanked at each end by two rectangularzg6 """"" ANCIENT CEYLON ' ' "

monolithic pillars, 13 inches wide in the face, the inner one being as high as the uppermost cornice, and being surmounted by a stone lion sitting on his haunches on a square capital with a Buddhist railing of two bars on its face; he is looking outwards, with half-open mouth. The outer one, which is very short, has vertical flutings on the face, and the lower half of a rayed sun emblem above them. The taller pillars have as ornaments on their face, a dwarf at the base, supporting on his head a vase out of which springs a tree decorated with a series of pairs of men and animals alternately, climbing upwards on each side of it. At the top, above the tree, there is a disk or' dharma-chakra/ a Wheel of the Law, on a pedestal, over which is a conical chatta in relief, with a snake lying head uppermost on each side of the pedestal. Above each snake is a Yak-tail fly-whisk,rthe emblem of a guardian deity. Behind the wahalkadas steps led to the two upper basal ledges.

According to Mr, Snrither's drawings, the diameter of the dome was 135 feet 6 inches. The upper basal cylinder had a diameter 9 feet 9 inches greater than that of the dome and was 3 feet nj inches high; the middle one was n feet 6 inches wider still and 4 feet 2 j- inches high ; and the lowest one was 12 feet wider still, and 5 feet 2f inches high. Thus their total height was 13 feet 4f inches. The height to the top of the ruin was 52 feet 7 inches ; and the paved basement platform on which it stands is raised 4 feet n| inches above the ground level, and supported by a retaining wall of stone. The total height when Parakrama-Bahu I restored it is stated in the Maha-vansa to have been 80 cubits.

This work and the Ruwanwaeli dagaba prove that Duttha-Gamini and his brother Saddha-Tissa may claim the credit of being the first rulers to appreciate the grandeur of the effect of an enormous white dome, far greater than anything of the kind previously erected in Ceylon or India, and admirably adapted to be an expression of stability, and permanence, and inaccessibility, such as the purpose of its construction demanded. The bold rounded outline gives one the feeling of a finish and completeness which, to my mind, the pyramid, with its salient angles, does not possess. The simplicity ofTHE EARLIEST DAGABAS 297

a very large dome is one of its charms ; its appearance of strength and nobility would be lessened by any decorative additions on its surface. The effect of the three platforms round the bottom of the dome is decidedly good ; they increase the idea of stability and add a finish to the base of the structure that is missing in early Indian works.

SELA CAITYA

Lajji-Tissa (119-109 B.C.), the second son of Saddha-Tissa, built a dagaba of stone, called in the Mahavansa the Sila Thupa, the Stone Dagaba, ' in front of the Thuparama' (i, p. 138). In the Dipavansa (p. 211) its site is better defined as being to the east of the Thuparama, and the structure is termed in that work the Digha Thupa, the Long Dagaba, the name doubtless indicating a different shape from that of the other dagabas in Anuradhapura.

There is only one small dagaba, now known as the Sela Caitya, in the position described, and although it is built ?of brickwork like all the others it appears to be the building erected by this king. If so, it must have been destroyed, and afterwards rebuilt in brick. Mr. Bell discovered that the material in the interior consisted of * earth and brick fragments, cased in by burnt brick/ an indubitable proof of its reconstruction.

The largest bricks found at it are apparently those of an early post-Christian date, being 9 inches wide and 2-53 inches thick, with Bt. 22*8. If the length was 6 times the thickness it would be 15-18 inches, and the contents would be 345 cubic inches. The size points to the ist century A.D, or late B.C. as the period when they were burnt. There are bricks of three other sizes at this dagaba, which belong to later restorations, probably of the second, twelfth, and thirteenth centuries.

The dome of this dagaba is 37 feet 8 inches in diaipeter at present, according to Mr. BeU, this being possibly nearly its original width ; it is a mere ruin 10 feet high on a platform raised 5 feet 3 inches above the ground level, and 46 feet 8 inches square. ? The' original form can only be surmised; its name, * the Long Dagaba/ shows that it must have been2,98 ANCIENT CEYLON

either of a high bell-shape, or more likely almost a cone, these being the only two forms that appear very high in pro-r portion to their width.

On sinking a shaft down the axis of the dagaba, Mr. Bell found that the relic-chamber was fa brick-cased cella, 3 ft, 6 in. in breadth by 2 ft. 6 in. high. Its bottom was on a level with the maluwa [enclosure] platform, and formed by a monolith slab, a foot thick, which covered a square yantrcigala [a slab with rectangular holes, often 25 in number, sunk in it to receive valuables] of nine partitions all empty.'1

We next come to the reign of King Watta-Gamini Abhaya, who after being dethroned in 103 or 104 B.C., in the first year of his reign, by South Indian invaders, afterwards re-occupied the throne from 88 to 76 B.C. There can be no doubt that he built a great dagaba called Abhaya-giri, at the extensive monastery of that name which he established. There is no account of the relics which it enshrines.

For some unknown reason it has always, in modern times, been pointed out as one which is to the eastward of the Sela Caitya and the Ruwanwaeli dagaba, a mis-identification that perhaps dates from the restoration by Parakrama-Bahu I, in the twelfth century. The state of the great dagabas at Anuradhapura at that time is graphically pourtrayed by the old historian : ?' These three Thupas that the Tamils had destroyed were covered with great trees in which lurked tigers [sic, literally, leopards] and bears. And because of the great heaps of bricks and clay and the thickets of the forest no man was able to have access thereto' (Mah., ii, p. 260). This tends to prove that the place had been totally abandoned for so long a period that the ruins were practically unknown, and this might give rise to their wrong identification by the officials sent by the king to restore them. In giving an account of this restoration the Mahavansa mentions that the Abhaya-giri dagaba was 140 cubits high and another large one, the Jetavana. dagaba, .160 cubits high; whereas we have the statement -of the Rajavaliya (p. 52) that the latter work was built 140 cubits high by Maha-Sena. Thus it is clear that 1 Annual .Report, 1895, P- 2-THE EARLIEST DAGABAS 299

there is an error regarding the nomenclature in one of these works. If, as I shall endeavour to show, the present names of these two dagabas are wrong and should be transposed, this would seem to transfer the mistake to the twelfth century, since in more modern times the place was never so much overgrown that these sites could not be visited by pilgrims, and it is on record that several later kings went on pilgrimage to Anura-dhapura. As the error involves both structures it will be convenient to deal with them together.

IDENTIFICATION OF THE JETAVANA AND ABHAYA-GIRI

DAGABAS

The account of the establishment by King MaM-Sena (277-304 A.D.) of the Jetavana wihara at which the Jetavana dagaba was built is given in the Mahavansa as follows :? 4 When Maha-Sena was a youth his tutor, who belonged to the (Vaitulya) Abhaya-giri Community of Monks, then engaged in a violent religious feud with the (Theravada) Community of the Maha Wihara, had induced him to adopt the doctrines of his own sect; and as the result of this partizanship,when he succeeded to the throne he prohibited the giving of alms to any priests of the Maha Wihara. Afterwards, having a high opinion of a (Vaitulya) monk called Tissa, he decided to build a special monastery over which this person might preside, and he 'constructed the Jetavana Wihara for him, within the sacred limits of the garden called Joti, belonging to the Maha Wihara. He then applied to the priests [monks] of the Maha Wihara to abandon their consecrated boundaries in order that the ground might be consecrated for the new temple' (i, p. 151).

It has been noted previously that the Mahavansa clearly explains (i, p. 64) that the Jdtivana is only another name for the Nandana garden, which I have shown to be a narrow enclosure wedged in between the city and the Mahamegha garden. It is in this strip of land that we must look for the Jetavana dagaba ; and it seems inexplicable that any doubt should be felt that the only great structure of the kind in that part, although it is now commonly called the Abhaya-giri300 ANCIENT CEYLON

dagaba,1 is the building in question. As the Nandana or Jotivana garden had been granted to the Maha Wihara by Devanam-piya Tissa, along with the Mahamegha garden, the application to the monks to make over part of their consecrated land for the new monastery can be easily understood.

This identification leaves only one other dagaba of the largest size to be named at Anuradhapura, and this must be the Abhaya-giri, which is now mis-called the Jetavana dagaba.

The description of its site given in the Mahavansa (i, p. 131), although doubtless sufficiently definite at a time when the different details of the surroundings were well known, is not very clear in these days when all is buried in jungle. A large body of Tamil invaders had marched straight on the capital from Mahatittha, close to Mannar, and they were encamped at a village near the city, called Kolambalaka, just as in the wars of the king's uncle, Duttha-Gamini, a similar force under Bhalluka, the nephew of the Tamil king Elara, had camped at the same place when advancing on the city from the same port (Mah., i, p. 100). The march of the two invading armies along the same route shows that this village was on the direct highway leading from Mahatittha to Anuradhapura, and therefore on the north-western side of the town.

Watta-Gamini led his forces out to meet the invading army, and fought a battle at Kolambalaka, in which he was defeated ; and ' mounting his chariot, fled through the Tittharama gate. This Tittharama had been built by Pandukabhaya, and had always been assigned as a residence to people of foreign religions. ... A certain Nigantha [a Jain anchorite] named Giri, seeing him in his flight, shouted in a loud voice, " The great black Sihala [Sinhalese] is flying." The Maharaja hearing this resolved within himself " when my wishes are realised I will build a wihara here1" (Mah., i, p. 129). Subsequently he regained the throne, and the account states that * Thereafter this monarch demolished the aforesaid Nigantharama, at which he was reviled in his flight, and on the site thereof built a wihara of twelve parivenas [monastic residences]. ... By

1 A local tradition which could point out the Dakumi dagaba as

the tomb of Elara cannot be accepted as possessing much authority,

V' / ^ v, «,, n ; \ ^ »/ V \ «t "^ fv ? o1 Jif/- \ **-«! * ? "%

/ $ i '' i *t V v ^ . ,THE EARLIEST DAGABAS 301

reason of the Arama having belonged to Giri, and by reason of the wihara having been made by the king Abhaya, therefore it was called Abhaya-giri Wihara' (Mah., i, p. 131). Thus the account shows that the monastery was between the fortified city of that period and the village to the north-west where the battle was fought.

In the account of Pandukabhaya's arrangement of the suburbs (Mah., i, p. 43) it is stated that a range of buildings for Vaeddas was established on the northern side of the cemetery, and to the eastward of these dwellings, that is, on the northeast of the cemetery,' he provided a residence [the Tittharama] for five hundred persons of various foreign religious faiths/ Unfortunately the site of the cemetery is unknown. It was crossed by the consecrated boundary fixed by Devanam-piya Tissa(Mah., i,p. 62), and the route followed by the king shows that it was on the western side of the city. A site to the northeast of it could not therefore be very ifar from the position occupied by the dagaba now wrongly termed Jetavana.

The clearest independent evidence in favour of this identification is contained in the account of Anuradhapura supplied by the Chinese monk, Fa Hien, in his narrative of his travels. He came from China to India for the purpose of obtaining copies of Buddhist manuscripts, and after spending six years there, he devoted two years to the same research in Ceylon, returning to China in the following year, 413 A.D. His words are (Dr. Legge's translation, p. 102) :?' When Buddha came to this country, wishing to transform the wicked Nagas, by his supernatural power he planted one foot at the north of the royal city, and the other on the top of a mountain [Adam's Peak], the two being fifteen ydjanas apart. Over the footprint at the north of the city the king built a large tope [dagaba], 400 cubits high, grandly adorned with gold and silver and finished with a combination of all the precious substances. By the side of the tope he further built a monastery, called the Abhaya-giri, where there are five thousand monks/ At this time, he states that the Maha Wihara had only three thousand monks, and it was thus inferior in size to the Abhaya-giri monastery.302 \ ' ANCIENT CEYLON

The Chinese traveller's reiterated statement that the monastery was at the north of the city must be conclusive as to the identity of the only great dagaba in that neighbourhood. I shall therefore refer to the two dagabas by the names that appear to me to belong to them, terming the northern one the Abhaya-giri, and the one to the east of the Sela Caitya the Jetavana dagaba.

There is, however, further evidence regarding the identity of the northern dagaba. When part of the debris collected round its base was removed by Mr. S. M. Burrows, late of the Ceylon Civil Service, he discovered a series of large stone relic-cases placed behind one or more of the wahalkadas which are annexed to the dome at the cardinal points. Several of these were uninscribed, but round two of them the following sentence was deeply cut in one line (see Fig. No. 153 for facsimile).

Siddham. Matu Tisa Maharajaha raji nimi tabi hada tani jani. Hail! Fashioned, established (for sacred purposes), put in the prepared place (in) the reign of the Great King Malu-Tissa.

The characters are those of the second century A.D., and the dedication belongs to the time of Kanittha-Tissa (165-195 A.D.). In an inscription at a monastery at Ussayppu. kallu, about nine miles from Marisi-kattu, and near the Modara-gam-oya, this king styles himself * the Great king Malu-Tissa, son of the Great King Niga/1 As Kanittha-Tissa was the

1 As this inscription is in a somewhat inaccessible part of the island I have given a facsimile in Fig. No. 153.

The transliteration is:?Siddham. Naka maharajaha puta Main Tisa ma (2) haraji ma ganane kariyihi wawa Luwimitayah(i) Cudataka wawiyi ca (3) jabo awiyi ca, mata karawiyi ca, tapawana awiyi caf me c(e)taka wawiya (4) bojiyapati, Karakalaya Kuba wihara kahi paca wata hiti. Ihata mula c(e)ta mawati (5) ya jina palisatiriya kotu dini.

Hail I The Great King Malu-Tissa, son of the Great King Naga, having formed and protected a tank of a great quantity of karishas (in extent) at Luwimitaya, and the Cndataka tank, and (thereby) caused rejoicings; and having protected the Ascetics* Forest; after having assigned the tank belonging to this dagaba, built the Karakalaya Kumbha wihara, suitable for the observances connected with theTHE EARLIEST DAGABAS 303

younger son of King Mahallaka-Naga, this proves that he is the monarch who terms himself Malu-Tissa, * Tissa-the-younger-brother' (maluwa in Elu or old Sinhalese) to distinguish his name from that of his elder brother, King Bhatiya-Tissa Tissa-the-elder-brother J (141-155 A.D.).1

It is clear that a relic-case with a dedicatory inscription of this king of the second century could not be placed in a dagaba which was not erected until the last years of the third century. Its discovery in the northern dagaba proves that this structure was in existence a century before Maha-Sena built the Jetavana dagaba, and is quite fatal to the nomenclature commonly accepted.

The dimensions of the bricks in the dom^ of the two dagabas

?s of the respective ERRATA. fais subject mOre

Page 302 line 17: * M«f*' should be - MaM to the east of the

Fag 3 i "breadth of 8-41

19 square inches, e of dated bricks to some date from the first to the fourth century A.D., inclusive.

In the northern dagaba, the bricks measure 18-92 inches in length, 9-62 inches in breadth, and 3*20 inches in thickness ;

requisites [clothing, food, bedding, and medicine]. Having constructed the chief dagaba at this place, he repaired the dilapidated buildings.

This inscription is a duplicate of one copied by Dr. E. Muller at Oalkowila (No. 98 in his list), the purport of which his defective copy led him to misapprehend. The, meaning of the word mata, which I have translated * rejoicings/ is doubtful. Luwimitaya (Lumitiya at Galkowila) = Lokamita ; compare Alu wihara = A! oka wihara. ' Protecting ' the tank probably means arranging for its maintenance and appointing a Guardian for it.

King Udaya II (952-955) nearly lost his life through his violation of the sanctity of the Ascetics' Forest (Mah., ii» pp. 82, 83). It was a ' Sacred ' forest to which ultra-ascetic monks, especially those termed Pansukulikas, retired, and it was evidently a sanctuary for offenders of all kinds. Its position is not known.

1 In Dr. Muller's Inscription No, 16, Kanittha-Tissa describes himself as * Tissa, the younger brother of the Great King Bhatiya-Tissa, (and) son of the Great King Naga.*THE EARLIEST DAGABAS 303

younger son of King Mahallaka-Naga, this proves that he is the monarch who terms himself Malu-Tissa, * Tissa-the-younger-brother' (maluwa in Ehi or old Sinhalese) to distinguish his name from that of his elder brother, King Bhatiya-Tissa "Tissa-the-elder-brother3 (141-155 A.D.).1

It is clear that a relic-case with a dedicatory inscription of this king of the second century could not be placed in a dagaba which was not erected until the last years of the third century. Its discovery in the northern dagaba proves that this structure was in existence a century before Maha-Sena built the Jetavana dagaba, and is quite fatal to the nomenclature commonly accepted.

The dimensions of the bricks in the domes of the two dagabas also strongly support the identification ; and in fact it was the discrepancy in their sizes compared with others of the respective periods which first led me to investigate this subject more than twenty years ago. Those of the dagaba to the east of the Sela Caitya.have a length of 15-82 inches, a breadth of 8-41 inches, and a thickness of 2-26 inches ; Bt. is 19 square inches, and the contents 301 cubic inches. The table of dated bricks already given proves that they must belong to some date from the first to the fourth century A.D., inclusive.

In the northern dagaba, the bricks measure 18-92 inches in length, .9*62 inches in breadth, and 3-20 inches in thickness;

requisites [clothing, food, bedding, and medicine]. Having constructed the chief dagaba at this place, he repaired the dilapidated buildings.

This inscription is a duplicate of one copied by Dr. E. Muller at Galkowila (No. 98 in his list), the purport of which Ms defective copy led him to misapprehend. The, meaning of the word mata, which I have translated * rejoicings/ is doubtful. Luwimitaya (Lumitiya at Galkowila) = Lokamita; compare Alu wihara = Aldka wihara. ' Protecting' the tank probably means arranging for its maintenance and appointing a Guardian for it.

King Udaya II (952-955) nearly lost his life through his violation of the sanctity of the Ascetics* Forest (Mah., ii, pp. 82, 83), It was a ' Sacred * forest to which -ultra-ascetic monks, especially those termed Pansukulikas, retired, and it was evidently a sanctuary for offenders of all kinds. Its position is not known.

1 In Dr. Mullet's Inscription No. 16, Kanittfaa-Tissa describes himself as * Tissa, the younger brother of the Great King Bhatiya-Tissa, (and) son of the Great King Naga.*304 ' ANCIENT CEYLON '

Bt. is 30-7 and the contents 583 cubic inches. The only dated bricks of this type are those in the Miriswaeti dagaba. While they are large even for the time when Watta-Gamini was king, it is practically impossible that such extreme dimensions should have been suddenly reverted to nearly four centuries later, as would be the case if this were the Jetavana dagaba.

So far as I am aware there is only a single line in the Drpa-vansa (xix, 17) which can be held to support the common nomenclature. This work says of King Watta-Gamini Abhayagirim patitthapesi Silathupam cetiyamantare, {He built Abhaya-giri between the Silathupa and the Cetiya/ Mr. Bell understands this to mean 'between the Sela Caitya and Mihintale' [Cetiya-giri].1 But Mihintale is eight miles away, and there is no authority for assuming that the old author omitted the word gin after Cetiya. I suppose that by the word Cetiya the author meant the Abhaya-giri dagaba, and that the Sila-thupa is the Sila Sobbha Kandaka Thupa, now called Lankarama, which the same king built. The line would then mean that 'he built the Abhaya-giri (wihara) between the Lankarama dagaba and the Abhaya-giri dagaba/ this being the actual position of a great part of the ruins of the monastery.

THE ABHAYA-GIRI DAGABA

Although the histories contain numerous references to the buildings which formed the Abhaya-giri wihara, little is said respecting the great dagaba itself. It is recorded of Gaja-/Bahu I (113-135 A.D.) that 'raising the Abhayuttara Thupa i he constructed it of a greater elevation ' (Mah., i, p. 132), a result that might naturally be expected if he raised it. Doubtless this refers only to the works above the dome, in the lower part of which no bricks of this period are to be seen.

It is possible that when King Kanittha-Tissa placed his relic-cases there he at the same time built the wahalkacjas. The decorative work on their flanking pillars is of a type intermediate between that of the Ruwanwaeli and Miriswaeti dagabas and that of the Jetavana dagaba, in some details 1 Annual Report for 1895, P» 2» footnote.

'-r' °"THE EARLIEST DAGABAS 305

agreeing with the first two and in some with the other.

King Voharaka-Tissa (215-237 A.D.) is said to have caused the chatta to be repaired (Mah., i, p. 144) ; the original work in it must have been done by Gaja-Bahn I. Probably the spire possessed one from the first, resembling in that respect those of the other great dagabas. Dhatu-Sena (463-479 A.D.) again repaired this chatta,

Moggallana presented a new cloth covering to the dagaba, and Kassapa II (652-661 A.D.) fixed a jewelled pinnacle on the spire, the top of which is thus shown to have passed through the chatta (Mah., ii, p. 32).

Sena III (955-964 A.D.) laid a stone paving round the dagaba at a cost of forty thousand kahapanas (Mah., ii, p. 83). It is surprising to find that at this important monastery such a necessary work had been neglected for more than a thousand years. It is safer to assume that the statement refers to some additional work in laying paving than to suppose that all the earlier zealous Buddhist monarchs had omitted to carry out such an obvious improvement. The sum paid for the work also appears to be totally insufficient to cover the cost of laying the whole paving.

The dagaba was repaired by Parakrama-Bahu I (1164-1197 A.D.) with the others at Anuradhapura, all, it is stated, having been seriously damaged by South-Indian invaders long antecedent to that period, that is, in the middle of the preceding century, when they held Northern Ceylon.

It is not specially mentioned as having been broken into during the reign of the invader Magha (1215-1236 A.D.), but as all the other large dagabas were injured at that time no doubt considerable damage would be done to this one also; and it was one of those repaired in the time of Parakrama-Bahu II (1240-1275 A.D.), when all are said to have been in a ruinous state and overgrown with jungle.

This was again its condition in the last quarter of the nineteenth century. It was then surrounded by thick jungle and covered by a forest of trees and interlaced undergrowth. This was felled and there-growth kept under control, but*no repairs have been undertaken yet.

x306

ANCIENT CEYLON

The dome of this dagaba was found by Mr. Smither to be a semi-globe, the centre of which is four feet above the basement or paved court-yard in which the building stands, which is raised six feet above the adjoining ground and is 587 feet square. The diameter of the dome is 310 feet at the top of the basal ledges.

It rests upon the usual three short basal cylinders, which

310 Fee*----

Top of J eta van a Sp«rer

FIG. 85. The Abhaya-giri Dagaba. (The triangle shows the size of the Third Pyramid at Gizeh.)

have a total height of 16 feet, the lowest one being 6 feet 6 inches high and 355 feet in diameter. The tee is 75 feet square and 33 feet high. It has a plain plinth, and a cornice of three plain overlapping bands. Each face has post-and-rail work in sunk relief, with a disk of the sun in the centre, 6 feet 6 inches in diameter. The railing consists of 12 pilasters each 2 feet 3 inches wide, and 14 flat rails, each 15 inches wide. The spire, which is 30 feet in diameter at the base, springs from a cylinder 15 feet high, and 30 feet in diameter; it is verticalTHE EARLIEST DAGABAS 307

for five feet and above that tapers gradually. Bands of cut stone 6 inches thick are inserted in it, with intervals 2 feet 6 inches high. The face of the cylinder is divided by pilasters into eight compartments, in each of which there is a shallow arched niche. The height from the platform to the top of the tee is 187 feet 6 inches, and the spire, which is broken, now rises 57 feet 6 inches higher.1

Four wahalkadas, 45 feet 6 inches long, were built at the cardinal points. Like those at the other great dagabas, they consist of a series of horizontal cornices or prominent mouldings, separated by plain cut stone-work, and were about 16 feet high. Two square decorated pillars were fixed at each end, carved on the faces with straight-stemmed trees having leaves in pairs, or an ornamental meandei springing out of a vase and having animals in its loops. On one pillar the animals are in pairs, one being on each side of the stem of the tree, and are climbing upward. On the side of each outer pillar are two Nagas, or in some cases other deities, in high false relief, in two panels, one above the other, a male above and a female below. These pillars may have been flanked by a slab carved with a multiple-headed cobra in high false relief.

On the south face of the dagaba a small building, measuring 24 feet by 15 feet, was constructed, probably to contain relics or statues.

THE JETAVANA DAGABA

The Jetavana wihara is often mentioned in the histories but they rarely allude to the great dagaba. A reference to the repair of the chatta by King Dhatu-Sena proves that like similar structures at the old city this one had this form of terminal in early times, and most probably from the date of its erection.

The plaster work on the dome was repaired by KingMahl-Naga (561-564 A.D.), together with that on the Ruwanwaeli

1 By the kind permission of the Secretary of State for the Colonies,, I am able to give a drawing of this dagaba, reduced from Mr. Smither's elevation of it, with the exception of the restored spire (see Fig. No. 85).308

ANCIENT CEYLON

and Abhaya-giri dagabas. Moggallana (608-614 A.D.) presented a new cloth covering for it as well as to these two works, and Kassapa II fixed a jewelled pinnacle on it, as on them.

Although the histories do not mention it, a stone pavement was wholly or partly laid round the structure in the tenth century ; and the record is preserved in some short dedicatory inscriptions cut on several of the slabs presented by private donors. One runs Kasa Mmiya taebu pahana,' the stone placed by the Lady Kasa.' Another is Nagd himiya panas pahanak»

FIG. 86. The Jetavana Dagaba. 1886.

' the Lady Naga, fifty stones/ A third states that the cost of two stones given by Kakawannaya was ran dasa kaladaf * ten kalandas of gold *; a fourth inscription records that a stone laid by another person also cost ran pas kalanaek, * five kalandas of gold'?the same price.

The dagaba must have been damaged in the eleventh century, since Parabrama~Ba.hu I restored it. In the time of Migha it suffered like the rest, and was again repaired during the reign of Parakrama-Bahu II. In the latter part of the nineteenth century it was in nearly the same ruinous state as the other two great dagabas.THE EARLIEST DAGABAS 309

The Jetavana is the widest completed dagaba in Ceylon,* the diameter of the dome at the upper basal ledge being 325 feet. The dome is a semi-globe, or rather part of one, as the centre of the globe was found by Mr. Smither to be 9 feet 6 inches below the base of the lowest of the three cylinders on which it rests. The platform round the dagaba being raised to this extent, the centre of the globe is at the original ground level.

As at the other large works, the three short basal cylinders rise 16 feet above the surrounding paved platform, and the diameter of the lowest one is 367 feet. They form three steps or ledges round the base of the dome, each 7 feet wide. The lowest one is 6 feet 9 inches high, the middle one 4 feet 9 inches t high, and the upper one 4 feet 6 inches high. The paved platform is 590 feet square.

The dome is surmounted by a ' tee/ 76 feet square and 32 feet 6 inches high, the distance from the platform to its top being 183 feet. The spire, 33 feet in diameter at the base, appears to have risen directly from the tee, with possibly at first a slightly narrower neck ; it tapered to 24 feet in diameter at 48 feet in height, the point where the top was broken off. The height from the platform to the top of the spire was 232 feet when Mr. Smither measured it. Bands of cut stone 6 inches thick were laid in the spire at intervals of 2 feet 6 inches.

The tee had the usual post-and-rail work on each face, the posts being n brick pilasters 2 feet 6 inches wide, while flat horizontal bands 15 inches wide form the rails. A sun-disk occupied the centre of each face.

At the cardinal points there are four wahalkacjas, each 48 feet long, flanked by two rectangular pillars at each end ; the

1 A still larger one was commenced at Polannaruwa by Parakrama-Balm I, but was not finished. The mound which constitutes its remains, to the north of the other dagabas there, was about 50 feet high and some 350 feet wide at the top when I examined it many years ago. The Mahavansa (ii, p. 259) calls it the Damila Thupa, * the Tamil Dagaba,' because it was partly built by Tamil prisoners of war. According to that work it was * the greatest of all the Thupas,' and was thirteen hundred cubits in circumference, or about six hundred feet in diameter at the bottom, if the cubit of that time 17! inches in length. The diameter of the dome is not stated.3io ANCIENT CEYLON

face work consists of alternate cornices or heavy mouldings and plain stone work. Elephants* heads project above the lowest cornice. The end pillars are decorated in the face by ornamental creepers, or Buddhist emblems or animals placed in the loops of a meander ; and on the outer side of the lower and outer pillar by Naga princes and princesses in false high relief in two panels, one over the other, the prince being in the upper one and the princess below him in the otherI (Fig. No. 8). At each end of the wahalkadas, and beyond these pillars, there is a limestone slab carved with a seven-headed cobra in high sunk relief, its body forming two loops on each side ?; a chatta or umbrella is usually carved above it. The Nagas and cobras were expected to act as guards of the relics, or the whole structure; according to the Dhatuvansa, Buddha on his third visit to Ceylon ordered the Nagas to protect his relics. A relic-room 26 feet by 18 feet was built at the western wahalkada.

In 1887 and 1888, the late Mr. R. W. levers, at that time the Government Agent of the Province, opened a horizontal tunnel to the centre of this dagaba, at a level of 33 feet above the surrounding pavement, and at the end of it sunk a shaft in the axis of the structure to a depth of 13 feet below the base. At a depth of 40 feet, or 7 feet below the pavement, a rough stone slab was encountered, under which was a small copper coin, having an animal, apparently a horse, on the obverse. It was stated by Mr. R. S. Poole, of the British Museum, to resemble the coins numbered 55 and 58 of Plate II of Sir Walter Elliott's Coins of Southern India, which were attributed to the Korumbars. It must have been placed there when the work was begun. The excavation of the shaft proved that in the centre the foundations only extended to a depth of 3 feet 6 inches below the ground-level. The whole inner work was found to consist of bricks set in a

1 They are only distinguishable from human beings by the cobras' heads which appear above or at the side of their heads, there being a five-headed cobra with the prince and a single-headed one for the princess. Some figures without these emblems may be intended for other

deities (male and female) and one may be Ayiyanar (Fig. No. 37).FIG. 87. Pillars at Wahalkacla, Jetavana Dagaba.THE EARLIEST DAGABAS 311

tenacious clay. No relic-chamber was discovered.1 A later excavation made by Mr. Bell showed the foundations to. be 26 feet deep; in brickwork, underneath which was concrete. This was at the periphery.2

THE LANKARAMA DAGABA

In addition to the Abhaya-giri it is recorded in the Mahavansa (i, p. 132) that King Watta-Gamini also built a dagaba to the north of the Ruwanwaeli' on an eminent place [that is, eminence] which was named Sila-Sobbha-Kandaka.' The only other dagaba of any importance, now termed the Lankarama dagaba, to the north-west of the Ruwanwaeli dagaba, appears to be the work in question. It seems to have been held in less estimation than the other great structures of thte kind, and if it is again mentioned in the histories the monastery at which it was erected was known as the Manis5ma wihara, or Soma-rama, which the same king is stated to have built. (Mah., ?,

P- 131).

We find the Manisoma wihara mentioned with the Thupa-rama, Miriswaeti, and DakkHna monasteries, this connection indicating that all were at Anuradhapura. Kanittha-Tissa is said to have constructed an 'edifice' over the Manisoma dagaba (Mah., i, p. 143), and Meghavannabhaya I (254-267 A.D.) made some repairs at this ' edifice/

The length and width of the bricks used in the dome of the Lankarama dagaba shows that they are pre-Christian, notwithstanding their reduced thickness. Their length is 17-37 inches, the width is 8-94 inches, and the thickness is 2-62 inches ; Bt. is 23-4, and the voliime becomes 407 cubic inches.

It is possible that the original shape of this dagaba differed from that of nearly all the other early dagibas in Ceylon, but resembled some early Indian works. In its present state the dome is a segment of a hemisphere 38 feet wide at the point where it leaves the upper cylinder of the three on which it rests, above which it rises 15 feet to the tee. This cylinder, 4

1 Manual of the North-central Province, by R. W. levers, 1899, p. 238.

2 Animal Report for 1894, p. 2.312 'ANCIENT CEYLON

feet high, is no wider than the dome, and thus does not form a ledge round it as in the other early works, but is in reality a continuation of the dome, with a vertical side. Of the other two cylinders the lowest one is 44 feet 2 inches in diameter and 3 feet 8 inches high, and the middle one 40 feet wide and 4 feet i inch high.

The tee was 9 feet 8 inches squaref without the plinth, and like the others was probably originally faced with post-and-rail work, although in an illustration taken from a photograph, in Fergusson's History of Indian and Eastern Architecture, published in 1876 (p. 194), it is shown with only two pilasters at each side of the face and a circular sun emblem in the centre. If its width then was the same as when Mr. Smither measured it in 1877 its total height was some 8 feet. It is not known if the spire had a chatta. As the dagaba is only twice mentioned after its erection the omission of the historians to refer to such a«terminal cannot be taken to prove that it had not this usual ornament, which was also a magical protection from evil. The height of the dagaba was found by Mr. Smither to be 33 feet 7 inches; he thought it may have been more than 50 feet when complete.

Like the Thuparama dagaba and the Ambatthala dagaba at Mihintale, this work is surrounded by disengaged thin graceful monolithic pillars, with separate prnamental capitals. While the Thiiparama dagaba had four rows of them that at the Lankarama had only three. The inner row consisted of 20, square at the lower part and octagonal above, each 16 feet 8 inches high ; the middle row had 28 similar pillars, 16 feet ii inches high, their tops being on a level with those of the inner line, on account of a drop in the flooring; and the outer row 40 octagonal pillars, 12 feet 5 inches high. All measure from 11 to i2j inches across, and they have a general resemblance to those at the Thuparama, but they have no tenons at the top. The animals carved on the capitals are homed lions in the inner row, sitting lions in the middle one, and dwarfs in the outer row, some of them playing flutes while others dance.

Although 'Mr. Smither knew of only the three above-THE EARLIEST DAGABAS T 3*3

mentioned structures surrounded by stone pillars, there may be a few others in Ceylon, but they are undoubtedly rare. It is possible that some may have had wooden pillars.

THE DAKUNU DAGABA

There was only one other monastery in Anuradhapura at which a dagaba of considerable size was built at an early date. This was the Dakunu or Dakkhina wihara, whose dagaba lay at the southern extremity of a curved north and south line passing from the Abhaya-giri, at the northern end, by the Thuparama, and Ruwanwaeli dagabas. Until Mr. BelFs excavations disclosed its real character it was merely a high tumulus-like mound completely overgrown with bushes and trees, and popularly supposed to mark the grave of the Tamil king Elara, who was killed by Duttha-Gamini at the capture of the city; although the Mahavansa (i, p. 99) says clearly enough that he was cremated and a tomb was built over his ashes at the spot where he fell' near the southern gate of the city/ the position of which has been already pointed out.

The account of the construction of the monastery is contained in the following words (Mah., i, p. 132), 'Of the eight warriors [chiefs of Watta-Gamini] the one named Uttiya built to the southward of the town the wihara called Dakkhina wihara/ This statement is repeated in the Dipavansa (p. 210) ; and it will be observed that the erection of the dagaba is not alluded to ; but this does not prove that it was not built then, as these two histories similarly do not record the building of the much more important Jetavana dagaba.

Kanittha-Tissa (165-193 A.D,) is stated (Mah., i, p. 143) to have constructed a covering for this dagaba; but in the Dipavansa he is said to have built the dagaba. There is thus an element of doubt as to the exact age of the structure, which one would rather expect to have been erected when the wihara was established.

The other references to it are found in the Mahavansa. King Vdharaka-Tissa (215-237 A.D.) is recorded to have caused the3I4 ANCIENT- CEYLON

chatta of the Dakkhinamula dagaba, which appears to be this one to be repaired along with those at some other important structures ; and the sub-king Kassapa is stated to have destroyed and rifled the dagaba during the reign of AggabodHi III (624-640 A.D). He appears to have restored it again after he succeeded to the throne.

In such a case as this we must look to the bricks for some light upon the question as to the period when the dagaba was really built. They average 16-36 inches long, 8-18 inches wide, and 2-31 inches thick; Bt. is 18-9, and the contents 309 cubic inches. It will be seen at once that these dimensions closely agree with those of the Jetavana dagaba; and they also unmistakably indicate a date after the Christian era. For this reason I accept the statement of the Dipavansa that the dagaba was built near the end of the second century A.D-This is the more likely since no dagabas of any considerable size appear to have been constructed by others than the monarchs of the country.

As to the similarity of the sizes of the bricks at this work and the Jetavana dagaba, which was erected fully a century later, it may be surmised that the dimensions adopted late in the second century continued to be employed, with slight variations, throughout the third century.

This dagaba has only been partly excavated by Mr. Bell. In his Report for 1898, p. 5, he refers to three basal platforms ? round it and mentions that traces of one of the ornamental had been found. A flanking pillar belonging/ to one of has the usual male and female guardians on one.-

in relief, and on another face a decorative tree sur-

by a bird with a crest and raised wings, like those oa the at the Thuparama dagaba. Mr. Bell states that

the circumference of the dome, which is bell-shaped, is 464 approximately; thus its diameter was about 148 feet at the top of the basal platforms. At the base of the outer-the diameter was about 179 feet 6 inches. There ? were m stone pillars round the dagaba. The tee was '

38 feet or very nearly a quarter of the diameter ,'

of theTHE EARLIEST DAGABAS 315

THE KIRIBAT DAGABA

At three and a half miles north of the Ruwanwaeli dagaba, and thus outside the city, there is another now termed the Kiribat, ' Milk-rice/ dagaba. Although Mr. Bell considered it to be * one of the oldest of the large dagabas at Anurar dhapura/ x the size of the bricks dug out of the shaft that was sunk by him down its axis is conclusive evidence against its early date, if they are similar to those in the rest of the structure. They are 6*80 inches wide and 2*26 thick, Bt. being 15-4. These are the dimensions of bricks used in the repairs of the * tee * at the little Pabulu dagaba at Polannaruwa, and they probably belong to nearly the same period as the ruins of the so-called Wijayarama monastery, which is near the Kiribat dagaba. Mr Bell has proved that this monastery is a work of the ninth century A.D.2 The dagaba is now a mere tumulus. Mr. Bell found that it was built on a raised platform 204 feet square, paved with bricks. It had the usual three basal platforms, each about 2 feet 6 inches high and wide ; and it had a diameter of 135 feet.

OTHER EARLY DAGABAS THE MAHIYANGANA DAGABA.

The Mahavansa (i, p. 5) records that a dagaba 30 cubits high was constructed by Uddhachulabhaya, a younger brother of Devanam-piya Tissa, at Mahiyangana, to enclose one 12 cubits high erected there by ' Sarabhu, disciple of the Thera Sariputta' immediately after the death of Buddha, in which was enshrined a collar bone (glvatihi) of the Teacher. This latter dagaba was said to be built over one formed of emerald, which was supposed to hold a handful of Buddha's ' pure blue locks/ presented by him to the god Sumana, ' the chief of the devas/ that is, Sakra, on his visit to convert the Yakkhas.

Duttha-Gamini during the early part of his war against the forces of Elara enclosed these in a larger dagaba 80 cubits high. The chatta was repaired by VohSraka-Tissa (215-237, A.D.),

1 Annual Report for 1892, p. 5.

31 Arch. Survey of Ceylon. Sixth Progress Report, p. 9.3I6 ANCIENT CEYLON

I have not visited the place, and the only account which I have of the dagaba, is the short one given of its state in 1848 by Sir Emerson Tennent, in which he says,' It is a huge semicircular mound of brickwork, three hundred and sixty feet in

FIG. 88. The Mahiyangana Dagaba. (From a photograph by DR. C. G. SELIGMANN.

circumference and still one hundred feet high, but so much decayed at the top that its original outline is no longer ascer-tainable. When Spilberg, the Dutch admiral, saw it on his way to Kandy in 1602 it was comparatively perfect, as white as marble, and surmounted by a gilded pyramid/ 1

The Mahavansa merely states that it was re-plastered by

King Parakrama-Bahu VI (1410-1462 A.D.). Through the kind-

of Dr. C. G. Seligmann I am able to give an illustration

taken from his photographs of this dagaba, "which shows its

present state (Fig. No. 88).

THE KAELANIYA DAGABA

Another very early dagaba was erected at Kaelaniya wlhara by King Yatthala-Tissa, nephew of Devanam-piya Tissa,

before the end of the third century B.C. (Raj., p. 24).

It commemorated the site where Buddha was feasted by'

the Naga king of Kaelaniya. .

1 Ceylon, 2nd Ed., Vol. ii, p. 421.THE EARLIEST DAGABAS 317

Although the monastery is occasionally mentioned in the histories, I believe the dagaba is only thrice referred to. The chatta was repaired by Voharaka-Tissa (215-237 A.D.). The dagaba was seriously damaged in the reign of the Tamil king Magha (1215-1236 A.D.), and was restored by Wijaya-Bahu III (1236-1240 A.D.). His son Parakrama-Bahu II (1240-1275 A.D.) then paved the court-yard, or part of it, round the dagaba.

The only other record with which I am acquainted is contained in Dr. E. Miiller's inscription No. 162, in which it is stated that King Dharma Parakrama-Bahu of Kotta, who

FIG. 89. The Kaela^iya Dagaba. -fa.

according to it began his reign in 1508 A.D., caused the dagaba to be restored and plastered.

The dagaba was again restored in its present form in 1779, probably, so far as regards the dome, according to its original shape. Its outline is of the type technically known in Ceylon as the ' Heap-of-Paddy' shape. It forms the end section of a wide cone with slightly convex sides, and is perhaps the earliest example of this class of dagabas. It has no wahal-katjas. For'the following particulars I am indebted to measurements and photographs which Mr. R. S. MacPhail, of the Irrigation Department, was good enough to obtain for me.3i8 ANCIENT CEYLON

The dagaba rests on three narrow circular basal platforms, the diameter at the base of the lowest one being 106 feet io|-inches. The top of this platform is 2. feet 6 inches less in diameter. The top of the middle platform is 98 feet in diameter, and that of the upper one is 92 feet i| inches in diameter. The lowest one is 3 feet 5 inches high, the middle one 4 feet 3 inches, and the tipper one 4 feet 2 inches.

The total height is 88 or 90 feet,1 the former being the measurement on a photograph, and the latter being calculated from the shadow of the structure. The dome is 85 feet 7 inches in diameter at the top of its basal moulding, which is 2 feet high; it is 46 feet 8 inches high by the photograph, according to which also the tee is about 19 feet 6 inches wide and 4 feet 9 inches high. The base of the spire, similarly measured, is 3 feet 3 inches high and 13 feet 6 inches wide, above which the spire and its brass pinnacle, which is terminated by a glass point, rise 21 feet 6 inches. The illustration (Fig. No. 89) shows the general shape.

THE DiGHA-VAPi DAGABA

A large dagaba was built at Digha-vapi by Saddha-Tissa, the brother of Duttha-GaminL This was a very important station in south-eastern Ceylon, where Saddha-Tissa was stationed for a considerable period before he became king. The dagaba was built to mark the spot where Buddha seated himself on the occasion of his last visit to Ceylon. It is not mentioned again in the histories.

There is said to be a large dagaba in the neighbourhood of the tank now called Kandiya-Kattu, which is almost certainly the ancient Digha-vapi. I have not had an opportunity of visiting the place, which, however, is likely to repay the trouble of an examination.

Major Forbes 2 quoted a note of Bertolaccf s regarding it, according to which the ruin was discovered in 1810. 'The size of the building is gigantic . . . the cone forming the

1 The resident monk stated the height to be 60 wadu-riyanas, or

'carpenter's cubits,* an evident mistake for riyanas, or 'cubits.*

2 EUven Years in Ceylon, Vol. i, p. 153, footnote.THE EARLIEST DAGABAS

319

pagoda [dagaba] is entirely covered with brick and mortar ; its basis is about one quarter of a mile in circumference, and the top and side are planted [overgrown] with large trees/

THE IDIKATU DAGABA

King Kalakanni-Tissa (40-20 B.C.) constructed ' a beautiful stone Thupa' in front of a hall which he built ' near the Mihintale monastery/ The structure now called the Idikatu (' Needles/) dagaba, at the base of the Mihintale hill, is believed to be this dagaba, although the upper work is of brick. The

-r. ?-""????--C"^.,

Moulding of

$0* Enclosure

FIG. 90. The Idikatu DJLgaba.

historians do not mention its original name, and I believe never afterwards refer to it. In a fine inscription cut on two upright slabs at the side of the great flight of steps * which lead up the Mihintale hill, MahindalV (975-991 A.D.) termed it the Katu Maha Saeya, and ordered its repair.

It is biiilt in a small stone-paved court-yard, measuring about 38 by 40 feet, which is raised five feet above the level, and supported at the sides by an excellently cut revet-*

i Some books state that these steps number 1500, bui: thisis erroneous. That number is only applicable to the whole of the steps on all parts of the hill, the monks say.320 ANCIENT CEYLON

ment wall, with bold mouldings. A quadrantal stone moulding i6| inches high encloses the floor of this court.

This dagaba was evidently of the modern bell shape, and it is possibly the earliest one of this form which is known in Ceylon. If, as appears probable, the upper part was hemispherical, the height to the top of the dome may have been 16 feet. For 6 feet 10 inches the face-work consists of a series of well-cut stone mouldings. There are no basal platforms, but three are clearly indicated by mouldings which project beyond the others. Above this the structure is of brick. The diameter at the base is 27 feet 4 inches, and at the top of the mouldings about 18 feet. Nearly all the upper part is broken down, as shown on my sketch. As usual, rectangular stone flower altars were fixed at the four cardinal points, close to the base of the mouldings. A single flight of steps 6 feet 10 inches wide, on the west side, led to the dagaba platform.

Around this building there are the remains of a large and very regularly arranged monastery, the dagaba being near the south-eastern corner of it. The whole is enclosed by a well built wall of uncoursed stone, four feet thick, with a coping of stone laid transversely. All the stones are wedged and slightly cut, the lower ones being squared and dressed on the beds and joints. Such a work as this wall is quite exceptional in Ceylon, nearly all the enclosing walls of monasteries being built of brick. In the illustration the Maha Saeya and the Aet dagaba are to be seen on the hill in the background.

THE AMBATTHALA DAGABA

This dagaba on the Mihintale hill was erected by King Maha-dathika Maha-Naga (9-21 A.D.), and is believed to mark the spot where the Buddhist apostle Mahinda, the son of the Emperor Asoka, stood when Devanam-piyaTissa first saw him. A headless and armless statue near it, facing the dagaba, is traditionally said to represent the king and to mark his own position on that occasion. It has no ornaments on the chest or waist, and the sole clothing is a plain cloth from the waist to the ankles. When Mr. Smither examined it the head was there, and he wrote of it (p. n) that the * head-dress consists ofTHE EARLIEST DAGABAS 321

a plain and slightly elevated pear-shaped cap encircled by a jewelled band or diadem ; the ears are adorned with pendant ear-rings, and the neck with a jewelled neck-piece. The base is carved to represent an expanded lotus flower and it is precisely similar in design to that found at the Thupa-rama dagaba/ Three octagonal pillars round it evidently supported a canopy over it.

It is recorded that after building the dagaba Maha-Naga held a great festival at which festoons of lamps were hung round the dagaba. FIG. 91.

Twenty-four thousand Tlie Ambatthala Dagaba. ^.

monks are said to have

been present at this great celebration (Hah., i, p. 136).

King Kanittha-Tissa constructed an e edifice' over this dagaba; possibly this was repaired by Meghavannabhaya I (254-267 A.D.). We may assume that the structure was rifled during the Tamil domination in the eleventh and thirteenth centuries, and that it was repaired, like the other principal dagabas, by the two Parakrama-Bahus. It is now completely restored, and the early work is covered by a coat of plaster, like that at the Thuparama dagaba. It is said to be built of stone. Its shape is intermediate between the Bell and the Heap-of-Paddy.

According to Mr. Smither its present dimensions are as follows :?The diameter of the dome at the top of the basal platform, or plinth, on which it rests is 23 feet, the plinth being 6 feet wider at its base. The height to the crown of the dome is stated by him to be 20 feet, but it measures only 18 feet on a photograph by Messrs. Skeenand Co., of Colombo, from which Fig. No. 91 is reduced. The height of the square tee is 3 feet in the photograph, and its side, exclusive of mouldings, 6J feet long. The spire rises about 9! feet higher, of which its cylindrical base occupies 3 feet 9 inches. The dagaba322 ANCIENT CEYLON

rests on a platform 97 feet in diameter, which is raised four feet above the ground level, and is supported by a brick wall. There are no wahalkadas ; these ornamental structures were confined to the five Anuradhapura dagabas already described. This dagaba is surrounded by two circles of slender monolithic octagonal pillars, 12 feet high, the capitals of which are decorated with a procession of lions or hansas marching to the left (which would be their own right when facing front), or a row of dwarfs facing front. They have no tenons on the top ; and whether the roof afterwards raised over the building rested on them or not, they appear to have been erected chiefly in order to carry the festoons of lamps that were hung from them at festivals. In shape they bear a close resemblance to those of the Thiiparama dagaba.

THE AET DAGABA

On the summit of the highest rocky point on the Mihintale hill a small structure, called the Aet (Tusk-elephants) dagaba was built by an early ruler whose deed is not specially recorded in the histories. There are, however, some references to the dagaba in inscriptions.

On the side of a slab at the Thiiparama dagaba an inscription in letters resembling those used by Wasabha (66?no A.D.) ends Ati c(e)tahi paca jar a dini,t"he meaning of which appears to be ' At the Aet dagaba he gave (anew) the former decayed (work)/ The name of the king is not visible on the stone. The words are quite clear, and the first two prove that this structure was already in existence when the inscription was cut.

Another very much worn inscription on the upright face of a rock near the dagaba, on the side of the path to it, cut in letters resembling those of the large rock inscription of King Gamirti-Abhaya,* the son of a king called Naga and grandson of one called Tissa, lower down the hill, ends c(e)ta padaya dim, ? He gave the dagaba steps/ It appears to refer to the last of steps for ascending to the platform of the dagaba. The name of the king who made the grant is ' Naka Maharaja '; he may be the Siri-Naga II (245-247 A.D.) who is 1 Dr. E. Muller's Ancient Inscriptions in Ceylon, No. 20.THE EARLIEST DAGABAS 323

mentioned as the father of the king in the lower inscription, if Dr. E. Mfiller's identification of its author as Meghavanna-bhaya I is correct.

The repair of the dagaba is referred to in the long inscription of Mahinda IV,1 which was specially devoted to recording the king's orders regarding the management of the services and funds of the Aet wihara, at that time evidently a very important establishment. We learn from it that the Aet and Idikatu monasteries were connected, and were held by monks who belonged to the Abhaya-giri Community. They were independent of the other two large Mihintale wiharas at which were the dagabas previously described.

On such a high and exposed site the dagaba would be extremely liable to be damaged by thunderstorms, and it is not surprising to find bricks of three or four periods employed in the lower part of the work. Some are of almost the same size as those in the inner room of the building termed the Dalada Maligawa, 'the Palace of the Tooth-Relic/ at Anura-dhapura, and thus may have been burnt early in the fourth century, possibly by Meghavannabhaya II, who is described as evincing great interest in the Mihintale monasteries.

I have met with no reference to the construction of this dagaba, unless it is one of the ten which were built on the hill by Wasabha, according to the Dipavansa (p. 216). There are a few brick fragments at it which are three inches thick; if they were burnt for it they must indicate that It is of pre-Christian date, but they may have been brought from some ruined building for use in the repairs.

I have no measurements of its dimensions ; it is a very small work, and only of interest on account of its situation and Mahinda's inscription.

THE TISSA DAGABAS

Regarding the group of early works?the Mahanaga, Yatthala, Sanda-giri, and Maenik dagabas?built at Tissa, the ancient Magama, in the Southern Province 2 little definite

1 Muller, op. cit., No. 121.

2 See Fig. No. 130 for their positions.324 ANCIENT CEYLON

information is available. The original shape of the upper part of their superstructures is lost, but on account of their early date it may be assumed that all, with the exception of the Maenik dagaba, had more or less hemispherical domes, with the usual tee, and a spire with a chatta terminal. All are without wahalkadas or surrounding pillars.

The dimensions of the bricks used in these four dagabas and another small unnamed one near the high level channel from the tank, throw some light on the period when they were built. They are given in inches in the following table.

Length. Width. Thickness, Bt. Contents.

Mahanaga dagaba . 17-35 8-84 2-83 25 434

Yatthala dagaba . 17-85 8-64 2.90 25 447

Small dagaba . 17-16 9-11 2-85 26 445

Sanda-giri dagaba . 17*H 8-67 2-81 24-4 418

Maenik dagaba . 16-57 8-86 2-80 24-8 411

It is clear that the sizes of the early bricks used at Tissa were much more uniform than those at Anuradhapura; and this being so they may indicate that the Maenik or ' Gem ' dagaba was erected by the king who constructed the Sanda-giri, the ' Moon-hill/ or his son; and that probably both these works are of somewhat later date than the first two in the list. The evidence afforded by the bricks is thus in favour of the statement in the Pujavaliya that Sanda-giri dates from the time of Kakavanna-Tissa, lEat is, early in the first half of the second century B.C. The Dhatuvansa, however, attributes the construction of the Sanda-giri wihara to King Maha-Naga. The later monarch may have added the dagaba to it.

THE MAH!NAGA DAGABA

The first dagaba constructed there was undoubtedly the largest one, at the Naga Maharama monastery, which was built by King Maha-Naga, the brother of Devanam-piya Tissa. The Mahavansa records his founding of this wihara ' bearing his own name ' (i, p, 83); it was therefore built early in the second half of the third century B.C. If I remember aright, I was informed that it enshrined the right temple bone of Buddha,

' '

/"U-x VvW * y\ 'c !,THE EARLIEST DAGABAS

325

but the Dhatuvansa, which gives many particulars of the Tissa works, does not contain this statement.

The superstructure of the dagaba was repaired or raised by King Ila-Naga (38-44 A.D.), who left an inscription recording his work on a stone that is now replaced in the dagaba, but was previously seen by Dr. Paul Goldschmidt, who copied the inscription.1

FIG. 92. The Mahanaga Dagaba.

King Voharaka-Tissa (215-237 A.D.) repaired the chatta on the spire (Mah., i, p. 144). During the Sola domination in the eleventh century this and the other dagabas at Tissa 1 that belonged to the three Fraternities * were broken into. The next Sinhalese king, Wijaya-Bahu I (1065-1120 A.D.) restored them.

They all appear to have been rifled again, more thoroughly than on the previous occasion, during the reign of the Indian king Magha (1215-1236 A.D.), and apparently were not afterwards restored. The tradition of this desecration has been preserved locally.

In 1883, when I first visited Tissa, the restoration of the Mahanaga dagaba had made considerable progress under the supervision of the incumbent of the wihara, by means of the subscriptions of the numerous pilgrims who flocked there from" the Southern and Western Provinces at the annual festivals; 1 Dr. E. Miiller's Ancient Inscriptions in Ceylon, No. 4.326 ANCIENT CEYLON

and the outer shell of the dome was already rebuilt. It had three basal ledges, above which it appeared to be nearly vertical for several feet in height, the whole dome being apparently almost a hemisphere. A square tee was afterwards built, and a spire was being added when I last saw the work (see Fig. No. 92). This was subsequently completed.

The method adopted for raising the materials for the superstructure was probably the same primitive one which was employed by the pre-Christian constructors. A long ladder of sticks and bamboos was erected up the surface of the dome, and on this a continuous line of men stood, each receiving the materials?bricks or mortar?from'the man below, and handing them to the man above, without moving from his post. At one time about seventy men were employed in this chain, all working without remuneration for the sake of acquiring ' merit' which would beneficially affect their prospects in their next existence. I suggested the use of a winch fixed on the pavement, but the old-fashioned method of their ancestors was adhered to.

On some of the bricks of the largest size I found letters engraved by the brick-makers before they were burnt. These were of the angular type which marks the earliest period of writing in Ceylon.

Through the kindness of Mr. T. Hamer, of the Irrigation Department, I am indebted to the Irrigation Guardian at Tissa for the following dimensions of this dagaba as now restored.

The diameter at the base is 164-5 feet. Above this there are three cylindrical platforms forming narrow steps round the dome; the lowest step is 6 feet high and 4 feet wide, the second one 4 feet high and 3! feet wide, and the upper one 2 feet high and 4 feet wide. The diameter of the dome, deduced from the circumference, like other diameters given, is 140-7 feet. The dome is 86 feet high. The tee is very wide, and is 60 feet square, a size which agrees with my own photograph. It is said to be 20 feet high, but in my photograph measures only 15 feet. The base of the spire is 53-4 feet in diameter and 15 feet high. The spire tapers for 41 feet, being 45-8 feet in diameter at the bottom. The gilt finial is n| feet high. TheTHE EARLIEST DAGABAS 327

total height thus becomes 185! feet. The height of the dome closely follows the Canon given below, which would require it to be 84-4 feet high.

THE SANDAGiRiDlGABA, the next in size to the Mahanaga dagaba, has become a mere high tumulus-like mound partly overgrown with jungle and trees. I have no details of it.

THE YATTHALA DAGABA

I possess no measurements of the Yatthala dagaba, the repair of which was begun when I was at Tissa. It is somewhat larger than the Maenik. dagaba. Its construction is credited by the Dhatuvansa to King Maha-lSfaga, and the size of the bricks used in it affords some corroboration of this statement. Possibly it %as completed by his son Yatthala-Tissa, with whom the name seems to connect it.

This dagaba had a deep vertical cut made by treasure seekers; it passed entirely through the upper part of the dome, which appeared to be of the usual hemispherical shape. The whole inner work thus disclosed showed no mark of a rebuilding such as is recognised everywhere by the employment of a large proportion of broken bricks, or bricks of a later date than those used in the original work. All were whole bricks of the same large size ; and on many of them letters of the very earliest type, with the long attached vowels, some of which perhaps do not occur elsewhere in Ceylon with letters of this early form, were inscribed, or in a few instances impressed by wooden stamps, before "they were burnt. A series of these bricks was sent by me to the Colombo Museum.

At first I supposed that these letters were the initials of the various brickmakers ; but now I feel little doubt that they were the initials of pious persons who paid for their manufacture and presented them to the builders of the dagaba, as an act of religious merit. If stamped letters and letters written in quite different styles are the initials of separate names, I found more than eighty persons represented.

When the restoration of the Yatthala dagaba was begun in 1883, the surrounding debris was first removed, and in it were found several articles of great interest, which had been taken328 ANCIENT CEYLON

out of the relic-chamber by the treasure seekers, and had evidently been thrown aside as valueless. That they formed part of the original contents of the relic-room, as it was left by Maha-Naga or his son, is rendered most probable by the fact that among them were two small moderately thick Puranas, or silver coins of the early Indian type, one nearly square and the other of a more irregular shape, without punch marks. They resembled some unpunched coins of this kind found in India, which date from a period considerably antecedent to the Christian era. There was also the greater part of an engraved carnelian gem, which had been set in a signet ring, and which is considered at the British Museum to belong possibly to the third century B.C., and to be certainly pre-Christian. I describe and illustrate it below, in a chapter on the earliest coinage of Ceylon (see Fig. 156).

The other articles found at the dagaba were four small relic-caskets or karanduwas, cut from gems. One (Fig. No. 93) was a chrysoberyl, another (Fig. No. 96) an amethyst or purple crystal, and the other two were of rock crystal, one (Fig. No. 94) being brownish in colour, and the other (Fig. No. 95) quite clear. They are all now replaced in the new relic-chamber, the restoration of this dagaba having been completed some years ago; but I was able to make drawings of them which are of value as illustrating some of the early types of dagabas.

Each karanduwa represented a dagaba with either one or two basal platforms, but only two were provided with a tee. The smallest one was especially valuable as the only example in Ceylon of a dagaba with a spire surmounted by a chatta, the horizontal extension of which is of course exaggerated in this stopper. The brown crystal had a lathe-turned stopper forming a plain spire of circular section, tapering so as to fit accurately into the cylindrical hollow which is drilled for receiving the relics. The stoppers of the other two karancjuwas were not discovered.

Each relic-case is provided with a tubular well, drilled into the stone from the top, as shown in dotted lines in the illustrations ; and in the smallest one two flakes of gold in which the relics were enveloped, making small packages about the size of aHalf Actual Sizes.

FIGS. 93-110. Articles deposited in Dlgabas. - r , ,. *

32933o ANCIENT CEYLON

grain of wheat, were found in situ, although the relics themselves had disappeared?miraculously, according to the opinion of the Buddhist monks of the place, who with many others believed that true relics of this nature are only visible to arfiats and royal personages.

THE MAENIK DAGABA

The Maenik dagaba has been restored in recent years, and I was informed by the Committee in charge of the work that the ancient dimensions would be adhered to. The Irrigation Guardian at Tissa has been good enough to send me the following particulars of its size.

It rests on three basal cylinders as usual, which form steps round it; the lowest one is 3 feet high and 2 feet wide, the middle one 2 feet high and 2 feet wide, and the upper one I § feet high and 2 feet wide. The diameter of the lowest one is 60 feet. The base of the dome is 51-9 feet in diameter. The lower part of the dome seems to be vertical for a short distance, up to a fillet or band which passes round the dagaba ; above this it is probably hemispherical. The tee is 12 feet square and 5 feet high; the cylindrical base of the spire is 6 feet high, and the finial is 8 feet high. The total height is said to be 80 feet.

I believe that the construction of this dagaba is attributed to Duttha-Gamim; the size of the bricks used in it does not contradict this date. He or his father may have built it.

THE SOMAVAT! DAGABA

According to the Dhatuvansa, Abhaya, king of the Giri district and son-in-law of Kakavanna-Tissa, built a dagaba in Ms part of the hill region, and called it the Somavati

dagaba, after his wife SomadevL

Its site is unknown. The old writer did not mention its size, but he stated that it had the usual three basal platforms, and was of the * Bubble' shape, that is, hemispherical.

THE SIRUVILA DAGABA.

The greater part of the Dhatuvansa is devoted to the history of the relics deposited in this dagaba, and to an accountTHE EARLIEST DAGABAS 331

of its erection, and the ceremonies held in connection with it. It is near the right bank of the Mahawaeliganga, and to the north of the Verukal branch of that river. It was constructed by Kakavanna-Tissa to enshrine the forehead relic (Mdta) and a hair relic of Buddha. Although there are now no Buddhists in that district, the inhabitants of which are all Tamils, or the so-called ' Moormen/ * some Sinhalese of other parts of the island still make pilgrimages in order to worship at this site.

According to the old work, which no doubt preserves, even although it considerably amplifies, an older account that, from the quotations, was evidently written in the Pali language, the forehead relic was first brought to Tissa in the reign of Maha-Naga, who erected a relic-house for it in the neighbourhood of his palace?' neither near nor far * away. It remained there until the last years of Kakavarma-Tissa, who was informed that it was his duty to fulfil a prophecy that he would enshrine it in a dagaba at Seruvila, then the capital of a subsidiary king?doubtless one of the c Paramakas ' of the early inscriptions.

He and his queen Wihara-Devi proceeded to the spot in a magnificent procession in order to carry out the work, after first handing over the charge of the government to his son Duttha-Gamini. In order to fix upon the correct position for the structure he resorted to a peculiar device. Two pairs of bulls were decorated with flowers and allowed to proceed alone in the jungle. They were found together in the morning at a rock which was adopted as the site of the dagaba after the result had followed similar experiments with a horse and an elephant.

The king found some difficulty in providing all the bricks for the work, but Sakka, that is, Indra, was good enough to relieve him of this trouble by sending Vissakamma, the builder of the gods, to make them for him.

: i Although the Moormen of Ceylon have been stated to be

from Indian Dravidians who had adopted Muhammadanism, I have

the best authority for saying that Arabs from Western

a racial affinity with them, and still occasionally settle them.332 ANCIENT CEYLON

When the relic-chamber in the upper part of the structure was ready for the relics, the king carried the forehead relic on his head and deposited it in it, and afterwards the queen similarly placed the hair relic in the room. After every one had put in the relic-room the jewellery and ornaments on his person, the chamber was closed by being covered with a stone slab.

The dagabahad tliree basal platforms, and was of the' Bubble * shape ; its dimensions are not stated. A wihara was also built at the spot and liberally endowed. The book describes its formal gift to the Community of Monks at a great festival at which Abhaya, the king's son-in-law, and other princes were present; and the words doubtless show us the orthodox method of making such grants. A large concourse of monks was there, and before these witnesses the king poured water over the right hand of the superior monk present. Then, in the words of the Dhatuvansa, ' afterwards the king made known (the gift of the temple), saying, ?" (My) lord, and members of royal families assembled together, in (accordance with) the (usual) arrangement for causing the acceptance of the wihara,. I have poured the water on the right hand " ; and the thera, having heard these words, declared his agreement, saying,, "It is good, Maharaja." '

THE NlKAWAE-KANBA DlGABA

High up on the precipitous eastern side of Nikawae-kanda,. a steep rocky hill in the North-western Province, an early monastery was established at a series of natural caves. Some of these contained statues, and one had also a small dagaba which had been demolished by treasure seekers, so that only a little of the lower part remained. Local tradition attributed the founding of the monastery to Prince Sali, the son of King Duttha-Gaminil; in that case it would belong to some date

1 According to the Mahavansa, a minister called Sali, who may have been this prince, built the Sali wihara during the reign of Watta-

Gamini. Its site is unknown ; if it was not at Anuradhapura it may "be this one.THE EARLIEST DAGABAS 333

after the middle of the second century B.C. He appears to have been the chief of this part of the country, and two .sites are pointed out as places at which he dwelt, one on a hill near the caves, called Gal-giriya-kanda, and the other to the north-west, called Raja Angana, where remains of various buildings of some importance are to be seen.

The tradition is supported by the sizes of the bricks in the lowest part of the dagaba. They have an average length of 16-41 inches, and a breadth of 8-16 inches, and a thickness of 3-39 inches ; Bt. is 27-6 and the contents 454 cubic inches. The contents indicate a date not later than the first century B.C., while the other dimensions probably belong to the period extending from the accession of Duttha-Gamini to the early part of the first century, this being evidently a time when larger dimensions than those of either earlier or later bricks were sometimes adopted. These bricks resemble those at .some caves at Nuwara-kanda, in the North-western Province, where an inscription was cut which appears to contain a reference to King Duttha-Gamini.1 These latter bricks are 16-76 inches long, 876 inches wide, and 3-09 thick; Bt. is 27, and the contents 454 cubic inches. On this evidence I conclude that the dagaba most probably dates from the latter half of the second century B,C-, or the first quarter of the next century.

In two of the Nikawae-kanda caves, which evidently have been abandoned for several centuries, there are two ancient wooden statues, larger than life, protected by a thin coat of plaster, one of which tradition identifies as the figure of Duttha-Gamini, while the other is believed to be that of his son, the supposed founder of the monastery. The former statue has the high royal crown, resembling in general appearance those represented in reliefs at Anuradhapura, which at least proves that he is a king. A fragment of an inscription of about the eleventh century, which was found on a broken slab in one cave, may belong to the time when works of restoration were carried out at the caves ; and possibly these wooden

1 See. The Earliest Inscriptions, No. 58.334 ANCIENT CEYLON

statues may be assigned to the same period.1 They tend to show that so long as eight centuries ago the same tradition regarding the founder was current.

About eight years ago, Selaratna Thera, the energetic superior of the monastery now at the foot of the hill, undertook the restoration of the cave wihara. Among other preliminary work, the heap of bricks in the lower part of the dagaba was removed, and it was then discovered that the persons who rifled it, possibly the followers of King Magha, had not found the true relic-chamber, which was covered by a large stone slab in the very bottom of the structure. When this was raised the undisturbed contents of the cavity under it were found to be as follows.

One bead (Fig. No. 106), a little flattened at the ends, belonging to a rosary, -80 in. long and -86 in. wide. It resembled porcelain in appearance, but is believed by Mr. C. H. Read, of the British Museum, from my sketch of it, to be of glass. Mr. Read was good enough to write of it,?" Beads of this design were in use in many places, and at several periods. We have some very similar obtained at Akhmim in Egypt? doubtless Egyptian make?of an uncertain date, but quite old enough to fit in with the date claimed for your Sinhalese dagaba. Others very like?N. European make?are commonly found in Merovingian graves and in the corresponding English cemeteries which may be dated as sixth century A.D. Of the two the former are more like your description." The ground colour of this bead is white, in which are waving parallel close black lines of varying thickness, except that in two irregular patches the colour is plain deep chrome yellow.

One rounded crystal bead (Fig. 105), 1-32 inches long and i-12 inches wide. One spherical deep blue glass bead, not measured but like a small ' marble.' One translucent blue glass moulded bead (Fig. 99), -72 in, square at the top and

3 In the cave temple at Dambulla a wooden statue of King Nissanka-Malla, which probably dates from the commencement of the thirteenth century, is in better preservation than these, and it proves that in a dry site free from white ants such wooden figures may last a thousand years or more.THE EARLIEST DAGABAS 335

bottom, and having similar square faces on the sides at each corner, with the intermediate angles bevelled into triangular facets. All the beads were drilled for stringing.

There were also two pieces of a kind of hard dark brown aromatic composition, cast in moulds, and having a flat bottom and a rounded boss on the top, and blue-lotus petal decoration round the sides (Fig. 97). Numerous tiny specks of gold were to be seen in them, and, strange to state, they distinctly emitted a slight fragrant aroma when I examined them. There were several other persons present at the time, and all noticed and remarked on this long-enduring scent.

Lastly, no less than twelve karanduwas or relic-cases were there, of which ten were clear crystals and two were formed of glass, one (Fig. 104) being dark green in colour, and the other (Fig. 107) opaque pale green. These were both cast in moulds, and were broken when I saw them. All the glass except the first-mentioned bead exhibited some iridescence on the surface. There was also a clear crystal of a lotus bud form (Fig. 98). The illustrations show the shapes of the articles, and the various types of karanduwas.

The largest crystal relic-case (Fig. no) held, it was said, one hundred and twenty minute pearls, all bored, and a tiny fragment of bone wrapped in a piece of thick gold leaf. Each of the others was said to have held a similar relic wrapped in gold leaf, and I have given a figure of one relic (Fig. 100) and a gold leaf package (Fig. 101) containing another, which were produced for my inspection. A much smaller package is visible in Fig. 108. All the cases had stoppers, but I was not shown those of the glass relic-cases. One of the crystal cases (Fig. 103) had a dark blue glass stopper, which was broken when I saw it. All the crystals were admirably cut by the aid of a lathe. There was nothing to indicate whose relics were enclosed in the cases, or by whom they were deposited.

These relic-cases furnish a valuable set of illustrations of some of the early types of dagabas. The tubular cavities in them are shown by dotted lines.336 ANCIENT CEYLON

TYPES OF DAGABAS

The construction of dagabas has been practised for so many centuries in Ceylon that it is not surprising to find that there is a recognised Canon which regulates their general proportions, although deviations from it occur in every work. To what date it belongs is unknown ; as it includes the height of the chatta it appears to have been written at some period not much later than 500 A.D., since the histories contain no reference to these terminals after the fifth century.

The Canon is found in a manuscript of which I failed to secure a copy, called the Waiddyanta-pota, and is written in a language which is chiefly Sanskrit, but partly Pali. It was copied for me in Sinhalese characters as follows : ?

Thupesu taram krita panca bhagarn. Gunan pamanan tribhaga tungam Ganthakara Ghatakaram Bubbulakara Dhanyakam Padmakarambala shat vidham

Thupesu taram krita panca bhagam Gunan pamanan catuvisa bhagam Trimala pancarddhaka garbbham aslitam Catussurakoshtha yugarddha yugmam Sash tan ta kuntam punarddha chatram Vadanticatah munihih puranaih.

1 Having divided the width across the dagaba into five

parts, (out of them) three parts are the height (of the dome). Bell-shape, Chatty-shape, Bubble-shape, Heap-of-Paddy, Lotus-

shape, and Nelli (fruit) are the six kinds (of dagabas).

* Having divided the width across the dagaba into five parts,

the length (of the dagaba) is subdivided into twenty-

four parts. For the three stories (or necklaces, take)

five and a half; the chamber (dome) eight; the four-

of the Celestials (devatas) a couple and

* TYPES OF DAGABAS 337

a half; the (other member of the enclosed*) pair (one and a half l) ; the last six for the spire ; a half more for the chatta. The sage of old prescribed (these proportions) as usually practised/

The Sinhalese names of the various parts of the dagaba are, tun-mal pesawa, or pesd-waMllu, (the three-story ornaments ' or 'the ornamental bangles/ the basal platforms ; gaeba, ' the chamber/ the dome ; hataraes kotuwa, ' the square enclosure/ or tee; devata ko$uwa, 'the godlings' enclosure/ or base of the spire ; kota, the spire ; sat, 'the umbrella ' or chatta ; kota kaemella, 'the end of the spire/ the pinnacle.

According to these rules, we see that the total height of a dagaba should be three times the height of the dome, which is three-fifths of its (widest) diameter. The length of the spire is fixed at a quarter of the total height, or three-quarters of the height of the dome. The height of the basal ledges would be eleven-sixteenths of the height of the dome ; that of the tee five-sixteenths of it; that of the base of the spire three-sixteenths of it; and that of the chatta one-sixteenth of it. As no chatta is now constructed its part is added,to the height of the base of the spire, making it one-quarter of the height of the dome. Modern constructors do not measure these heights vertically on their drawings, but upon a sloping line extending from the centre of the top of the spire to the edge "of the lowest basal platform, the outermost line of the circumference. By this means all the heights will be reduced.

Apparently these were the recognised proportions for dagabas of all the six shapes, and they usually guide modern designers, who, however, I have been informed, commonly add one extra part to the height of their domes, which are now always of the Bell shape, thus making them four-fifths instead of three-fifths of the diameter.

When we apply these proportions to the pre-Christian works there is usually no agreement with them, excepting sometimes in the altitude of the dome.

1 These words in brackets are necessary to make up the correct

number of twenty-four parts.338 ANCIENT CEYLON

The chief difficulty in testing the heights lies in the uncertainty regarding the length of the ancient cubit. Dr. Davy (1816-1820) remarked that' Carpenters and some other artists have measures of their own. The carpenter's angula [inch] is equal to the space between the second and third joint of the fore-finger; and his wadu-riyana [carpenter's cubit] is composed of twenty-four angulas and is divided into four parts/1 In this case the carpenter's cubit would be about two feet long. Twenty years later Major Forbes stated that the carpenter's cubit was two feet three inches in length,2 but the reputed height of the Kaelaniya dagaba, restored in..'1779, does not support this. Captain Robert Knox, writing of the measures used in the middle of the seventeenth century, said, ' A Rian is a Cubit, which is with them from the bone on the inside of the Elbow to the tip of the fourth Finger. A Waddo rian is the Carpenter's Rule. It is as much as will reach from one Elbow to the other, the Thumbs touching one the other at the tops, and so stretching out both Elbows.'3 At the present day, Sinhalese artizans make these measures agree with the English scale by using a riyana of eighteen inches, subdivided into inches, and eighths of inches called nul, which are again divided into fourths ; and a wadu-riyana of three feet, that is, two riyanas.

The old Pali vocabulary, the Abhidhana-padipika, has three words, ratana, kukku, and hattha, as synonyms which mean a cubit, or two spans, vidatthi. The vidatthi was a measure of twelve angulas or fingers ; and it wiH be found on trial that twelve fingers' breadths thus measured by laying the hand flat, the usual method in Ceylon, exactly make up the length of the span from the end of the extended thumb to that of the little finger.4 In the Pali edition of the Mahavansa both hattha and ratana are employed in stating the heights

1 An Account of ike Interior of Ceylon, p. 244. I have corrected his spelling of wadu-nyana,

2 Eleven Years in Ceylon, p. 223.

3 An Historical Relation, of the Island Ceylon, p. 98,

4 According to Winslow's Tamil Dictionary twelve fingers' breadths make one span.TYPES OF DAGABAS 339

of the dagabas, these words being alike translated in the Sinhalese edition by the term riyana. Thus it would seem that the cubit used in the histories is one of two spans or nearly 18 inches ; but this is far from agreement with the actual heights of the dagabas.

I found that ten men had a mean height of 5 feet 45- inches, and that their average cubit, measured as described by Knox, the method always used at the present day, was 17-88 inches.

We. may obtain a measurement of the ancient cubit by means of the early bricks ; the largest ones are always termed riyan-gadol, f cubit-bricks/ an expression which indicates that their length was determined by the measure of the ancient cubit.

For ascertaining the length in this manner I have taken all bricks in my tables the volume of which exceeds 400 cubic inches, this being apparently a fairly trustworthy indication that they are pre-Christian, and I find that the length of those used at nineteen different works averages 17-56 inches, or only one-third of an inch less than the length measured on the arm. But bricks shrink considerably in drying, and we do not know whether it was the length of the brick when thus contracted, or the length of the mould in which It was formed, that represented the early cubit. If it was the latter we should require, with the ordinary clay of Ceylon, an addition of at least three-quarters of an inch, or even an inch, to the size of the burnt brick in order to arrive at the true length of the cubit. It is probable, however, that allowance for this shrinkage was made in the size of the mould; at the same ratio as in the men I measured, a cubit of 17-56 inches would be that of men who were 5 feet sJ inches high, which is very nearly the actual height of the present Kandian villagers.

Notwithstanding this contemporary evidence of the length of the early riyana used by the brickmakers, a comparison of the heights of the dagabas given by Mr. Smither with those stated in the histories, shows that another cubit must have been in use from the earliest period. Thus, in the case of the340 ANCIENT CEYLON

Ruwanwaeli dagaba the total height, if the cubit were inches long, would make the top of the lowest chatta, or of the pinnacle if there was one originally, only eight feet above the tee shown in my restoration; and at the Abhaya-giri the pinnacle would be inside the existing remains of the spire.

Nor is this difference between the existing and the former recorded measurements merely due to the inaccuracy of the latter. When we seek evidence of the length of the other cubit we are confronted by one striking fact. No one can take many measurements of the ancient works in Ceylon without being astonished at the frequent occasions on which these are found to be an exact number of English feet. For buildings, such sizes as 36 feet by 24 feet, or 18 feet by 12 feet, or measurements of 20 * or 10 feet, are quite common. The sides of square pillars usually measure 12, or 15, or 18 inches. The culverts of ancient sluices are in most cases exactly 12, or 18, or 24, or 27 inches wide, and some are exactly 2 feet, 2 feet 6 inches, 3 feet, 3 feet 6 inches, and 4 feet high. Mr. Sinither's dimensions of the dagabas are also incontestable evidence which points in the same direction.

These numerous examples indicate that the cubit of the early mason^ and carpenters was not that which was used by the brickmakers, but was either exactly eighteen inches long, or much more probably exactly two feet long. The recorded heights of the early dagabas when compared with the existing remains prove that it was considerably greater than eighteen inches long, and it must therefore have been of the other dimension, that is, two feet in length. The ancient cubit was always equal to twenty-four angulas; it was the mode of measuring the angula or * finger * that varied. One trade, the brickmakers, and probably also the general public, employed the width of it, and apparently their cubit was 17! inches long; other trades, the builders, stonecutters, and carpenters, had a longitudinal scale, as described. by Dr. Davy, and their cubit thus became two feet in length.

1 I have already stated that the outer shell of the Ruwanwaeli dagaba Is 19 feet iff inches thick.FIG. 112. Miniature Stone Dagaba.TYPES OF DAGABAS 341

On the whole, it would appear that while definite rules depending on the widest diameter of the dome sometimes regulated the height of the dome, and less commonly the total height also, the early builders allowed themselves considerable latitude in determining the proportions of the various parts which made up the whole dagaba. The height of the basal platforms was evidently fixed independently of the rest of the structure. In the large Mihintale dagaba, the Abhaya-giri, the Lankarama, the Dakunu, and the Jetavana dagabas, as well as in a later miniature stone dagaba next described, the breadth of the tee is not far from one quarter of the diameter of the dome. The height of the tee, and the thickness of the lower part of the spire were about one-tenth of the diameter of the dome in very large dagabas.

MINIATURE DAGABA ON THE RUWANWAELI PLATFORM

An interesting small limestone dagaba, cut out of a single stone, was placed on the pavement at the side of the Ruwan-waeli dagaba; and not being a receptacle for valuables it escaped serious damage by treasure seekers, with the exception of the spire, the upper part of which is broken off.

It appears to be the work referred to by Nissanka-Malla (1198-1207 A.D.) in his inscription on a large slab called the Galpota, 'the Stone Book/ at Polannaruwa,1 in which he records that * he made a stone dagaba [at Ruwanwaeli] as a worship-place for the Gods.' Excepting in the heights of the dome and the basal platforms, this work partly adheres to the Canon, thus proving that this scale was already in existence.

The diameter of the dome, according to Mr. Smither's measurements, is 38 inches ; by the Canon its height should therefore be 22-8 inches, but it is actually only 19 inches. The total height, which is now 4 feet 2% inches, must have been very nearly according to the Canon. The basal ledges being lower than the Canon requires, the difference has been

1 Dr. E. Mutter. Ancient Inscriptions, No«- 148.342 ANCIENT CEYLON

added equally to the tee and the base of the spire. Round the latter a series of reliefs of standing figures, separated by pilasters, explains why this member is termed the ' enclosure of the devatas/ the figures evidently being those of the devatas who guard the structure. The series of shallow niches in the base of the spire at Abhaya-giri may have had plaster representations of devatas like those on this work, and it may be assumed that the other great dagabas were all similarly supplied with guardian deities.

The three basal platforms are loj inches high, the tee 8 inches, and the base of the spire 6 inches. Elephants' heads project round the base of the dagaba, and there is a lion looking outward at each corner.

EXAMPLES OF TYPES

Of the six kinds of dagabas, I am acquainted with no pre-Christian example of the modern Bell shape with the exception of the Idikatu dagaba at Mihintale. The primitive type of. bell which was usually copied may have been unlike later bells; and it is most probable that the Lankarama dagaba, in which a more or less hemispherical dome rested on a short vertical-sided cylinder of the same diameter, may represent the earliest Bell-shaped edifice. If not, we should have in it a seventh type, which the old authority would be unlikely to omit from his list.

In excavating at an early monastic building at Anuradhapura, Mr. Bell found a small copper bell with high vertical sides and a rounded top,1 which proves that this shape was employed in ancient times for such articles. Ancient Egyptian and some early Indian bells were somewhat similar, with rounded tops. There can be no doubt also that the common wooden bell which is hung on the necks of cattle, and has nearly vertical sides, adheres to a primitive type. The Bell dagaba of modem times copies the present form of the bell in varying proportions, and is now decidedly the favourite shape with the designers of these structures.

1 Archaeological Survey of Ceylon. Fourth Progress Report, p. 4*TYPES OF DAGABAS 343

The ' Chatty V or Water-pot shape has been abandoned, and was never popular. It is formed from the sphere, by fixing the base line at about one-quarter of the height, so that the dome becomes three-fourths of a globe. The only Sinhalese representation of it with which I am acquainted is the Tissa karanduwa No. 93, in which the greatest diameter is at about one-third of its height. This style of building appears to have been more practised in Indial than in Ceylon, but not in very early works.

The Bubble, or plain hemisphere, was the favourite design in early times, and it is still found in the case of many small dagabas reconstructed inside cave temples. Nearly all the very early dagabas in Ceylon and India followed this type, which with its simple well-rounded outline is perhaps the most effective one for very large structures. A large number of the smaller dagabas had no tee or base for the spire, which in such a case was in the form of a monolithic stone pillar, usually octagonal in section, with a rounded top ending in a blunt point; it rose directly out of the top of the dome.

The Heap-of-Paddy shape is doubtless a very ancient type of dagaba, which perhaps represents a form of early eastern tumulus or cairn. It consists of the end portion of a wide cone, with slightly curved sides and a rounded top. The restored Kaelani dagaba is of this type and the Sela caitya at Amiradhapura may also have been of this shape. It is not found in Tissa, nor in the karanduwas; but a few examples are to be seen at secluded wiharas, and although these have been restored after their original construction it is most unlikely that the form would be changed from another type to this unusual one. The tee in this dagaba is commonly a very small one, but at the Kaelani structure the length of its side is very nearly equal to a quarter of the width of the dome.

There is a dagaba of this character at Ottappuwa, in the North-central Province, which is attributed by tradition to Devanam-piya Tissa. The bricks in it are 18-60 inches long, 9-52 inches wide, and 3-12 inches thick ; Bt. is 29-7 and the contents 552 cubic inches. These dimensions point to some

1 An example is to be seen in the Amaravati relief, Fig. No. 81.344 ANCIENT CEYLON

time late in the second century, or early in the first century B.C., as the date of the work. An inscription l left at it by ' Siri-kana raja1 (30-33 A.D.) proves that it was in existence before his time.

FIG. 113. Ottappuwa Dagaba. ?siv

At a dagaba of this shape at Wellangolla, in the Northwestern Province, the bricks are 3 inches thick, a size that belongs to pre-Christian times. At another, at Kahatagaswela, in the same Province, the bricks average 15-22 inches in length, 7-84 inches in breadth, and 2-65 inches in thickness, Bt. being 20-8, and the contents 316 cubic inches. This size indicates that they were burnt in the first three centuries after Christ.

The Lotus (bud) dagaba might be expected to be of common occurrence, from the popularity of the flower as a decorative * motive'; but it is one of the rarest forms of dagaba, and I have not met with a single building of this shape. It is found, however, in the Nikawae-kanda karanduwa numbered 104, in which it represents an unopened lotus bud. The crystal numbered 98 may be an unfinished karanduwa of this type, there being no relic-cavity in it

The Nelli dagaba is supposed to represent the form of the fruit of the Nelli tree (Phyllanthus emblica), which seems to have been highly esteemed in former times, since it is used as a popular simile in writings and inscriptions. King Nissanka Malla mentions that he had inspected all Ceylon and ' had as precise a view of the whole as if it were a ripe Nelli fruit in his hand.2 This is a small round fruit with a green rind,

1 See the final chapter, and Fig, No. 153 for facsimile.

2 Ancient Inscriptions, No. 143..TYPES OF DAGABAS 345

growing on a low tree which is abundant in part of the Nilgala district; it is there about the size of the ordinary * marbles * of schoolboys.

I presume that the round-topped dagabas in which the curve of the side of the dome is an arc of a circle, the centre of which is beyond the vertical axis, represent this form. This outline is seen in the Tissa karanduwas numbered 94 and 96 and Nos. 108-110 at Nikawae-kanda; and possibly such designs as the Tissa karanduwa No. 95 and the Nikawae-kauda karanduwas Nos. 102,103 and 107 were grouped under this heading. These forms appear to have been rarely adopted in actual construction in Ceylon.

For convenience of reference, I append an amended list of the dimensions of dated bricks, in which I have inserted the sizes of the bricks at the structures to which definite ages have been attributed in this chapter.AMENDED LIST OF DATED BRICKS.

No Name of Structure. Date. Length " in Inches. Breadth in Inches. 1 Thickness j in Inches. Breadth x Thickness. i (Contents inj Cub. Ins. 1

B.C.

Large dagaba, Sanchi, India 3rd Cent. 1 6 -00 ro-oo 3-OO 30 48 480

I Maha Saeya, Mihin-

tale .... N.C.P. ? I7-92 8-87 2-91 25-8 52-1 461

2 Mahanaga dagaba, Tissa .... S.P. 17-35 8-84 2-83 25 49-1 434

3 Yatthala dagaba,

Tissa .... » 17-85 8-64 2-90 25 51-7 447

4 5 Sanda-giri dagaba, Tissa .... Maenik dagaba, Tissa ? Early 2nd Cent. 2nd Cent. 17-14 16-57 8-67 8-86 2-8 1 2-80 24*4 24-8 48-1 46-4 418 411

6 Miriswaeti dagaba,

Anuradhapura . N.C.P. « ? 10-41 3*00 31-2 ? ?

7 Ruwanwaeli dagaba,

Anuradhapura

(inner work) . . ? j ? 9.67 2-79 *7i ? ? -

8 Ruwanwaeli dagaba,

Anuradhapura, (outer work) . . ? 14-06 8-99 2*90 26*1 40-8 366

9 Ruwanwaeli dagaba,

Anuradhapura,

10 (elephant wall) . ? Nuwara-kanda . . N.W.P. " I7-OO 16-76 8^76 2*85 3-09 27 48-4 52 454

ii Nikawae-kanda dagaba ? Late 2nd Cent. 16-41 8-16 3-39 27*6 55-6 454

or

early ist Cent.

12 Abhaya-giri dagaba,

Anuradhapura . K.C.P. Early ist Cent. 18-92 9-62 3-20 30*7 60-5 5**3

13 Lankarama dagaba,

Anuradhapura . ? f 17-37 8-94 2-62 23-4 45-5 407

14 Danibulla cave temple 1 C.P. ist Cent. I9'45 ? 2-45 ? 47*6 ~?

A.D.

15 Mfflaewa-gala dagaba N.C.P. ist or 2nd Cent. I5'57 8-00 2-72 21.7 42-3 339

i6J Daknnu dagaba,

Anuradhapura . ? 2nd Cent. 16-36 8*18 2-31 18-9 37*8 309

17 Jetavana dagaba,

Anuradhapura . ? Late 3rd Cent, 15*82 8*41 2*26 19 35-7 301

i8j Hurulla tank, central

sluice .... ? »» ? 8*40 2*70 22*7 ? ?

1 Hurulla tank, high-

level sluice . ? tt ? 3-35 2-51 21 ? ?

I9j Padawiya sluice ,, t* 14-02 8-50 2*46 2O9 34-5 293

20 vann^.t"|i PSlania ,, »» I3*6l 8-48 2-70 22-9 36-7 3"

21 Nlrammulla sluice N.W.P. )r : ? 8-45 2-70 21-5 ?

22 MSligawa

(inner room) . N.C.P. Early 4th Cent. 14*10 8*45 2*52 21*3 35*5 300

2J Tdpiwaewa sluice ? Late 4th Cent. ? 7*40 2*43 18 ? ?

24 Slgiriya gallery ? Late sth Cent. 13*09 7-32 2-31 16-9 30-2 220

25 Rankest d%abaf

Polannamwa . I2th Cent. 12*52 8*40 2*00 16-8 25 210

a6J TMpIjaina hall (In-

side) . . , . 12'OQ 8-36 2'OO 167 24 2OI

27 Fortification Wall

Polannaruwa . f n 12*63 8*20 2'OO 16.4 25-3 207

There is reason to beHeve that about the tenth century the breadth

was 8 inches and the thickness a little less than 2 inches, but I

amenable to give the of any dated bricks of that period.

1 For the afe erf this see The Earlmt ItucriptfoHSt No. 75.

345X THE EARLIEST IRRIGATION WORKS

THE special feature of the ancient civilisation of Ceylon was its irrigation works, which with the exception of part of the mountain district were made throughout the whole country. Their purpose was to store or convey the water which was required for the rice fields that were formed at every suitable place in the island.

Two different systems of irrigation were adopted, depending on the circumstances of each case. According to one the water was impounded in reservoirs, from which it was gradually passed out, either directly onto the fields where it was wanted, or by means of excavated channels down which it flowed to them.

According to the other system, part of the water flowing down the rivers was turned into longer excavated channels which conveyed it to more distant lands, or reservoirs, temporary dams or permanent masonry dams being constructed across the rivers below the off-takes of the channels, in order to divert into them a larger quantity of water than could be secured without such aid when the flow of the rivers began to diminish after the end of the seasonal rains of the two monsoons. The north-east monsoon lasts from October to March, and its regular rains end in January; the Southwest monsoon lasts from April to September, and its rains cease in June.

This latter method of irrigation by means of channels cut from rivers is of the greatest antiquity, having been practised in North-western and Central India, and most probably also Southern India, from immemorial times. It originated in the Euphrates valley, where the cultivation

347348 ANCIENT CEYLON

of the fertile lands on both banks of the river was largely dependent on it.

The first record of any irrigation work there is contained in an inscription left by Eannadu, King of Shirpurla in Southern Babylonia, who ruled in about 4000 B.C., and who mentions his construction of several canals, one of them being known as' Lummadimshar/ at the side of which he made a reservoir, the first on record, ' a basin (containing) 3600 gur [each being eight bushels] of water/ Another of these canals is specially stated to have been cut ' from the great river ' (Euphrates).

Entemena, nephew of Eannadu, recorded the opening of several fresh canals, and also the prolongation of the Euphrates canal to the river Tigris. Urukagina, King of Shirpurla, who reigned in 3900 B.C., according to the latest conclusions, * and not in 4500 B.C. as was supposed by Dr. Radau (Early Babylonian History, p. 47), also cut a canal there. His own words regarding it are, ' For Nina her beloved canal Ninaki-tum-a he has built/ Nina was the Goddess Bau, the Great Mother.

In nearly all cases these early canals were distinguished by special names. It is most improbable that this would be the case when irrigation channels were originally made, and as one of the first ones of which a record has been preserved has its own title it may be concluded that the construction of such works dates from nearly 4500 B.C., or possibly an even earlier time.

In India, we find the digging of channels referred to in very early times (Rig Veda, iii, 33, 6; iv, 19, 2), perhaps in the third millennium B.C. ; and the benefits derived from them would be so apparent that doubtless many others continued to be opened from that period down to historic times, even although no actual record of them has been preserved.

While it is almost certain, therefore, that the first Gangetic must have been acquainted with this manner of irrigation they came to Ceylon, there is nothing to indicate that brought with them a knowledge of the construction of reservoirs, which as a general rale were neither required 1 King aad Hall. Egypt and Western Asia, p. 189.THE EARLIEST IRRIGATION WORKS 349

nor made in the districts inhabited by their ancestors in India. Although an inscription at Junagadh1 has recorded that one was formed by Pushyagupta, the brother-in-law of the great King Candragupta,2 and was afterwards repaired by the latter's grandson, the Emperor Asdka, it appears to have been a comparatively small work, of which little or no trace now remains. It is possible that the Sinhalese acquired a knowledge of the art of reservoir construction in Southern India. In any case, there can be no doubt that the credit of its development and extension in the island is due to some of the first Sinhalese rulers and their responsible advisers.

The nature of the flat plains around the sites of the primitive capitals of Southern India could never have encouraged the construction of reservoirs with high embankments, which, in fact, are still non-existent on them. All that could be attempted there in very early times in the way of making reservoirs would be the formation of shallow village tanks, with embankments from six to ten or twelve feet high, for retaining a supply of rain-water for bathing purposes, and for the irrigation of the adjoining fields attached to each village.

It was only in the districts surrounding the early capitals of Ceylon that the necessary conditions existed for promoting the construction of larger works of this character?-a series of shallow valleys down which flowed seasonal streams of moderate size, and a heavy rainfall lasting for only a short period in each monsoon. It may be assumed, therefore, that the formation of all reservoirs of a class with embankments much higher than those of simple village tanks was originally due to the constructive genius of the Sinhalese themselves.

At an early date they undertook the raising of great earthen embankments, often some miles in length, across many suitable valleys, thus intercepting the flow of the streams, and storing up during the rainy seasons, in the reservoirs thus formed, immense sheets of water for the irrigation of large

1 The Indian Antiquary, Vol. vii, p. 257.

a According to Mr. V. A. Smith, this king reigned from 321 to 297 B.C. (Early History of India, p. 44). Sir F. Max Mullet's date is 315-291 B.C. (The Dhammapada, p. xxxvi).350 ANCIENT CEYLON

tracts of land lower down in the valleys, that were found to be suitable for the cultivation of rice, the only culture for which the water was utilised.

In addition to the benefit which the country in general derived from the works, a considerable part of the produce of the irrigated lands was devoted in many instances, after the introduction of Buddhism, to the maintenance of the Buddhist monks. Thus it soon came to be thought an act of great religious merit to construct such reservoirs, and the continuance of the practice by all the pious monarchs of the island was then assured. These were the larger works, such as private enterprise could not attempt to undertake.

In the meantime, the formation of minor works at the villages, by the combined labour of the inhabitants, was doubtless encouraged,1 until in the end such 'village tanks' were

1 Moral Suasion, when applied by the district chiefs, must have been an exceedingly effective instrument. Those who possess an intimate acquaintance with the village life of the East will easily understand that in early times the life of the villager who ventured to set his inclinations in opposition to the will of the chiefs would become an extremely uncomfortable one?as is often the case even now. At the present day, in Ceylon it is not necessary that the chief or headman should take any active steps against the recalcitrant person in order to bring this about. In every village there are at least two parties,, often bitterly opposed to each other ; usually they consist of the friends and followers of the local headman for the time being, and the adherents to the ex-headman. When a villager is once known to have incurred, the displeasure of the local chief, and more especially if he be a man. who has played an unpopular role in the village for any reason, his enemies proceed to take advantage of the opportunities which this affords them to annoy him. His cattle are stolen, sometimes his corn-stack is burnt down in the night, or his house robbed during his absence. When he appeals to the headman for assistance in apprehending the culprits he is informed that they will be arrested on his discovering them and producing some proof of their guilt; and there the matter often ends, as the headman is not interested in it, and takes no steps to find out the wrong-doers. In many instances, false charges are trumped up against the objectionable person, or false or doubtful claims instituted over his lands, which he often has the greatest difficulty in rebutting, there commonly being some weak points in his own proofs of his ownership. If resort be not made to these extreme measures there are many other ways of inflicting petty annoyances on him the cumulative effect of which almost renders his-life a burden to him.THE EARLIEST IRRIGATION WORKS 351

established at practically every little settlement in the drier districts of the island.

The first irrigation works made in Ceylon obviously would be these village tanks, containing sheets of water that covered from two or three acres to one hundred acres or more, the size depending on the amount of the water-supply, the requirements of the village, and the formation of the ground. At first, only the simplest works of the smaller class, with very low embankments, would be undertaken; but when a better knowledge of the art of raising such banks of earth to hold back greater depths of water was acquired, schemes of a more comprehensive character would be attempted, until at last no reservoir was looked upon as too great to be constructed, and the lengths of the embankments extended for any distance up to a maximum of nine miles, while their heights in a few instances rose to more than fifty feet.

The histories, which were compiled by monks who, especially in early times, were chiefly interested in recording the erection of Buddhistic edifices, and the other religious acts of the various monarchs of Ceylon, contain no reference to the formation of the communal village tanks, and too few notices of the construction of even the larger class of works, some of the most important of which are never mentioned in them, at any rate under names that can be recognised at the present day. Im such cases we have nothing to mark the age of the works that cannot be identified in the histories, except the evidence obtainable from the dimensions of the bricks that were commonly used either in some part of them, or in Buddhist monasteries which depended for their existence on the water-supply afforded by the works, and the presence of a considerable population which that ensured. Notwithstanding the possibility of error in fixing the age of a work by such data, the general trustworthiness of these contemporary records is so unmistakable that in the absence of other evidence I shall make full use of them in determining the probable dates of some of the works.

The first notice of the construction of a reservoir in Ceylon is found in the Mahavansa (i, p. 37), where it is stated that352 ANCIENT CEYLON

Prince Anuradha, the brother-in-law of the second king, Panduwasa Deva, made one on the, southern side of the capital, Anuradhapura. This was early in the fourth century B.C. It has not been identified, and we may assume that it was merely a small work intended for the use of the village at which the prince resided.

The record, the truth of which there can be no reason for ?doubting, is interesting as showing that the Sinhalese had already become acquainted with the art of making reservoirs. Considering the intimate connection existing between the first Sinhalese king and the king of Madura?Wijaya having married the latter's daughter?such a knowledge could easily have been acquired from Southern India before this date?perhaps even by the early Nagas. Wijaya must have obtained his information regarding Madura and its sovereign's family through traders who were visiting the two kingdoms; there could be no other travellers to carry news in those days. Thus there would appear to have been a regular intercourse between the two countries from an early period; it is improbable that it would spring into being simply because Wijaya had become king of Ceylon, since mere settlers from the valley of the Ganges would have no personal acquaintance with Madura, and its ruler, and its trading requirements. They can only have heard of them from traders who had been at Madura. Such persons would doubtless observe the advantages accruing from the presence of village tanks on the line of their journey?dried up as the country becomes when there is no rain?and the knowledge of them would thus be transmitted to Ceylon.

It may appear to be such a simple matter to raise a long bank of earth in order to hold back a certain quantity of rain water for bathing purposes or for watering an adjoining rice field after the rains have ceased, that any people living in hot countries where the rains are only seasonal and are followed by several almost rainless months might be expected to be struck by the idea of making these little reservoirs for themselves, without its transmission from anoijier country ; but as a matter of fact the notion of reservoir-making appearsTHE EARLIEST IRRIGATION WORKS 353

to have been originated in only one country, and never to have been invented independently elsewhere, at any rate in the Old World. When I visited West Africa, the natives of the Gambia valley who have cultivated rice for so long a period that they have developed many special varieties of this grain, informed me that such an idea as storing water for its irrigation had never crossed their minds. They had never heard of such a practice, and had no notion regarding the manner in which such works should be constructed, even on the smallest scale. Probably this was the position in other countries.

It is most likely, therefore, that the art of reservoir construction owes its origin to the early peoples of the Euphrates valley, and that it spread westwards and south-eastwards from that centre, reaching the Dravidian districts of India possibly before the Aryan invasion of the country, and being transmitted thence to Ceylon.

The next work for storing water, of which any information is given in the histories, is of an entirely different class from the village tank of Anuradha. Possibly it was the first reservoir ever made with an embankment of an importance that must have required special acquaintance with the principles of reservoir construction. The honour of occupying this prominent position rests with either Panda-waewa l in the North-western Province, or Abhaya-waewa at Anura-dhapura, or possibly another reservoir at that city.

PANDA-WAEWA

In the North-western Province, near Hettipola, a small village at the junction of two roads, and sixteen miles east of Chilaw, the large deserted tank called Pancja-waewa is found. On its southern side and close to the end of its embankment, there is a fortified site which apparently was once that of a town of considerable size, but is now completely overgrown with forest and jungle. .It is known as Panduwas Nuwara, ?and is locally believed to have been a city founded by Pancjuwasa

1 The Sanskrit and Pali word vdpi, the Elu or early Sinhalese words wawi and wiya the Sinhalese waewa,.and the Tamil kulam have the same meaning, and signify * tank *. or * reservoir/

A A354

ANCIENT CEYLON

Deva in the first half of the fourth century B.C., and at one time his capital.

Beyond the name and the tradition, there is no evidence that he actually founded a city at this site ; but this at least may be said in favour of the tradition?that it Is in the highest degree unlikely that if the town was established by a later monarch he would perpetuate the memory of a much earlier ruler, in preference to his own, in bestowing a name on' it.

A long and exaggerated account of the city is given in an old manuscript termed Pradhana Nuwarawal, 'Principal Cities,' which describes other large towns of this part of the

FIGS. 114, 115. Pa^tda-waewa. '.

country, such as Kuranaegala, Yapahuwa, and Kandy. It proves that the site possessed the same name early in the fourteenth century?when the work appears to have been

*~and that it was then believed that the city was founded by the same ' Pancjuwas Raja/ who, it states, also 'for the support of the city made a great tank.' It

that the original city had the honour of being con-

» It the day and hour, but not the year, of the death of a king

Bahu, and ends after mentioning the accession of hm brother, a monk, under the name of Buja Parakrama The Mag at Dambadeniya, and removed thence

to he foundedTHE EARLIEST IRRIGATION WORKS 355

stracted by Vissakamma, the divine builder, acting tinder orders issued by Indra.

The fortifications consist of an enclosing wall forty feet thick, faced with brickwork on both sides, and having immediately outside it a ditch which is still some seven feet deep, and more than ten feet wide in the bottom. This wall is well defined and still several feet high; in plan it is a regular rhombus, 1000 feet long on two opposed sides, and 950 feet long on the other two.

Without doubt this fortification is of much later date than the time of the second king of Ceylon, but that is not proof that a settlement did not exist there long prior to its construction. In reality, it shows that some pre-existing station of sufficient importance to be worth strong fortifications was already established there when the wall was built. I possess no measurements of the bricks used in the work, No remains of buildings are known inside this fortified space ; this indicates the lapse of several centuries since the place was abandoned.

The connection of the position of the city with the date of the construction of the reservoir lies in the fact that the town was built not only close behind the end of the embankment, but so near the edge of the reservoir that when the latter was quite full the water extended into the ditch which surrounds its wall. This shows that the reservoir was already formed before the exact site of the town was decided upon, so that if the name of the city and the tradition respecting it be regarded as sufficient evidence that it was founded by Panduwasa Deva at this spot, the construction of the reservoir must also be attributed to this monarch, although neither the one nor the other is mentioned in the histories. The chief difficulty in accepting the identification lies in the area of the reservoir, the water of which would cover no less than 1050 acres when it was ML It was not until at least a century later than the time of Pancjuwasa Deva that any other reservoir of this size appears to have been made in Ceylon ; and on a review of the probabilities of the case I should be inclined to think that a town built there by that356 ? ANCIENT CEYLON

king would be nearer the stream across which the embankment of the tank was raised. If the tank was made at a subsequent date, the town, such as it was?probably a mere congeries of wattle-and-daub huts?would be re-constructed in a new position at the edge of the water. If this opinion be not adopted we are driven to the conclusion that this great reservoir was formed in the time when the second king ruled over the country. As evidence in favour of its early date we have the fact that Anuradhapura was established at the distance of nearly a mile from the adjoining river, the Malwatta-oya.

Another difficulty which also throws doubt on such an identification is found in the fact that if the reservoir was made by Panduwasa Deva, we must be prepared to admit that either brick-making or stone-cutting, or both, as well as the art of building with those materials, were sufficiently understood in Ceylon at that early period for the designer to venture to construct a masonry outlet or,sluice for the purpose of regulating the flow of the water and passing it out for the use of the rice fields that would be cultivated lower down the valley by its aid. Such a great body of water would never be retained for the mere use of the inhabitants of the city; and the tank must have been originally intended for irrigating rice lands in addition to providing the people with a supply of water for drinking and bathing purposes. For fulfilling such an object some kind of substantially built outlet at a low level would be a necessity.

On the other hand, there can be no doubt that the arts of stone-cutting and brick-burning were well advanced long before the erection of the first dagabas, and the cutting of the earliest inscriptions and the fronts of the cave shelters of the Buddhist monks, in the third century B.C. No mere learners could have done the works in brick-burning and building, and in stone-cutting, which are still preserved. King Panduka.bh.aya, who probably became king at about the end of the fourth century B.C. (that is, less than seventy years after Pancjuwasa Deva), certainly formed a reservoir which had an embankment higher than that of Pan4a~waewa, ThusTHE EARLIEST IRRIGATION WORKS 357

there must always remain a possibility that the tradition regarding the origin of this latter work is correct. I shall therefore give a short account of the works at Panda-waewa, which may be the first great reservoir ever constructed, if we omit from consideration the great lakes of Egypt, since they were merely immense natural hollows into which water was turned.

The histories contain almost no information regarding this reservoir. They state that King Dappula II (807-812, A.D.) built a hospital there,' with a fruitful village attached thereto/ for its support (Mah., ii, p. 57) ; and it also appears to be mentioned in the same work under the name Setthivapi, ' the Hetti(pola) Tank/ as one of the reservoirs repaired by Para-krama-Bahu I (1164-1197 A.D.). King Nissanka-Malla (1198-1207 A.D.) left a record of this restoration in an inscription of four lines 1 cut on a stone at the outlet of the low-level sluice. Perhaps the work was only completed in his reign. The final breach in the embankment was made in the early years of last century, and its history is instructive as showing how many other great reservoirs in Ceylon may have burst. According to the information which I received thirty years ago from persons who had heard the story of the catastrophe related by those who remembered it, a track made by cattle that crossed the embankment had become worn down into a deep hollow which was left unfilled. The natural consequence of such neglect followed. A sudden and extremely high rise of the water (which flooded some of the nearest houses at the side of the reservoir), following a very heavy rainfall, caused its level to mount up during the darkness of a rainy night until it overtopped the low place ; and when daylight broke the embankment was found to be completely breached at the spot, and the reservoir was empty. The last Sinhalese king subsequently entrusted its repair

1 It begins, Srimat Parakram® Bhttssa, Mahd Bhitssd, Kdlinga Sinha dapa nerana patettvanu, and the last part is dis* mewara gad% wagyam. * The Prosperous King Parakrama, the Great King, the Kalinga Lion, putting aside and subduing pride . . . the overcoming of the impediment to the work of the country.*358 ANCIENT CEYLON

to one of Ms chiefs, but he was recalled in 1815, before the work was begun, owing to the rebellion which ended in our occupation of the Kandian kingdom at the request of the chiefs; and the reservoir has been left in the same useless state down to the present day, although plans for its restoration were prepared thirty years ago.

The embankment was carried in a north-and-south line nearly straight across the valley of the Kolamunu-oya?a stream that rises about fifteen miles away?until it passed across this river; it was then turned round to the southwest so as to abut against a large and nearly flat rock, near the southern side of the valley. From the southern side of the rock it then resumed its southward direction for a short distance, after which it was turned up-stream at a right angle for 2100 feet until it ended at high ground not far from the site of the old city, Panduwas Nuwara. Owing to the configuration of the ground, a considerable amount of earthwork was saved by this sudden alteration in the line of the bank.

The rock, which is about 250 feet across, was utilised as a waste-weir or flood-escape; and it is evident that the valley had been carefully examined, and the site of the embankment chosen with the special view of making use of this rock as a safe place for the escape of floods. It is quite certain that the bank was raised to its full height at the first construction of the reservoir. I made a search in vain for any channel such as the floods must have excavated had they been discharged out of the tank at a lower level; there can be no doubt that from the first they were passed over the rock. That the reservoir remained in working order until perhaps the twelfth century is a proof that the height to which floods would rise over the rock had been correctly estimated,

The embankment is 8400 feet long, or if miles; and.is 22 feet high above the sill of the low-level sluice, above which the crest of the rock rose 13 feet. In later years an additional

of 2 feet of water was retained by means of a temporary dam raised along the front of the rock after the main

had ceased, by the aid of short rough stone pillars,THE EARLIEST IRRIGATION WORKS 359

over which a foot-bridge may have been fixed. The top of the bank was 8 feet wide, and the sides sloped at the rate of 21 feet horizontal to i foot vertical. This section is weaker than that of any other large pre-Christian bank that I have seen. It is almost the only respect in which the bank differs from those of other very early works of a similar size, and it may indicate its greater age ; later experience evidently showed the old engineers the advisability of adopting a broader section and flatter slopes. Along the slope facing the water a layer of small boulders is laid as a protection against erosion caused by waves. This may be of later date than the original work ; such a protection is found at all the larger embankments in Ceylon, with one or two exceptions.

One sluice, with a rectangular stone culvert for discharging water, was built in the low ground on the northern side of the stream, and another at a high level near the northern end of the embankment. They appeared to differ in no respect from similar structures in other reservoirs in Ceylon; they may have been reconstructed long after the original work was done, as the position of the inscription of Nissanka-Malla indicates. I shall refer to the question of the type of the Sinhalese sluice in describing one at a somewhat later reservoir where I was able to examine the original work.

When the reservoir retained a depth of 13 feet of water at the low-level sluice the area covered by it was 1050 acres, and its capacity was 311 million cubic feet; the extra depth of two feet increased the area to 1360 acres, and the capacity to 416 million cubic feet.

Although the size of this reservoir was surpassed by other pre-Christian ones, and left far behind by many post-Christian works, we cannot fail to be astonished at the boldness and originality of the early engineer who ventured to construct such an earthen bank across a valley down which floods of considerable volume passed in the rainy seasons. Owing to the heavy rainfall of the gathering ground, which averages about 85 inches per annum, the maximum flood may amount to 12,000 or 14,000 cubic feet per second. Every engineer will recognise that to get rid of this volume of water360

ANCIENT CEYLON

in safety would be a serious problem; the old designer of the works must have been a highly intelligent man to overcome it so successfully. Besides this he made every effort to reduce the quantity of the earthwork to a minimum; to effect this the line of the bank was twisted about in order to avoid low ground, in a manner never found in later works of large size.

ABHAYA-WAEWA

In about 300 B.C., King Pandukabhaya, the grandson of Panduwasa Deva, made the Abhaya tank at Anuradhapura (Mah., i, p. 43) ; this is the earliest constructive work which

FIG. 116. Basawak-kulam.

can be identified with certainty in Ceylon. Subsequent references to it in the histories, as well as an inscription left at it in the tenth century, containing the orders of King Kassapa IV prohibiting fishing in it, in which it is mentioned by name, prove that it is the existing reservoir at Anuradhapura which is now termed Basawak-kulam. The first duty assigned to me on my arrival in the island in 1873 was the survey of this interesting reservoir, at that time almost useless, having a large breach through the embankment, in front of which a low temporary dam of sticks and earth held'back a little water ; its restoration was undertaken immediately afterwards from the designs then prepared.THE EARLIEST IRRIGATION WORKS 361

It is sometimes mentioned casually in the early part of the histories?in the time of Pandukabhaya and subsequently? but always as a reservoir in working order ; and it appears to have remained unbreached as long as Anuradhapura was inhabited? that is, for more than 1500 years, a respectable record for a work of such early date. Of no structures can it be said more truly than of reservoirs, that the most successful works have no history. Decade follows decade, century succeeds century, and while the work is performing its functions satisfactorily there is nothing in its life that is worth recording, except the levels of the water in it year by year. Naturally, therefore, we find nothing noted regarding the state of this tank.

Compared with Panda-waewa its area is insignificant; when full it only covers 255 acres, although it appears to have been a little larger originally. Yet it was well designed to fulfil its purpose, the storage of rainfall close to the town, for the water-supply of the city and for bathing purposes. It made the best of a very poor catchment area ; had it been supplied with a higher embankment it would have failed to secure much more water in years of ordinary rainfall. Owing to the smaE area from which the surplus rainfall flowed into it there would be no difficulty at it, like that experienced at Panda-waewa, from very high floods, either during its construction or afterwards.

The plan of the tank on Fig. 70 shows that a much shorter bank might have been carried across the valley in a south-east line from the flood-escape to a projecting point on the opposite side of the reservoir ; but this would have removed the water nearly half a mile further from the early city, whereas the evident aim of the designer was to construct the tank as close to the town as possible. He therefore ran the bank to a position lower down, where on the eastern side the ground level was below that of the water to be retained. From this point he turned the line in an up-stream direction, at nearly a rectangle, until higher ground was encountered. This turning of one end of the embankment upstream is a special feature of the Aauradhapura reservoirs, Pancja-waewa, and Sangili-362 ANCIENT CEYLON

Kanadara tank, described below, and is not found elsewhere in Ceylon, I believe, excepting in the tanks of the Mannar district, where the configuration of the ground, which is practically a sloping plain, rendered it unavoidable.

The embankment is 5910 feet long, or iJ miles. As now restored, its crest is 22 feet above the sill of the sluice; but originally it appears to have been six feet higher, judging by the levels of its more elevated portions. It was considerably eroded, and for a great part of its length the top was below the level adopted at the restoration. The width of its crest was only from six to eight feet, but the slopes on both sides were flatter than at Pan Ja-waewa, being at the rate of 3-1 feet horizontal to one foot vertical. The slope adjoining the water was protected by a layer of small boulders (Fig. 138).

A single sluice was built near the western end ; it consistedr as usual, of a stone-lined rectangular well near the water-level, and a stone culvert for discharging water. This was a work of later date than the embankment, a number of pillars and other stones removed from pre-existing buildings being used in its construction, After it was built a small rice field was formed on the low side of the embankment.

Floods were allowed to escape round the west end of the embankment, through a slight hollow 22 feet wide, the level of which was 19 feet above the sill of the sluice. The present flood-escape is 3 feet 8 inches lower. The original area of the reservoir was about 330 acres, and its capacity about 133 million cubic feet.

There is nothing in the design of the embankment which is indicative of its antiquity. The slopes of the sides were similar to those of many later works, and the weak section which appears to be a primitive characteristic pf Panda-waewa is thus absent. At a little later date it will be seen that it became the custom to make them still flatter. In view of the general features of the design, I am of opinion that several other embankments of considerable size had been constructed in Ceylon before the works at Abhaya-waewa were undertaken.THE EARLIEST IRRIGATION WORKS 363

JAYA-VAPI

Another reservoir made by Pandukabhaya at the same city, before Abhaya-waewa was formed, appears to have been a somewhat large one. The Mahavansa relates of it (i, pp. 42, 43), ' Causing his uncle's canopy of dominion to be brought, and having washed it in the natural tank that was here, this Pandukabhaya caused himself to be anointed king with the water of that very tank. , . . Having deepened the above-mentioned marsh, he made it contain a great body of water. By his having been anointed with that water as a conqueror (Jaya) it obtained the name of Jaya-vapi/

The old name having been changed, this reservoir has not been identified. If it had an embankment and was not

FIG, 117. Pool in Tissa-waewa.J

merely an excavated pool, and if it also covered a large area and was near the town, as the extract would lead one to suppose, it may have occupied the site of Tissa-waewa, the next tank of which I give particulars.

The illustration (Fig. No. 117), from a sketch made by

me in 1873, shows a large natural pool in the bed of Tissa-waewa. Jt is not unlikely that an embankment may have been raised on the low side of this sheet of water, along the line of the present embankment of Tissa-waewa, so as to retain a better supply during the dry seasons, before the construction of Abhaya-waewa. There is some evidence of this in the name of a long channel that was subsequently cut in order to lead water from the great Kala-waewa into Tissa-364 ANCIENT CEYLON

waewa; it bore the name Jaya-ganga, ' the Jaya river' ; this may indicate that it was the channel that conveyed water into the Jaya tank.

A third reservoir at Anuradhapura, termed Gamani-vapi, Is referred to (Mah., i, p. 43) as being in existence to the northward of Abhaya-waewa during the reign of the same king. The name has been changed, and the tank has not been identified. The time of its construction is uncertain ; as the father of Pandukabhaya was named Gamani the tank may have been constructed by him. According to the Maha-vansa he lived at Anuradhapura. Its winding embankment Indicates a possibility that the shallow tank now called Pera-miyan-kulam is this work.

TISSA-WAEWA.

Soon after the middle of the third century B.C. King Deva-nam-piya Tissa formed the Tissa tank at Anuradhapura (Mali., i, p. 79). The account of the incident which led to the erection of the Miriswaeti dagaba by King Duttha-Gamini proves that this is the reservoir on the south-western side of the dagaba ; it still bears the original name. According to the story in the history, the king had gone to bathe in Tissa-waewa, and had set up his sceptre in the ground at the side of it. When he had finished his bath and wished to take away the sceptre it was found to be miraculously fixed, and immoveable. The dagaba was built by the king immediately afterwards, close to this reservoir, and enclosing the sceptre, in order to commemorate the miracle.

The valley in which the reservoir was made is very shallow, and the design took a peculiar form in consequence. A straight bank was raised across the lower part of the ground for nearly three-quarters of a mile, running nearly north and south. From each end of this a long arm was carried in an up-stream direction, forming an obtuse angle with the central part, and being continued until ground was met with sufficiently high to prevent the escape of floods. If Pan Jukabhaya raised an embankment at this place, it must have occupied the line of the central straight part of this bank.THE EARLIEST IRRIGATION WORKS 365

The embankment of Tissa-waewa is 11,000 feet, or miles, in length, and about 25 feet high across the bed of the valley, which is flat for a long distance. The width of the crest averages 12 feet, but in parts is 18 feet. The outer side slopes at the rate of 3 feet horizontal to i foot vertical ; the upper part of the inner slope, adjoining the water, falls at the rate of 2-8 feet horizontal to i foot vertical until at the level of the flood escape it reaches the stone facing of boulders and wedged stones which is laid to protect it from erosion by waves ; this is at an inclination of i| feet horizontal to I foot vertical (Fig. 139). The bank is a well made and substantial work, which with a little attention may last practically for ever ; it appears to be in its original state, and is a credit to the men who raised it. There is no sign that it has ever given way except at the low-level sluice, where there was a small breach when I first saw the work in 1873.

A low-level sluice was built in the northern arm of the embankment, and a high-level one at about the middle of the southern arm. These had the usual rectangular wells and stone culverts. The well at the low-level sluice was nine feet wide in the line of the culvert.

A place for the escape of floods was left at each end of the bank, unprotected by masonry. The level at both was about 15 feet 4 inches above the sill of the low-level sluice ; it is now raised to 17 feet 6 inches. Their total width only amounted to about 50 feet, the catchment basin of this tank being a very small one.

The area of Tissa-waewa was about 396 acres ; as now enlarged it may be 550 acres. Its capacity is unknown.

YAVUNIK-KULAM

The construction of Vavunik-kulam, a reservoir in the Northern Province, should probably be assigned to nearly the same period. Its original name was Peli-vapi, so-called

because it was formed by raising a long embankment across the valley of a stream now termed the Pali river. The single reference to it in the Mahavansa (i, p. 107) shows that it was in existence before the time of Duttha-Gamini, and this may366 ANCIENT CEYLON

carry its construction back to the third century B.C., as the Tamil ruler Elara, who was killed by that king, and who is stated to have been on the throne for forty-four years, from 205 B.C., to 161 B.C., is not known to have made any reservoirs. Of course there remains a possibility that it was formed during the reign of Elara, and that the early annalists omitted to record the fact.

The reference to it in the history is as follows : f To the northward of the capital, at the distance of seven yojanas, in the sand-banks of the stream flowing into the tank of Peli vapi gama, four superb gems, in size about a small grindstone1 and of the colour of the Umma flower, were produced.' The name, the distance from the city, and the reference to a stream with sand-banks render the identification certain, there being no other reservoir on the river, and no other stream with sandbanks at that distance north of Anuradhapura.

According to my hand copy of the inscription left by King Wasabha (66-110 A.D.) at Peramiyan-kulam, on the northern side of Anuradhapura?(No. 7 of Dr. E. Miiller's Ancient Inscriptions)?it appears to have been granted by that king to the Community of Monks. The words in my copy are, Pali nakamka wawiya ma tera Majibaka dini. * He gave the Pali-nagara tank to the great thera Majjhima/ In the forest near the northern end of the embankment, Mr. C. F. S. Baker, the engineer who surveyed the tank, met with some ruins which may indicate the site of this ancient city, Pali nagara.

Detailed surveys have shown that when the reservoir was full the water covered an extent of 1975 acres, and the tank then had a capacity of 596 million cubic feet.

The work was of a different class from those already described, its object being solely the storage of water for the irrigation of the rich lands lower down the valley. Thus it may have been the first large reservoir entirely devoted to such a purpose. The fact that some of the most productive land in northern Ceylon would be irrigable by means of it, accounts

1 I do not what is meant by this, the grindstone not being :

an article.THE EARLIEST IRRIGATION WORKS 367

for the selection of this valley as the site of one of the first large irrigation schemes.

But the provision of water for this purpose would be useless were there not cultivators ready and willing to utilise it; it follows, therefore, that an adequate population who understood rice-growing was already established in this part of the island at this early period. We may safely assume that all could not be the descendants of settlers from the valley of the Ganges ; and if not, the others must have been Dravidians, that is, most probably Nagas, the Vaeddas being a race who were unacquainted with rice cultivation. The raising of such an embankment as that of Vavunik-kulam would necessitate the presence of many hundreds of labourers accustomed to earthwork; the amount of work done itself indicates that there was already a large resident population in the district.

We may feel confident that other irrigation reservoirs of considerable size had been formed before the benefits derivable from large schemes of this nature had become sufficiently well known to induce the sovereign, who of course was the moving spirit in such matters, to undertake the construction of the long embankment of this tank solely for the furtherance of agriculture in an outlying part of his dominions.

The design of the work was of a simple character. A straight embankment was carried across the Pali river, from the northern side of the valley, in a south-south-eastern direction, for a mile and a half. After arriving close to the southern side it was deflected into a south-western line for three-quarters of a mile, so as to include in the reservoir another subsidiary shallow valley, this part of the work being doubtless a subsequent addition to the original scheme. The extended bank ended by being turned round again into a south-eastern curve until it encountered higher ground. The total length is I3.35o feet, or about 2} miles.

Although the river rises only twenty miles away, and the catchment area has a rainfall which amounts to less than 50 inches per annum, the embankment has been badly breached in five places, and the reservoir has been abandoned for many368

ANCIENT CEYLON

centuries, and its bed is now overgrown with jungle. There are also unmistakable signs of former breaches that have been repaired. The only possible natural way in which five breaches can

113. Average S%cti««» c Bank.

FIGS. 1 1 8, 119.

be caused at the same time in an embankment of a reservoir is by the rising of the water until it flows over the top of the

bank at any points where the earthwork is a trifle lower than elsewhere. This is what must have occurred at Vavtmik-THE EARLIEST IRRIGATION WORKS 369

kulam; but it does not indicate, as might appear to be the case, that the space provided for the free escape of all ordinary floods was insufficient. On the contrary it may have been enough, under ordinary conditions, for a reservoir in such a site. In the northern part of the bank a flood-escape at least 80 feet wide, and possibly much more, was left open. At. the southern end a width of 450 feet was allowed at the end of the extended bank, and therefore most probably a wide flood-escape existed at the end of the original bank.

The experience of the last twelve years has shown that it may be suggested with confidence that the bursting of the reservoir was due to one of those violent cyclonic rainstorms which sometimes occur in this Province, and against which in most cases it is impossible to make provision, even if it could be foreseen. On the occasion of such a downfall in another part of the Northern Province in December, 1897, the actual depth of rain which fell in 24 hours, as recorded by three observers at Nedunkeni, one being the Medical Officer of the station and another his dispenser, was 31-72 inches. The enquiry which I personally made on the spot regarding the manner in which this fall was gauged satisfied me that it correctly represented the quantity of rain collected in the rain-gauge, and that in addition a small amount must have been intercepted by two high trees as the wind veered round in their direction. The storm about three hours before this record commenced, the

total amount which fell in twenty-seven or twenty-eight hours must have been 34 or 35 inches deep.

It is almost unnecessary to state that the damage throughout the tract which experienced this cyclonic storm enormous. Roads were washed away, and one iron presented a curious spectacle, standing isolated over the river that it spanned, with the approaches, that is, the road on an embankment at each end of it, more or less carried away. The tanks of the district suffered most; more than 160 were burst, in all cases by the flood-water's pouring in a over the crests of their banks.

One work called "^Periya-kulam, that had by

B B37 ANCIENT CEYLON

the Government, and has an embankment half a mile long, with sufficient space at one end for the escape of all ordinary floods, is in a catchment area that extends only some five miles in length above the embankment. The bank is about 20 feet high. The flood rose until it poured over the whole length of this embankment, and when I afterwards visited it there were several small logs and one large one stranded across its crest, left there by the water on their way over its top. Of course a deep breach was made at a place where this bank at last gave way.

All ordinary precautions against floods must be unavailing when such an outburst as this occurs. The design of the Vavunik-ku}am scheme cannot be considered defective if it failed to meet such a contingency. The tank may have been in good working condition for many centuries before the former breaching of the bank took place, and many more may have passed before its final destruction.

The designer did, in reality, take quite unusual steps to ensure the safety of the reservoir. Though the crest of the bank is only ten or twelve feet wide in the better sections, the up-stream side slopes at the rate of 4-8 feet horizontal to i foot vertical, and the outer one at the rate of 4-6 feet horizontal to i foot vertical. It may be doubted if there are more than three or four other reservoirs in Ceylon with such flat slopes in their embankments. The whole bank is made of good material, and the side adjoining the water is protected up to the ordinary water-level by a layer of small boulders. Under ordinary conditions the work might have survived intact to the present day; but the person responsible for the design could not be aware?as, in fact, no one in the island knew twelve years ago?that this part of Ceylon is liable to experience such frightful rain-storms as that which I have just described?which was perhaps the heaviest that has visited the modern world.

The ctepth of water retained in the reservoir between the sills of the sluices and the level of the flood-escapes was about 18 feet, and the crest of the bank was 8 feet higher.

Only two sluices were found at this work, one being near eachTHE EARLIEST IRRIGATION WORKS 371

side of the valley. They consisted, as usual, of a rectangular well, and a rectangular stone culvert which passed under the embankment. I did not see them ; they were lost in the thorny jungle which enveloped the whole bank, and their sites were unknown when I visited the place. Their wells measured about 10 feet by 15 feet in plan, and were built of brickwork. According to the drawings, the northern one was 80 feet and the southern one 140 feet from the centre of the bank, these being distances that are far greater than those at other similar embankments, in which the well is usually placed near the point where the water-level meets the up-stream slope of the earthwork. This variation from later practice indicates the early date of the works.

PAVAT-KULAM

On account of the dimensions of the bricks used in one of its sluices, another reservoir in the Northern Province, now called Pavat-kulam, twenty ^eight miles south of the last, also appears to be a work of either the third century B.C. or the following century. As its original name is unknown it cannot be traced in the histories, even if it is mentioned in them.

This reservoir was made at the junction of two streams which flow westward through the district to the south of Vavuniya, the total length of the catchment area being about 16 miles; the average rainfall amounts to a little more than 50 inches per annum.

Evidently the valley had been well explored before the position of the embankment was decided upon. Advantage was taken of the presence of a long and high rocky ridge which projected into the valley from the northern side, and the embankment was run in a south-south-westerly direction from its end to a continuation of it two miles away, on the opposite side of the valley, meeting on the way two high rocky detached portions of the ridge. There are thus three separate banks which fill up the gaps left in this rocky ridge. The total length from end to end is 9700 feet, or if miles, of which the artificial bank occupies about a mile and a half. TheS.L Corner of Bisokofuwa. S. Sluice

tfii I v * r t

i^^s^r »«. Plan *««i Section *f S Low letel Sluice

FIQS, 120-1^5, Pftva|-ku|am,THE EARLIEST IRRIGATION WORKS 373

site was undoubtedly the best one in the whole valley for the formation of a storage reservoir.

The tank held a depth of iSfeet of water above the sill of the lowest sluice, up to the permanent level of the flood-escape; its area was 2029 acres and its capacity 779 million cubic feet. The scale of the work therefore resembled that of Vavunik-kulam, but the quantity of earthwork in the bank was much less than at that reservoir. The quality of the soil in the irrigable tract is not so good as at the northern work, and as some difficulty was experienced in providing sufficient space for the passage of floods it is probable that Vavunik-kulam would be the first to be selected for construction. The sole object of the work was the storage of water for the irrigation of rice fields.

The embankment has a total height pf 28 feet in the deeper part of the reservoir; its crest was 8 feet above the permanent water-level. The top is usually about 10 feet wide, but on many sections (at which it may have been worn down) it is from 15 to 25 feet in width. The side-slope on the up-stream face is about 3-2 feet horizontal lo I foot vertical, and on the outer face 2-6 feet to i. The engineers were evidently beginning to recognise that it was unnecessary to give the outer face as flat a slope as that of the inner one. The inner slope was protected as usual by a layer of small boulders and wedged rabble stones, extending downwards from the water level The top has been generally worn down three or four feet below its original level.

In order to allow the passage of floods three places were left open. Owing to the steep ends of the rocky ridges, the designer found it a difficult matter to provide sufficient space for this purpose, and as a matter of fact he must have underestimated the requisite extent; I calculated that with a probable flood of 11,500 cubic feet per second the water would rise within two feet of the crest of the embankment. In such a long bank the settlement or gradual wearing down of the top of the bank to this extent in some places might escape notice, and the result was that the embankment was breached two or three times. The sites of three repaired breaches are374 . ANCIENT CEYLON

visible, one of them being of large size, and at the present day there are two deep ones through which the rivers flow.

At the southern end of the bank there is a flood-escape 125 feet wide in the line of the bank. Its floor and the ends of the embankment at it are covered with large wedged slabs of stone, carefully laid, those on the floor occupying a transverse breadth of 60 feet. This maybe a work of later date than the construction of the reservoir.

This ' waste-weir' is provided with a series of pairs of stone pillars, irregular in size and shape, a short one about two feet high being in front of a taller one five or six feet high, and a few inches distant from it. By the aid of these, a temporary dam of sticks and earth could be raised across the waste-weir after the floods had passed, so as to retain an extra depth of two feet of water. This would increase the area of the reservoir to 2400 acres and its capacity to 972 million cubic feet. A road-bridge of stone slabs laid on stone pillars enabled the stream from this flood-escape to be crossed when a considerable volume of water was passing down it (Fig. 125).

The other flood-escapes were simple overflow channels at rocky sites, one being 25 feet and the other 100 feet wide. At the former the ends of the bank are protected by squared stones laid in steps from the floor upwards.

The extent to which the reservoir was utilised may be gauged by its being provided with four sluices, in addition to a high level culvert under the floor of the southern waste-weir. One of these was a high-level sluice near the northern end of the embankment; the others were much lower, one being in the northern bank, another in the southern section, while the remains of the inlet of the third one can be seen tiear the middle of the work.

The southern sluice was the lowest, and was 18 feet 2j inches below the level of the waste-weir; the northern low-level sluice was i foot 8| inches higher; the northern high-level sluice was 5 feet ij inches above the lowest one; and the culvert at the waste-weir was 9 feet 9 inches above it.

The sizes of the wells at these works were as follows, the longer dimension being the measurement parallel to the lineTHE EARLIEST IRRIGATION WORKS 375

of the bank :?At the southern low-level sluice, n feet 2 inches by 8 feet 10 inches ; at the central sluice, about 8 feet square ; at the northern low-level sluice, 8 feet 2% inches square, and at the northern high-level sluice, 13 feet 8| inches by 8 feet. The culvert at the waste-weir was built of stone, and the well of the northern low-level sluice was also limed with stone, with a substantial backing of brickwork. At the others, the well of the central sluice probably, and those of the other two sluices certainly, were built of stone in the lower part,

FIG. 126. BIsokotuwa. N. Low-level Sluice.

(One end removed.)

with a backing of brickwork, but after the first two or' three courses were finished in brickwork only, the side of the southern low-level sluice adjoining the central line of the embankment being, however, faced throughout with stone slabs (Fig. 124). All

the brickwork was laid in excellent mortar made with lime burnt from coral. The stonework in all the sluices is of

the type of all later works, and consists of long thin slabs of considerable breadth, passing from one side of the wells to the other when laid in their walls. These slabs were placed on edge when used as linings of the wells, and in all sites they were fitted together with great care, The and beds376 ANCIENT CEYLON

of the stones were well, though not finely, dressed, but the backs were left rough (see Figs. 124 and 126).

The dimensions of the bricks employed in the low-level sluices provide the only clue to the age of the reservoir. At the southern low-level sluice their length was 17-36 inches, the breadth 8-60 inches, and the thickness 2-89 inches ; Bt. was 24*8 inches, and the contents 433 cubic inches. The size clearly points to some date not later than the early part of the first century B.C. The figures agree very closely with those of the Sandagiri dagaba at Tissa, which was built by King Kaka-vanna-Tissa in the first half of the second century B.C., and they are also nearly those of other very early dagabas.

The bricks in the northern low-level sluice may be of a slightly later date, as the variations in the length and breadth prove that moulds of a different size were used for them, that is, that they were not burnt at the same time as the others. They have a length of 1670 inches, a breadth of 8-29 inches, and a thickness of 2-94 inches ; Bt, is 24-4 square inches and the contents 408 cubic inches.

The thickness of the bricks at both these sluices is relatively much greater than in those of the Lankarama dagaba, and if that work be excluded the dimensions indicate some period either in the third century, or?if we accept the Sandagiri bricks as our guide?in the second century B.C. Considering the advanced type of the designs for the sluices, the latter is the more probable time.

The southern low-level sluice was of special interest. The unbroken state of practically all the bricks used in the face of the well, and the fact that they were all of one size, prove that this part of the work was the original structure, just as it \fas left by its builders.1 When I saw it twenty-four years ago, it was still fulfilling the purpose for which it was constructed, although the culvert was damaged ; and a small

1 Mr. R, A, Powell, of the Public Works Department, the engineer

who supervised the re-construction of the sluice, has informed me that in the * backing * of the brickwork he found bricks of several sizes. This must indicate some subsequent repairs to that part of the work, although the lining or face of the walls appeared to be intact.THE EARLIEST IRRIGATION WORKS 377,

rice field was supplied with water which passed through it. It is greatly to be regretted that it was taken down and rebuilt according to an 'improved' design a few years afterwards, when the reservoir was partly restored by the Public Works Department. This is the more to be lamented for the reason that in all likelihood it was the only work of the kind of such an age in the island, unless the sluices at Vavunik-kulam are also in their original state.

As in all later sluices, the work in this one consisted of three parts, (i) a rectangular open well built near the point where the water level met the inner slope of the embankment, (2) an inlet culvert through which the water passed into this well, and (3) a discharging culvert from the well to the foot of the outer slope of the bank.

The well is called in Ceylon a biso-kotuwa, which literally means' Queen-enclosure/but probably would be more correctly termed bisi-kotuwa, f the enclosure where (the water level) lowers/ The sketch (Fig. No. 123) shows the manner in which the inner work of three of the sides was built at this sluice. The flooring was formed of long well-fitted slabs of cut stone, like those in the walls. I do not know the thickness of the brick walls ; at other sluices it is often from five to six feet. Mr. Powell stated that the walls were surrounded by very good clay c puddle * for a thickness of two feet or more, and that the brickwork was of such excellent quality that he could not avoid regretting that he had been instructed to pull it down. This well was 14 feet deep; originally it was probably built up to the level of the flood escape, that is, a little more than 18 feet above the sill level. The inlet culvert was 52 feet 6 inches long, and had a peculiar bend in its line, as shown in the plan (Fig. 122). I have seen nothing of the kind elsewhere.. Across its entrance there was a block of brick masonry 7 feet thick and 9 feet long, . which rose 6 feet high above the sill of the culvert and had foundations 3 feet 6 inches deep. No similar construction has been seen at other works. The culvert was rectangular, 2 feet wide and 2 feet 6 inches high at the Met, and 2 feet 6 inches wide and 3 feet 6 inches high at its junction with the3;8 ANCIENT CEYLON

bisokotuwa or well. It had walls 2 feet thick, and was covered by slabs about 9 inches thick. Its floor was at the level of the bottom of the well.

The outlet or discharging culvert was of a very interesting form. For a length of 14 feet 6 inches it was divided into two culverts, each 2 feet square, separated by a wall 2 feet thick (Fig. 124). From the end of these double passages their outer walls were continued in straight lines to the outside of the embankment, gradually approaching each other until they were 2 feet 6 inches apart at the outer end. The height of the passage for the water was, however, gradually increased from 2 feet until it became 3 feet 6 inches at the end of the culvert. The walls were 2 feet thick, and on them were laid large coverstones of varying thicknesses, from about 9 inches to a foot; these were from 5 feet 9 inches to 8 feet long, or more, and like those of the floor were dressed on the face and sides. Across the outer end of this culvert there was a brick wall like that at the inlet, 6 feet high and 12 feet long. The culvert walls were built throughout of large stones, well dressed on the faces, ends, and beds, and fitted together very carefully. For all these measurements of the culverts I am indebted to the drawings of them made when the new work was about to be built.

When compared with later sluices, practically the only difference occurs in the form of the inlet. In most sluices the inlet channel is a very short one, and in large works its entrance is protected by a high wall across it, with sloping wings built at a batter, to support the soil at each side of the approach to it. The increase in the height of the discharging culvert from the well to the outlet occurs at some large works only ; in most cases the section remains the same throughout. It is astonishing to find this early work adhering so closely to the best type of later designs.

No means of regulating the out-flow of the water is visible at any of the ancient sluices in Ceylon, and considerable speculation has arisen regarding the purpose for which the wells were invariably built across the line of the culverts, in the up-stream slope of the embankments. It has beenTHE EARLIEST IRRIGATION WORKS 379

thought that the intelligent engineers who designed these great works may have believed that the culvert was relieved from internal pressure caused by the water in it, when the water was allowed to rise freely in these open wells. This opinion is easily proved to be incorrect. The bisokotuwas, as I prefer to term the wells (the word 'well' usually implying a work with a very different function), are much larger than would be needed for such a purpose, and at the northern high-level sluice at Pavat-kulam we find a larger one than at the low-level sluices, Even when other arrangements were adopted which would really tend to relieve the culvert from excessive pressure?as by enlarging its sectional area from the well to the outlet?we still find the well always present.

As one whose duties permitted him to gain an intimate acquaintance with the ancient works, I have never concealed my admiration of the engineering knowledge of the designers of the great irrigation schemes of Ceylon, and the skill with which they constructed the works; and my friend and predecessor the late Colonel C Woodward, R.E,, expressed the same opinion to me more than thirty years ago, when recommending me to study them thoroughly. When we find, therefore, that the open well is never absent at any sluice in a reservoir, excepting only such works as the culvert under the Pavat-kulam waste-weir, we may safety conclude that it fulfilled a very important function.

Since about the middle of last century, open wells, called * valve-towers' when they stand clear of the embankment and ' valve-pits * when they are in it, have been built at numerous reservoirs in Europe. Their duty is to hold the valves, and the lifting-gear for working them, by means of which the outward flow of the water is regulated or totally stopped. Such also was the function of the bisokotuwa of the Sinhalese engineers; they were the first inventors of the valve-pit, more than 2100 years ago.

It will be readily understood that in an when iron-casting was unknown, and even the smallest plates of iron could be heated only with difficulty in the early forges, no iron or.380 ANCIENT CEYLON

iron-bound sluice valves were made, and that it must have been no easy task to control the out-flow of the water at reservoirs which had a depth of thirty or forty feet, as was the case at several of the larger works. Yet the similarity of the -designs of the bisokotuwas at all periods proves that the ?engineers of the third century B.C., if not those of an earlier period, had mastered the problem so successfully that all others were satisfied to copy their designs.

An examination of the bisokotuwas reveals two invariable -and peculiar features in them : they are always rectangular, and the faces of their walls are never rough or uncut. The commonest type of them is an oblong enclosure, ten or twelve feet long by eight or ten feet wide, built across the culvert at a short distance nearer the water than the point where the water-level meets the slope of the bank. It has thick walls of brickwork laid in mortar, round which there is an excellent watertight backing of tempered clay, or 'puddle/ Where the plan is an oblong the longest sides are always built across the culvert. At most works the brickwork is faced or lined inside the well with admirably cut thin slabs of stone, laid horizontally, and invariably on their edges, which fit closely together. Usually they extend as monoliths along the whole length of each wall, and all have well-cut faces, free from any twist. In some cases there is no facing to the brickwork.

The wedging and accurate cutting of these long stone slabs, which are always of gneiss, must have proved a difficult work in pre-Christian times ; we may guess that their preparation was the most arduous part of the construction of the sluices. As they have rarely a greater thickness than ten or twelve inches, even when they are two or three feet broad, and ten or twelve, or more, feet long, while the brick wall behind them is often six feet thick, it is clear that in most cases they were not used merely in order to increase the strength of the wall. They may have been inserted partly to protect the front of the brickwork, but the accurate cutting of their faces shows that this was not their only purpose.

In my opinion they were intended to permit the accurateTHE EARLIEST IRRIGATION WORKS

fitting, close to the face of the wall, of a further lining of woodwork. This alone will account for the excellent manner in which their faces were cut. It would transform the well into a nearly watertight box.

At a few sluices I have observed indications of the manner in which other woodwork was fixed inside these wells. It was evidently in the form of substantial beams or posts, the-duty of which must have been partly to support the wooden lining of the walls, and partly to carry some form of lifting-gear by which a door or valve might be raised or lowered^ so as to regulate the discharge of the water.

This part of the woodwork appears to have varied in design at different sluices, but generally there was a vertical post about one foot square on each side of the entrance to the-outlet culvert. These must have been supported by horizontal beams which also held up the wooden lining of the walls, some of them probably resting against other wooden posts standing near the corners. At one sluice at Minneriya tank, a work of the third century A.D., square sockets were cut in the floor in order to receive tenons left at the ends of the vertical posts.

At a bisokotuwa at Katiyawa, in the North-central Province,, which tradition attributes to the time of Duttha-Gamini, that is, the second century B.C., there are two lateral recesses, two feet square, at the lower corners of the side walls next to the centre of the bank. In this, as at some few other sluices, there is a wide step of ashlar work in the bisokotuwa, at each side of the inlet culvert, extending up to the side walls; its use is unknown. There is also a cut ten inches wide through the projecting coping stone, above the entrance to the outlet culvert. The age of the work is indicated by the bricks at another sluice ; these are of the large type which belongs to the second half of the second century B.C., or the first part of the first century.

Wooden doors or valves, which might vertically In

wide grooves, must have been placed so as to or

both the culverts leading out of the wells. Most probably were worked by means of levers supported by38a ANCIENT CEYLON

It must have been in order to reduce the friction at these valves that, while there was a single inlet culvert at Pavat-kulam, there were two outlets for discharging the water from the bisokotuwa, each having a sectional area rather less than half that of the inlet, and thus permitting the use of a door or valve of much smaller size than would be otherwise necessary. This is unmistakable evidence that difficulty had been experienced in other works, before Pavat-kulam was made, in overcoming the friction due to the pressure of the water on valves of larger area. A similar arrange* ment is found at many later works.

As no example of the woodwork of the sluices has been preserved, its exact details can only be conjectured; but it is clear from the indications given above that the purpose of these carefully-built bisdkotuwas was to act as true valve-pits. Whatever form the design took it was a triumph of the ingenuity of the ancient Sinhalese engineers, and the more surprising when we find one of the earliest sluices furnished ivith it. Evidently from the first it was a device the general form of which later generations were unable to improve.

It was this invention alone which permitted the Sinhalese to proceed boldly with the construction of reservoirs that ?still rank among the finest and greatest works of the kind in the world.1 Without some efficient means of regulating the discharge of the water through the sluices, the provision of reservoirs for storing it could never have extended beyond the minor tanks. Thus, it may be inferred that the bisokotuwa, with its valves, had not only been designed but had been found to work satisfactorily before the engineers would venture to undertake the construction of Pavat-kulam and Vavunik-kulam, both of which in many years would be of limited use without it. Whether the works of Paiiijuwasa Deva or Pan$ukabhaya were furnished with this means of

1 There are eight or nine post-Christian reservoirs In Ceylon which have areas exceeding 4000 acres ; detailed surveys have been made of one (Maha Kanadara-waewa) which, covered 5670 acres, and of

another (The Giant's Tank), which was not completed, that apparently would have had an area of 6400 acres ; as now restored, the latter

covers 4425 acres with water at a very low level.THE EARLIEST IRRIGATION WORKS 383

regulating the outflow of the water is unknown. In any case it appears to date from either the fourth or the third century B.C.

Every engineer must feel astonishment to observe that the designer of this early sluice enlarged the sectional areas of his inlet and outlet culverts from their entrances to their outlets. He was evidently aware that as the water passed along the culverts the friction of the sides retarded its velocity, and thus rendered an increased space for it necessary in order to avoid undue pressure against the sides and roof. Without such enlargement the resulting increased pressure would tend to force the water through the joints of the masonry, along the back of which it would then flow, gradually removing the soil in suspension until in the end the bursting of the reservoir might be brought about. It is extremely likely that the existing breach at the great Padawiya tank was caused in this manner, and I feel no doubt whatever that other embankments gave way from the same cause; but the designer of Pavat-kulam cannot have had many opportunities of observing such effects, and it is therefore the more surprising to find him taking these precautions against them.

The use of well-tempered clay * puddle' round masonry that was subject to water-pressure was perfectly understood at the time when Pavat-kulam was constructed. It continued to be employed in similar positions at nearly all later sluices, and sometimes round the culverts also. It was always of excellent quality.

SANGILI KANADARA-WAEWA

A smaller reservoir for storing water for irrigating rice lands was formed at an early date in the valley of the Sangili Kana-dara-oya, a small river on the eastern side of the Malwatta-

oya valley. It had not special features like the last, but

was a good average example of a class of reservoirs made solely for irrigation, and occupying a position between the larger village tanks and the great works like those last described. The embankment, instead of running straight across the bed of the valley as usual, was raised for a great part of itsANCIENT CEYLON

length in a north-and-south curve, having its convex side facing the reservoir. Its northern end was turned towards the west for 2000 feet, so as to carry part of the flood-waters clear away from the work ; its southern end, on the othet hand, was deflected sharply eastwards for 1700 feet, to meet high ground.

ItS. $«ctj«n of

FIGS* 127-129. Sanglli-kanadara Tank,

The bank has a total length of 8100 feet, or about a mile

and a half ; and its crest was xyf feet above the sills of the

sluices. Its top was 10 feet wide; and the sides

OB the up-stream face at about 4 feet horizontal to i vertical, and on the outer face at 3-5 feet to i.

The adjoining the water was protected by small boulders op to a of one foot above the level of the waste-weir.THE EARLIEST IRRIGATION WORKS 385

There were three sluices,two at about the same level being in the deeper part of the bank, and one at a slightly higher level near the middle of the southern arm. The middle one (Fig, 128), of which only I have particulars, had a rectangular bisdkotuwa 10 feet 5 inches long, parallel to the bank, and 6 feet 3£ inches wide. There was a single rectangular stone culvert, 13! inches wide and 12 inches high in the inlet portion, and I2J- inches wide and 12 inches high in the outlet part, the latter being raised 3! inches above the floor of the bisokotuwa. The walls, 9 feet 5 inches high, were built of brickwork, and their lower part, for a height of 5 feet 10 inches, was lined with thin monolithic stone slabs, laid on edge.

The brickwork portion of the sluice was repaired or rebuilt several times, there being bricks of four different sizes in it. Some which were 2 inches thick belonged to the tenth or twelfth century A.D., and point to the last restoration of the work. ;

Others, 17 inches long and 2*33 inches thick, may have been burnt in the first or second century A.D. Those of a third type were 18-18 inches long, 9-12 inches wide, and 3-22 inches thick, Bt. being 29*3 and the contents 534 cubic inches. These belong to the period of very large bricks, extending over the second half of the second century B.C. and the early years of the first century-

The fourth type had a length of 1775 inches, and a thickness of 2*75 inches, Bt. was 24-4, and if the width was half the length the contents would be 433 cubic inches. Apparently these bricks cannot belong to the same period as the last ones, and if, as is probable, they are of prior date, they may have been burnt in the third century B.C. It is possible, therefore, that the reservoir may have been constructed at that early period. The flat slopes of the bank also indicate a very early date.

Three flood-escapes were provided; one at the northern end of the bank, measuring 450 feet in width, but probably scoured out and much widened by floods; one at the southern end about 80 feet in width, these being on the natural surface of the ground; and a waste-weir of stone masonry

cc386 ANCIENT CEYLON

built in the angle at the commencement of the southern arm of the bank. Most probably this was of later date than the original formation of the reservoir. It was 270 feet long, and in its form of construction it resembled many subsequent works of the kind (Fig. 129).

In the deepest part it had five courses of wedged and partly-cut stones, the top one being 27 inches deep at the outer face, and the others 18 or 20 inches; each course was set back 3 inches from the face of the course below, and was sunk an inch deep into it. The top of the weir was 17 feet wide, and it had a backing, along the side adjoining the tank, of brickwork apparently laid in mud, to prevent leakage through the stonework. At a distance of iof feet from the outer face there was a row of dwarf cut-stone pillars, about 12 inches square and 2 feet 7 inches high, fixed in the top of the weir at irregular distances, which ranged from 10 J feet to 17 feet. These were evidently placed there in order to assist in raising a temporary dam of sticks and earth after the floods had ceased, so as to retain an additional depth of perhaps 2 feet of water in the reservoir, an extremely hazardous proceeding when the level of the crest of the weir was itself dangerously near that of the top of the embankment.

The crest of the weir was 13 feet 6 inches above the sills of the low-level sluices, and the top of the bank was only 4 feet higher. At the weir level the area of the reservoir was 800 acres, and its capacity 200 million cubic feet. With an ?extra depth of 2 feet of water temporarily retained, the area became 918 acres, and the capacity 275 million cubic feet. The tank has recently been restored with a water-level about 2 feet below the original height of the weir, and an area of 646 acres, which was probably nearly the primitive size of the work.

THE SOUTHERN RESERVOIRS

During the third century B. X, King Maha-Naga, the brother of Devanam-piya Tissa, and tributary king of southern Ceylon, appears to have formed a reservoir called Tissa-vapi, at his capital, Magama. He or his immediate successors, in ther388

ANCIENT CEYLON

latter part of that century or the first half of the second century B.C., constructed also the Duratissa-vapi, * the Far Tissa * tank, as well as another called Digha-vapi, * the Long Tank.' To these may probably be added one now termed Yoda-kandiya, * the Giant Embankment/ the original name of which is unknown.

TlSSA-WAEWA '

The southern Tissa-waewa was made in a shallow valley about a mile and a half east of the Kirindi river, which flowed

FIG- 131. Tissa-waewa? S.P.

past the capital The town occupied the ground between the reservoir and the river, and for some distance lower down the valley, and also extended on the eastern side of the tank. The chief purpose of the work was the storage of water for the use of the city; it is not certain that any rice fields were irrigated by means of it, at any rate in very early times.

Although the area from which water flows into the reservoir is very small, being only some five square miles, it is considerably larger than that of Abhaya-waewa, at Anuradhapura. The rainfall amounts to about 47 inches per annum. The earlyTHE EARLIEST IRRIGATION WORKS 389

designer of the work evidently gave this matter careful consideration, and decided that under such conditions it would be safe to allow a smaller margin than usual between the water-level and the crest of the bank; he fixed this at 5 feet, and his opinion has been justified by later experience.

The reservoir was formed by raising a straight bank about half-way across the bed of the valley until it met a low ridge with two slight elevations on it. From that point it was deflected slightly up-stream, so as to follow this ridge and save earthwork.

In the first century A.D., King Ila-Naga (38-44 A.D.) improved the appearance of the work by abandoning the ridge, and in place of it continuing the straight portion of the bank in one line to the eastern side of the valley* The mounds on the ridge now form two small islands.

As there is no record of its restoration, the tank may have remained in working order until the end of the twelfth century, beyond which time the histories do not contain any references to Magama. At last, however, probably owing to continued neglect of the ordinary works of maintenance, it was breached; and the town, which had evidently dwindled into an unimportant settlement, was totally abandoned, the residents being too apathetic to carry out the small and simple work of repair that was necessary. The whole bed of the reservoir, the embankment, and the former rice fields or the lands on the low side of the bank, as well as the site of the old city, then became gradually overspread by a thick forest growth, infested by wild buffaloes, elephants, and bears. It is clear that the breaching of the embankment must have occurred several centuries ago.

The embankment was about three-quarters of a mile long, and after King Ila-Naga's improvements was practically straight from end to end. It had a top which appears to haw been always used as a cart-road (as at present), and was from 15 to 20 feet wide, with the flat side that characterise

many other early works. The inner at the rate of

about 5*1 feet horizontal to one foot vertical, and the outer one 4*4 feet to one. The level of the of the390 ANCIENT CEYLON

about i8| feet above the bed of the reservoir, in which apparently a depth of 13! feet of water was retained, the area covered being 652 acres and the resulting capacity 160 million cubic feet.

I have met with no reference to any sluice at this reservoir. In one built in 1871,.and afterwards replaced by a larger one, there were no stones of the kind that one would expect to find if the materials of an old bisokotuwa had been utilised in it. It is possible that a small brick sluice may have been constructed long after the original works were made,

A single escape for floods,, about 100 feet wide, was left at the natural ground level on the eastern side, half a mile from the bank, behind some high ground against which the end of the bank abutted. From this, surplus water passed down a depression for three-quarters of a mile, and entered another reservoir now called Yoda-waewa, which appears to be the Duratissa tank of the histories.

At the restoration of Tissa-waewa in 1871, as a great part of the top was found to be much worn away the higher parts of the bank were cut down to the extent of three feet, and the depth of water retained was ten feet. After more than a quarter of a century, however, it was found necessary to raise the water level once more to what seems to have been the height originally fixed by the old Sinhalese engineers. This is a high mark of appreciation of the excellence of their designs and their suitability under the conditions which control such works.

The reservoir was of such Yital necessity to the city that after experience had proved that it often remained unfilled during dry years, important measures were adopted in order to ensure its getting a better supply of water. For this purpose a permanent stone dam was erected across the Kirindl-oya, the river which flowed past the capital, at a distance of two and a half miles from the upper part of the tank. A short shallow channel, with a bed about ten feet wide, was then opened from a point immediately above it in the river, up to a site whence the water conveyed by it could flow into the tank by gravitation, without further works beyond the closing of a hollowTHE EARLIEST IRRIGATION WORKS 391

which led back to the river. The age of this part of the scheme is unknown, but it must be an extremely early work, and possibly the dam was the first one of the kind built in Ceylon. The stones of the dam had been removed before I visited the place; but a few notes on it, in a Report written in 1858 by Mr. G. D, B. Harrison,1 are of interest. It was then altogether broken down by floods. He stated that it had had a height of fifteen feet, and that it' was built of large roughly-hewn blocks of stone, few of which are less than a ton in weight,

133

FIGS. 132, 133.

while many are far more. They appear to have been set dry? or without being imbedded in any mortar. ... A great volume

of water must have passed over the anicut [dam] during the rainy season, and carried with it large trunks of trees, with

a force sufficient to destroy anything but the most massive

masonry/ Floods, or rather the impact of the great tree

trunks that they brought down, did, in fact, eventually destroy the dam, .as well as nearly every other work of the kind in Ceylon.

The body of water which is approaching a dam built across a river extends considerably below the level of the crest ; but

1 Report of the Irrigation Commission, 1867, p» 229.392 ANCIENT CEYLON

immediately before arriving at the up-stream face of the masonry the lower part of the moving water rises so as to pass over it. With it rise any bodies that were being carried near the surface, such as large trunks of dead trees ; these are tilted obliquely upwards, and at that angle may strike the upper stones of the face of the dam. In that case, when the water is moving with great velocity and the tree trunk is

very large__(I have seen one of over ten tons stranded on one

of these dams)?there is great probability that one of tire stones of the top course will be displaced as in Fig. No. 132. It is in this manner that the ancient Sinhalese masonry dams have been breached almost without exception. Among the numerous ancient structures of this kind in Ceylon I have observed all stages of this destruction, from the displacement of the first stone on the up-stream face to the total demolition of the work.

The special point of interest in the Kirindi-oya dam is the astonishing fact that instead of being taken across the river by the shortest possible line, as one would expect, it was built at an oblique angle, which, from the traces I saw, I judged to be nearly forty-five degrees from the direct line. There is a possibility that this does not prove that the principle of the oblique dam, and of its greater discharging power than one built square across a river?the knowledge of which was only acquired in comparatively recent years in Europe?was understood in Ceylon in very early times; Mr. Harrison, in commenting on this oblique dam, stated that in India there was an idea that one built at such an angle would be less exposed to the action of the current than one built square across the river. The Sinhalese possessed profound practical knowledge of the best methods of dealing with water, and the illustrations in Fig. No. 133, of typical dams 40 feet wide, the usual size of the larger ancient works, show clearly that they correct if they believed that such a dam must have much greater stability than one of the same width built square across a river; and especially must be more capable of withstanding violent shocks due to the impact of great tree trunks, than the dam. It is evident that in the oblique dam theTHE EARLIEST IRRIGATION WORKS 393

blow of a log carried by the water would have much less tendency to displace a stone than in the other. Nevertheless nearly all the later dams were built square across the rivers, probably because that was the line of the rocks, on which they were founded.

It is certain that the dam and its channel are not of much later date than the Duratissa tank next described, which in many years could not be expected to fill without their assistance.

THE DURATISSA RESERVOIR

This work is first mentioned in the second century B.C. ; it Is stated that King Saddha-Tissa (137-119 B.C.) built a wihara there (Mah., i, p. 128). The construction of the reservoir is not referred to in the histories; it must have been made by a previous ruler during that century, or late in the third century B.C. Its purpose was chiefly the irrigation of rice lands. There is little doubt that this is the reservoir now called Yoda-waewa.

In the first century A.D., King Ila-Naga is stated to have executed some works of enlargement at it. There are also later references to it, the last one being in the reign of Para-krama-Bahu I (1164-1197 A.D.), when it is included with other large works which he restored; apparently it was then in a breached state.

The embankment, about 3400 feet long, or nearly two-thirds of a mile, was taken in a north-west and south-eastern direction across the mouth of a subsidiary valley to the south-east of the Tissa tank, its south-eastern end abutting against Mgh rocky ground at the point where the valley joins the low lands that stretch for seven miles between the Tissa tank and the sea. The bank was about 14 feet Mgh above the sill of the sluice, and the depth of water retained by it was about nine feet, at which level the area was 1230 acres, and the capacity 336 million cubic feet. The top of the bank was about 15 feet wide, and was doubtless utilised as a cart-road.

A single sluice was built at the south-eastern end of the bank. It consisted of the usual short inlet culvert, bisokotuwa,394 ANCIENT CEYLON

and two discharging culverts, As restored, probably according to the original dimensions, the inlet culvert was 3 feet 6 inches wide and 2 feet 8 inches high; the bisokotuwa was 13 feet 2 inches wide, in the line of the bank, and 12 feet 6 inches long, in the line of the culverts ; and each outlet culvert was 2 feet 6 inches square. These were separated by a pier 2 feet thick. They passed the water into a channel with a base 10 feet 6 inches wide, excavated in decomposed rock. A small dagaba was built on a rock at the side of this channel; possibly this was at the monastery founded by Saddha-Tissa. The work in the sluice and culverts was of the usual type of stonework.

A place for the escape of floods was left open at the side of the sluice, between it and the rocky hill against which the end of the embankment abuts. It was only about 60 feet wide. High floods apparently were allowed to escape round the other end of the bank.

The reservoir received its water-supply partly from some short streams that flowed down from adjoining rocky hills, one being about four miles and another six miles in length ; but its chief and unfailing source of supply was from the flood-escape of Tissa-waewa, over which the water brought down from the Kirindi-oya dam flowed into Yoda-waewa. After these head-works had been constructed there was little fear of any loss of crops in the lands to which this reservoir supplied water; and it is evident that the prosperity of Magama was largely dependent upon them.

Since the restoration of the Tissa tank and Yoda-waewa about 7500 acres of wild forest below them have been converted into rice fields ; and the place, instead of possessing, as formerly, one of the most deadly climates of the island, is now fairly healthy. Numbers of healthy-looking children are to be seen about the houses of the cultivators. There is no place in Ceylon where a greater change has been effected by irrigation.

YOBA-KANDIYA RESERVOIR

This work was formed in a very shallow valley on the western side of the Kirindi-oya, down which a small streamTHE EARLIEST IRRIGATION WORKS 395

flowed into that river. It is opposite the Tissa tank, on the other side of the valley. Nothing is known of its history, the ancient name having been lost. Its sole object seems to have been the storage of water for irrigating rice lands.

The embankment runs in a general north-west and southeast line, and is 11,400 feet long, or 2& miles ; its line forms a long curve and reverse curve, a shape for which there appears to be no special reason in the contour of the ground The side slope3 are extremely fiat, and it is this peculiarity that induces me to include it as one of the very early works. On the inner side the rate of inclination was about 7-4 feet horizontal to I foot vertical, and on the other about 6-8 feet to i. The top seems to have been 15 or 20 feet wide; but all is now very much worn down, and when originally made it may have been higher at these places, and therefore narrower at the top. The depth of water retained in the reservoir if, as at the works on the opposite side of the valley, the flood-escape was at a level of five feet below the crest of the bank, was about 12 feet 6 inches. At this level the area was 1407 acres, and the capacity 380 million cubic feet

No sluice has been discovered in the embankment,1 which is also so much worn down, except at a few points, that it is impossible to recognise the ancient flood-escape.

Doubtless the work would be of a little later date than those nearer the capital, which have been described last. It must have been carried out after the population of the neighbourhood had increased, and required facilities for the extension of cultivation. It is almost certain that water was obtained from the Kirindi-oya, for filling the tank; but no direct channel into it has been discovered in the thick forest and juagle which covers the valley, although one was taken off from the river at a distance of some two and a half miles above the dam which diverted water to Tissa-waewa. After flowing some distance in a cut channel, the water may have been

1 Mr. Hamer, the eugiaeer who has charge of the works of "restoration that axe now being undertaken, informs me that he has not yet dug out the soil in the bed of the main breach. The sluice may have been at this site.396 ANCIENT CEYLON

allowed to find its own way into the reservoir by gravitation. It was in the lands below this work that Gona-gama was situated, the village near the site at which Wijaya is supposed by me to have landed. The pool which still preserves the ancient name is four or five miles from the tank.

DIGHA-VAPI

This reservoir is mentioned in the Mahavansa (i, p. 93) as being in existence during the reign of Kakavanna-Tissa, the father of Duttha-Gamini, that is, some years prior to 161 B.C. Its importance in those early times may be judged from the fact that the king's second son, Tissa, who succeeded Duttha-Gamini on the throne, was specially, stationed at it * to superintend the agricultural works in progress/?-possibly a reference to the reclamation of the irrigable lands to which it supplied water.

The place is occasionally mentioned in later times. In the middle of the seventeenth century, at the time of the first arrival of. the Dutch in Ceylon, the country about it was termed * a rich, prosperous, and populous district' (Hah., ii, p. 332).

This reservoir has never been satisfactorily identified; but as it was certainly in south-eastern Ceylon, and a work of great importance, there is every probability that it is the tank now known as Kandiya-kattu or Maha Kandiya, a reservoir which has been supposed to be capable of irrigating 10,000 or even 20,000 acres of rice fields. The ' prosperous and populous * neighbourhood of the work is totally abandoned, with the exception of two small hamlets; all has relapsed more or less into its original wild forest.

According to the topographical survey, the reservoir is supposed to be narrow, but very long in the direction parallel to the bank. It was formed near the foot of the Kandian mountains, by raising a low embankment across a hollow on each side of a central stretch of high ground, so as to retain a great sheet of water that was perhaps six miles in length parallel to the banks, but possibly less than one mile in width on the average. Although so large, it seems to have had a very limited catchment area, but water may have beenTHE EARLIEST IRRIGATION WORKS 397

diverted into it from an adjoining river. I have not visited the place, and therefore cannot describe the works.

The southern part of what is now the Eastern Province was of so much importance in pre-Christian times that it may be accepted as certain that several other reservoirs were in existence there in the first three centuries before Christ. At present, however, there are no data by which they can be identified,1 and if they are mentioned in the histories their original names are unknown. Some of the works were among the earliest to be restored in modern times, and their masonry structures were pulled down and rebuilt, leaving no trace of their primitive state, of which also no descriptions were preserved.

BATALAGODA-WAEWA

In the account of one of the 'Lost Cities/ Parana Nuwara, I mentioned that the reservoir made at it is of pre-Christian date. Its age is proved by the dimensions of the bricks found at its southern sluice, its flood-escape, and a building which may have been a wihara, close to the southern end of the embankment. Among the nearest dimensions which I have found elsewhere are those of the bricks used in a ahll at Vetfik-kinari Malei, a low hill in the Northern Province, the

inscriptions Nos. .41, 42 and 43 of my list are found at caves, and may belong to the second century B.C. The and thickness of the bricks in the Ruwanwaeli at

Anuradhapura are also similar* Thus the reservoir was when the large bricks were in vogue in the or

part of the first century B.C.

These sizes are?

Breadth* Thickness. Bt.

C« «Bt* X.H.

Southern Sluice . 9-90 ins, 2-83 ins. 2§ 47f

Flood-Escape . 90 (one only) 2-1% 25-7 44^

Wihara ' . . 9-50 (one only) 3'° 28"5 5*3

Vedikkinarl Malei 9-30 s-9° 2; 470

1 Tradition attributes the construction of one or two

ones to Dnttha-Gamim.FIG. 134. Plaa of Batalagoda Tank,THE EARLIEST IRRIGATION WORKS 399

? #- -There is a worn inscription in characters of the tenth

century on a pillar at the embankment, which indicates that it was then restored, or was in working order; and a longer one on alarge slab left there by Queen Kalyanawatl (1202-1208 A.D.), and cut in the third year of her reign, in which she relates that she had examined the sites of * the known sluices/ and had rebuilt one of them, besides causing three breaches to be filled up.

There is no tradition regarding the date when the tank burst again; possibly it was not very long after the time of its restoration, as part of the embankment was covered with large forest trees when I undertook its repair in 1890.

FIG. 135. Batalagoda Tank.

The reservoir was doubtless constructed chiefly for the use of the inhabitants of the early city called Parana Nuwara ; but partly also for irrigating some adjoining rice fields. The bank blocks up the valley of a minor stream; but instead of taking it square across the stream in the usual way the designer wisely adopted an oblique line, in order to utilise some elevated ground, and effect a saving in earthwork. He merely closed up a hollow on each side of this central high ground, and by doing so made the reservoir of greater capacity than if the direct line across the valley had been followed.

The bank was originally 6000 feet, or about i| miles, in length from end to end, but the actual length built was only about 4000 feet. The top was from 10 to 12 feet wide, and' the sides sloped at the rate of 3 feet horizontal to one foot vertical400 ANCIENT CEYLON

in the face adjoining the water, and 2J feet to one in the outer face. The up-stream face was not protected as usual by a layer of small boulders. The total height was about 30 feet. The top was considerably worn down, so that the original level of the flood-escape was uncertain; if, as is probable, it was 13 feet below the crest, the area of the reservoir was 470 acres, and its capacity 141 million cubic feet. As now restored, the tank covers 635 acres.

The sluices were completely destroyed before the modern restoration. Apparently only the upper eight or ten feet of water were drawn off for irrigating or other purposes. A tradition, to which the inscription of Kalyanavati appears to contain a reference, states that the reservoir once possessed seven sluices; it seems to have been without any foundation in fact. It is unlikely that there were more than two, one of them being near the southern end.

Floods calculated at 4000 cubic feet per second are expected to be received by the reservoir from a catchment area of only ten square miles, in which the mean rainfall is about 78 inches. The ancient designer of the works, who may have had experience of floods in much drier districts only, must have greatly under-estimated them ; and totally inadequate space was left for their escape. The breaching of the embankment on several occasions must have been the result. Bricks of four sizes in the southern sluice show that it had been rebuilt three times, and there were three breaches in the bank at the time of the last restoration, as well as in the thirteenth century. These prove that the floods found their way over the crest of the bank on both occasions.

NUWARA-WAEWA

It is probably to the early part of the first century B.C. that the construction of Nuwara-waewa, 'The City Tank/ the

last of the early reservoirs of Anuradhapura, must be assigned. It is on the east side of the Kadamba river or Malwatta-oya, and a mile and a half distant from the present town, in a shallow

flat valley, with a drainage area of about 29 square miles,

from which no excessive floods were to be expected, the rain-THE EARLIEST IRRIGATION WORKS 401

fall amounting to only 55 irithes per annum. The work was utilised partly for irrigating rice fields and partly for supplying water to adjoining monasteries and suburbs.

The embankment follows the example of that at Tissa-waewa, Anuradhapura, that is, the higher portion, a mile long, crosses the bed of the valley, while at each side long arms stretch up-stream at obtuse angles, to sufficiently elevated ground to prevent the escape of floods round their ends. At the southern end of the main bank a long mound of high ground rendered any earth-filling unnecessary for three-quarters of a rnHe; the southern arm began on the opposite side of this. The total length was three miles.

The embankment was 37 feet high in the bed of the valley, above the sill of [the low-level sluice, and from 12 to r6 feet wide on the top. The side facing the water sloped downward at the rate of 3 feet horizontal to one foot vertical, to the top of the wedged stonework or * pitching ' that protected the face from erosion; this began at about 4! feet below the crest of the bank, and was laid at a much steeper inclination, perhaps 11 or 2 to one. The outer face sloped at about 2§ feet horizontal to one foot vertical. The main bank appears never to have given way excepting at one insignificant breach, which may have been cut, but there is some leakage through the soil under it.

This reservoir was provided with two sluices, one being at a low level, and the other having a sill 3 feet i inch higher. At the low-level sluice, the bisokotuwa measured n feet in the line of the culvert, and 15 feet in a transverse direction; it had walls 3 feet 6 inches thick, which rose 14 feet above the sin. It was lined with stone slabs.

There were two inlet and two outlet culverts built of stone. The former were only 17 feet 6 inches long, and were separated by a ma$onry wall 6 feet 6 inches thick; they were 2 feet wide, and 4 feet 2 inches high. An open paved inlet channel, 71 feet 6 inches long and 15 wide, led up to them; this had 3 feet 6 inches thick.

The outlet culverts about 156 feet long, and were

by a wall 7 feet thick. They on a floor 18

D B402

ANCIENT CEYLON

inches thick. Each culvert was 2 feet wide and 2 feet 9 inches high; their outer walls were 18 inches thick, and they were covered with large stone slabs. *

The bisokotuwa of the high-level sluice was built of brick and not lined with stone ; it measured 8 feet 4 inches transversely, and 7 feet 10 inches in the line of the culverts. It was 22 feet high, and had walls 3 feet thick.

136 Plan and Section Of Nuvtara-wcitwa Low-tortl Sluice

131. ft an and Section * Nunara-«va«wa High-level Sluice.

138. Average Section »t BaiawaK-itulam Bank f59. Average Section o* Ti*sa Bank.

TWRy' FIGS. 136-139. Nuwara-waewa Sluices, and Amiradhapura Banks.

The inlet culvert was of a peculiar form. It began inside the reservoir, at 115 feet from the toe of the bank, as a single rectangular stone culvert, 2 feet 9 inches high and 2 feet 6 inches wide, with walls and floor 18 inches thick, and cover-stones one foot thick. At 148 feet from its entrance it was converted into two culverts, 2 feet wide and 3 feet high, with the wall between them, the side walls, and floor 2 feet thick, and cover-stones 18THE EARLIEST IRRIGATION WORKS 403

inches thick. These were 25 feet long up to the interior of the bisokotuwa.

There were two outlet culverts, 14 inches wide and 20 inches high, separated by a wall 2 feet 8 inches thick, having side-walls and cover-stones 18 inches thick, and a floor 2 feet thick. They were 154 feet long, and the total length from the entrance of the sluice to the outlet was 335 feet. A thickness of 2 feet of clay puddle was laid round all the masonry. For these particulars I am indebted to drawings of the sluices made by Mr. W. Wrightson, C.M.G., who carried out their restoration.

The bricks used in this sluice afford the only means of fixing the age of the reservoir. I was unable to measure their length; the breadth is 9-85 inches, and the thickness 3-15 inches, Bt. being 31 square inches. If the length was six times the thickness it would be 18-90 inches, making the contents 586 cubic inches. When these dimensions are compared with those of the bricks laid in the Abhaya-giri dagaba, they are seen to agree extremely closely with them. At the latter structure the length of the bricks is 18-92 inches, the breadth 9-62 inches, and the thickness 3-20 inches ; Bt. is 307 inches, and the contents becomes 583 cubic inches. I conclude, therefore, that the reservoir was made during the reign of Watta-Gamini, in the first twenty years of the first century B.C., or at very nearly that time.

It was repaired at subsequent times. One of these is indicated by bricks which measure 8-48 inches in breadth and 2-64 inches in thickness, to have been about 300 A.D. At a later restoration the bricks were 7-50 inches wide and 2-30 inches thick, a size which points to about the fifth century.

A flood-escape was provided in the high ground to the south of the main bank, at a rocky site. It was 136 feet wide. The sides of the cutting were protected by 'dry stone walling, probably at a later date than the formation of the reservoir. The permanent depth of water retained appears to have been 17 feet; but it seems probable, as the crest of the embankment was 20 feet higher, that a temporary dam of sticks and earth was raised at the site, so as to hold up a considerably greater depth of water. The top of the stone pitching which protected404- ANCIENT CEYLON

the slope of the bank is 14 or 15 feet higher than the rock at the flood-escape, a height that would be unnecessary if an additional depth of water had not been retained. Had this not been the case the southern arm of the bank would also not have been required.

With a depth of 17 feet the area was 2160 acres ; at six feet higher, the level now adopted, which appears to have been nearly the former higher level, it became 3180 acres, according to my tracing of the contours. The capacity then was about 1500 million cubic feet.

Immediately after the reservoir was made the flow off the catchment area must have failed to fill it year after year, and an additional supply of water was discovered to be necessary. This was obtained by taking levels?-(as we may assume)? up the adjoining Malwatta-oya, until a point was reached sufficiently high to permit water to be diverted from it into the reservoir. Above this spot a ridge of rocky ground approached close to the river, and indicated the most suitable place for the dam which was required. At this site, therefore, a strong masonry dam (Fig. 140) of wedged and more or less cut stones was built across the river.

Nearly all the stones were removed in 1873, for use in a road-bridge that was erected over the river. The remains show that the dam was at least 33 feet wide and nearly 160 feet long; it was well and solidly built. It rose about 8 feet high above the bed of the river. At the north end, an abutment roj feet high, of rough stones, laid in four courses, protected the end of the bank of the channel that was cut for conveying the water to Nuwara-waewa.

From this point a channel about 40 feet wide, capable of passing a depth of four feet of water, was opened till it met with a small stream that flowed into the reservoir, at 5 j miles from the dam. The bed of the channel had a gradient of about one foot in 5000 feet, a slope adopted in several later instances.

At 150 feet from the dam, an escape for floods was provided at a rocky site, in order to pass out surplus water when it entered the channel. This was 44 feet wide, and overTHE EARLIEST IRRIGATION WORKS 405

it a bridge 12^ feet wide was constructed, supported by two pairs of wooden pillars for which socket-holes were cut in the rock. The ends of the bank at each side were protected by boulders.

At 6 miles, a bridge 154 feet long crossed the stream down which the water flowed. It was carried on three lines of stone posts, fixed in rows of three, which were 6 or 7 feet apart. Over each set of three posts a stone beam about 12 inches square and 10 feet long was placed; on these, longitudinal wooden beams must have been laid, for carrying the planking of the bridge, as shown in my restoration (Fig 141).

The bricks found at the dam are a proof of its age. They are 9-05 inches broad and 3*25 inches thick; Bt. is 30*9, and the length may have been 18 or 19 inches. It is evident that they belong to the period when the larger types of bricks were burnt, that is, that they must belong to the early part of the first century B.C., since they cannot be of earlier date than Nuwara-waewa.

For several centuries the water-supply provided by these works was sufficient for the requirements of the district and the people below them ; but at length, as the population increased, it became insufficient in dry years. Doubtless it was observed that in flood times the greater part of the water passed over the dam in the river, and especially that when freshets occurred at times when the water was urgently needed, only a limited part of the flood could be secured.

NACCADUWA TANK

A careful examination of the valley showed that at 3§ miles below the dam in the river, two ridges projected into it, leaving a gap of only a mile between their ends. In order to increase the water-supply it was then decided to raise an embankment across the valley at this spot, closing up this gap, and impounding the floods in the reservoir thus formed, which is now termed NaccadGwa. It was a bold scheme, as floods estimated to amount to 14,000 feet per second were to be expected, and there was no suitable rock over which they could be allowed to flow; but it was carried out successfully.FIGS. 140-146. N&ccSdfiwa Tank.

400THE EARLIEST IRRIGATION WORKS 407

The embankment, running nearly north and south, is 5550 feet long, or a little more than a mile. It was 36 feet high above the sill of the sluice, and 55 feet above the bed of the river ; its top was about 20 feet wide, and both the sides sloped at the rate of 2 J feet horizontal to one foot vertical. The slope facing the water was protected by a layer of small boulders.

A single sluice (Fig. 144) was built near the point where the bank crossed the river. It had the usual bisokotuwa, 10 feet 10 inches long in the line of the culverts, and 12 feet 6 inches wide. Its walls were 7 feet thick, but n feet thick on the side facing the tank ; they were 16 feet high. The floor, and the walling for a height of 3 feet, were built of stone; above that level all the work was of brick.

Two inlet culverts, separated by a wall 2 feet thick, passed through the wall of the bis5kotuwa. They were 2 feet wide, and according to the drawings under 2 feet high. An inlet channel 9 feet wide, and 27 feet long led up to them. Its sides were protected by sloping walls of rubble stones, built at a batter.

There were two outlet culverts built of stonework, each 22 inches wide and about 18 inches high, separated by a wall 22 inches thick; they were covered by large thin slabs of stone.

In order to pass out the floods, a fine masonry dam, 44 feet wide at the crest, and 167 feet long (Fig. 145), was built at the point where the embankment abutted against the northern ridge. Its top sloped upward considerably from the back to the overfall, and the back was protected by a mass of brickwork to prevent leakage, although all the inner work of the dam consisted of boulders and wedged stones laid in good lime mortar, as weU as brickwork.

In the deeper part, the work in the down-stream face consisted of seven courses of stones from one foot to 19 inches thick, each course projecting two inches beyond the one above it, which was sunk into it for about an inch. The upper course' projected six inches, so as to form a coping; all the stones in it at the overfall were laid as * headers/ while those at the rear face were "stretchers/ A peculiar feature, which also occurs at some sluice inlets of stone masonry,4o8 ANCIENT CEYLON

was a number of hammer-headed stones laid as headers in the down-stream face, so that the projecting * head;' of the hammer rested against the course above and the course below, to prevent them from moving outward (see Fig. 146).

Stone abutments were built at each end of this dam or waste-weir, with a backing of brickwork laid in lime mortar.

Part of the flood water which escaped over the dam was caught near the point where it rejoined the river, and passed down to Nuwara-waewa, by a channel about 50 feet wide. Possibly other water was permitted to flow down to this channel by a cut opened round the northern end of the waste-weir. Even by this means only limited use can have been made of the reservoir for supplying water to Nuwara-waewa, since only a shallow layer of the upper water can have been drawn off for it; and it is clear that the old channel opened from the dam in the river must have continued to be indispensable. The new tank only supplemented the old works to a small extent; part of its water was used for irrigating the land on the opposite side of the river.

The crest of the flood-escape atNaccaduwa tank was 21 feet 6 inches below the top of the embankment, and 14 feet 2 inches above the level of the sluice. The tank had an area of 2015 acres, and a capacity of 525 million cubic feet. It is now restored so as to retain an increased depth of 8 feet 5 inches, at which the area is 3920 acres and the capacity 1600 million cubic feet.

The bricks used in the sluice measured 8-50 inches in length and 2-58 inches in thickness, Bt. being 21*8. These are the same dimensions as those of some bricks used in the repairs of the high-level sluice at Nuwara-waewa, and they show that the work at both reservoirs was done at about the same time. According to tradition, Naccaduwa tank was made by Maha-Sena (277-304 A.D.) ; the bricks strongly support this date.

The upper part of the bisokotuwa was built chiefly of a later type of bricks, which have a length of 12-55 inches, a breadth of 7*40 inches, and a thickness of 2-04 Inches ; Bt. is 15-1; and the contents 189 cubic inches. They nearly re-THE EARLIEST IRRIGATION WORKS 409

semble the bricks of the twelfth century found at Polannaruwa, but are not so wide; it is possible that they are of a little earlier date. There are also very large rectangular wedge-holes in some stones of the waste-weir, of a type which I have not found elsewhere excepting in twelfth century work, especially that of the time of Parakrama-Bahu I. It is probable that he restored the work, and rebuilt the masonry weir.

When we examine the lists of reservoirs constructed by Maha-Sena and restored by Parakrama-Bahu I (Mah., ii, p. 263), we see that if Naccaduwa be included among them it must be one of two works, (i) the tank called Tissawa, Wadunnawa, or Vaddhana, or (2) MahadaragaUa, Of the rest that are found in both lists, I can identify all but the tank called Cira-vapi or Walahassa, the first name of which, meaning * Small Tank/ shows that it cannot be Naccaduwa. With another Tissa tank at Anuradhapura, this one is not likely to have been termed Tissawa; thus it may be MahadaragaUa.

How long the reservoir remained in order after the twelfth century is unknown. When the recent restoration was undertaken it had a deep breach at the river, and evidently it had been abandoned for many centuries. The whole bed and the embankment were overgrown with high forest, and I was informed that a year before my first visit two bear cubs were captured inside the bisdkotuwa ; this will give an idea of the wild state into which the place had relapsed.

OTHER EARLY WORKS

It is stated in the Mahavansa (i, p. 34) that King Kala-kanni-Tissa (42-20 B.C.) ' formed the great canal called Vanna-kanna, as well as the great Amadugga tank/ but neither of these works has been identified, and the history gives no information regarding their position, nor are they agaia mentioned in it.

A reservoir called Pa$4^~^api is referred to as being in existence during the reign of King MaM-dathika MaM-NSga (9-21 A.D.), and apparently it was made in pre-Christian times ; but aothing is known of its construction. The recurs4io ANCIENT CEYLON

twice or thrice afterwards in the histories, and especially as that of a reservoir which was greatly enlarged by Para-krama-Bahu I, whence it acquired the name Parakkam Samudda. * The Parakrama Sea ' (Mah., ii, p. 148). As the context shows that it was not in the part of Ceylon over which his cousin Gaja-Bahu ruled at that time, it may be the great abandoned tank now called Pandik-kulam, in the southern part of the Uva Province, which I have not examined. It is certainly not Panda-waewa, in the Northwestern Province.

The measurements of the bricks at some reservoirs of smaller size indicate that they also are of early date ; but it is unnecessary in a work of this nature to give a description of such tanks, which cannot be identified in the histories.

Although other works of great interest were constructed at a later period, I include in the present account only the more important schemes which can be shown to have been originated in pre-Christian times.

THE ALLEKATTU DAM

In addition to the Malwatta-oya dam for turning water into Nuwara-waewa, and the Kirindi-oya dam for supplying water to the Magama tanks, I know of only one other pre-Christian masonry dam across a river in Ceylon. It is termed in Tamil the Allekattu, and is built across the KaUaru, the river that flows from the breaches in Pavat-kulam, and forms the principal feeder of the Malwatta-oya in the Northern Province. The dam is two miles above the road bridge over the Kallaru on the road from Mannar to Madawachchiya.

The evidence of its age depends chiefly upon the sizes of the bricks found at it, but partly also on the primitive style of the design. The bricks measure 9-45 inches in breadth, and 3-^ inches in thickness;. Bt. is thus 28-3 inches. If the length was 18 inches, the contents would become 510 cubic inches. It Is clear that they belong to the period when these large bricks were burnt, in the second half of the second century B.C., or the early part of the first century.

The dam, which Is roughly but substantially built, is carriedTHE EARLIEST IRRIGATION WORKS 411

in a north-and-south direction square across the general line of the river, along a ridge of gneiss. It follows the highest line of the rock, and in consequence has two slight bends. Many of the outer stones are roughly dressed, and nearly all are

FIG. 147, The Allekattu ? Dam.

wedged into a shape that in section bears at least some affinity to a rectangle. The inner work consists only of round or shapeless boulders, apparently laid without mortar; they may have been embedded in clay, like those at some other works. All the stone was obtained in the bed of the river, close to the site.

The discharging length is 220 feet. The down-stream face is from three to six feet high, and has no batter; it consists of two, three, or four courses. The top of the dam, which is horizontal throughout, is ig feet wide in the northern part where it is complete, and is formed of six roughly-parallel rows of flat and partly dressed slabs. The northern end has an abutment which is four feet high, and two courses were similarly built at the southern end, with a slight backing of wedged stones and boulders,

Although founded on rock, it was breached by floods in two places, and a third cut was made by them round the southern end,

A small channel was cut on each bank for conveying water to some irrigable land, or perhaps for village tanks

lower down the valley.

Four up the river dam,4i2 ANCIENT CEYLON

called the Kurinja-kulamTekkam, was also built. It is of a rougher type than the last, but may be of later date. Only bricks of a smaller and later size than those at the other work are found at it. A mark of its later date is the upward slope of the top from the up-stream face to the overfall.

This dam is 266 feet long, 20 feet wide at the top, and from 7 to 10 feet high. It consists of roughly-laid gneiss blocks, nearly all being uncut and many being unwedged, which were gathered in the river, close to the work. The downstream face has a considerable batter. Though rough in construction this dam is still unbreached, but the river has cut a new course for itself down the southern channel that was opened from it, re-entering its former bed after flowing down it for some 800 feet.

Probably there are other works of this kind, of pre-Christian age ; but in the absence of bricks of the period of their formation there is no way of identifying them. It is certain that the number is small, since nearly all the river dams of Ceylon exhibit a later type of construction, and consist of masonry laid in lime mortar.

As all the works that I have described are the earliest schemes of the kind, in Ceylon or elsewhere, which can be identified, I have thought it advisable to give exact measurements of them as far as they are available, so as to preserve these in a form suitable for reference by engineers or others who study this subject. The general reader of course cannot be expected to feel much interest in these details, many of which, were they not inserted here, would be lost for ever.Part III ARTS, IMPLEMENTS, AND GAMES

PAGES

XI THE EARLIEST INSCRIPTIONS .... 415

XII THE EARLIEST COINS . . .... 459

XIII THE ANCIENT WEAPONS AND TOOLS . . . 523

XIV THE ANCIENT GAMES . . . . . - 569 XV THE CROSS AND SWASTIKA ? . . . . . 643XI THE EARLIEST INSCRIPTIONS

SINCE 1883, when Dr. Edward Miiller compiled and published for the Ceylon Government the first complete account of the ancient inscriptions then known in the island, much progress in copying others has been made, especially by Mr. H. C. P. Bell, of the Ceylon Civil Service, the present Government Archaeologist, whose excellent and systematic work is of the greatest antiquarian value in preserving complete records of the constructive and epigraphical work of the ancient Sinhalese. There were numberless sites in the jungle where inscriptions have been cut that neither the lamented Dr. Paul Goldschmidt, who was the first to completely overcome the difficulties attending their decipherment,1 nor his successor, Dr. E. Miiller, had heard of; and up to the present day many fresh inscriptions continue to be discovered, and doubtless others will be found for many years to come. This is especially the case with those inscribed on rocks lying on the slopes of the less known hills isolated in the depths of the wild jungle, and often at considerable distances from any villages. Even where such sites occur in the immediate neighbourhood of the jungle hamlets it is generally found that little is known of them by the inhabitants, who have no inducement to make a systematic search for ancient remains.

It would be easy to mention many instances of the annoying manner in which comparatively long inscriptions elude observation even when in close proximity to others that are weU known. On many rocks one may walk over an inscription without suspecting its presence, until some ray of sunlight illuminating one side of the shallow letters and

1 Translations of some inscriptions had been made by Professor Rhys Davids before that time.

4!S416 ANCIENT CEYLON

throwing the other into shadow makes the whole stand out in comparative clearness. This fact indicates one of the difficulties of correctly copying the more worn inscriptions. It is often necessary to have light from two different quarters in order to read them; the morning rays, to catch one side of some letters, the afternoon rays to display others. It too often happens that the passing archaeologist finds it impossible to devote so much time to the decipherment.

In my own experience an excellent illustration of this difficulty occurred. On two mornings I had examined an inscription (No. 88) cut on the flat top of a rock at a distance of four miles from my temporary station, and had obtained a satisfactory hand-copy of three lines of it; yet though it was evidently incomplete and I had had considerable practice in copying such letters I failed to see any continuation of it. On paying it a third visit one afternoon I found that the light, falling from a different direction, lit up the whole remaining line in such a manner that it could be copied with ease.

A trained eye is also necessary in order to distinguish slight artificial cuts from the natural markings of weather-worn rocks. On one occasion I pointed out to a friend who had accompanied me a very early shallow inscription about five feet from the ground on a weathered vertical face of a large rock, and proceeded to copy it without difficulty; yet my friend assured me that he was unable to distinguish a single word of it. All appeared to him like the natural hollows in the face of the rock.

Dr. E. Miiller ascribed the earliest inscription known in Ceylon up to 1883 to either King Duttha-Gamim (161-137 B.C.), or to King Watta-Gamini (88-76 B.C.) ; and stated, without giving reasons for his opinion, that the king's title, * beloved of the Gods/ rather pointed to the latter monarch,1 The date of the first one known at the present day is certainly the third century B.C., and almost contemporary with those of the celebrated Indian emperor Asoka.

It is found at a low rocky hill called Naval Nir£vi Malei

1 Andmt Inscriptions in Ceylon, p* 25.THE EARLIEST INSCRIPTIONS

417

* The Hill of the Jambu Well/ about eight miles north-east of Vilankujam, in the Northern Province. The hill itself is quite inconspicuous and is hidden in the midst of wild thorny

FfQ.t&8. TevBnddn fuliyonKulam HiU.

Inscribed imiWers m Hitt.

jungle frequented by bears, three of which* an adult and two cubs, escaped from an open cave at it on the occasion of my first visit to the place. The top of the hiH is crowned by rocks

EE4i8 ANCIENT CEYLON

and large boulders, a few of which are also on its slopes; the hollows under their sides formed shelters which were improved for the occupation of the monks who took up their residence in them.

There are two other low hills to the south of it, called respectively Tevandan Puliyankulam Malei, and Erupotana-kanda, the three being nearly in a line about one and a half miles long. Erupotana-kanda is a hill somewhat like Niravi Malei, but higher, with numerous large boulders on its slopes. The other hill is formed by an immense steep-sided rock, with a high vertical precipice to the east, and a gradual ascent on the north and south-west sides. There are large boulders on its top, which extends in a long north and south line.

On the detached boulders which are scattered about all three hills numerous cave inscriptions are cut, which indicate 'that this little known part of the island was once the residence of a large community of Buddhist monks. When we seek to learn why such a site should have been selected for cutting what must have been at the time some of the earliest inscriptions in the island, it is found that the explanation seems to lie in the fact that this place was on the line of an early highroad leading from the capital, Anuradhapura, nearly due north-east to the port from which vessels sailed for the eastern coast of India. It is not surprising to find that some of the earliest monasteries were established on this well-known line of communications. The numerous cave shelters and the traditional associations of the Naval Niravi site caused it to be chosen for perhaps the most important of them. At other rocky hills near the same line there are either early inscriptions or other Buddhist remains ; while numerous fragments of an early type of pottery and the early coins found at Mulleittivu, on the north-east coast, and described in another chapter, ?prove that this town also was a pre-Christian settlement.

Of the inscription in question fortunately no less than three copies were cut, each over the entrance of a different rock-shelter, or cave, that had been cleared out and prepared for the occupation of the ascetic monks to whose use it was made over. As is seen in other caves that have been used for thisTHE EARLIEST INSCRIPTIONS 419

purpose down to the present day, the inside was doubtless whitewashed, or even plastered, and a brick or mud wall was built so as to form a protected or enclosed room under the shelter of the rock. At two of the caves a deep cut, termed a katdra, was also made along the rock, above the front of the cave,'and for a short distance below this the face of the stone was cut away, as is usual in nearly all such cases, in order to prevent the rain-water that trickled down the front of the upper part of the rock from entering the room. The cave inscriptions are almost always found in this dressed face of the rock, and two of these are also cut in it, each in a single line.

Two copies are cut over caves or recesses at the north side (No. 2) and south side (No. 3) of the same rock, a large block standing on the top of the Naval Nuravi hill; Fragments of bricks found at them are of three sizes, 3-10 inches, 2*55 inches, and 2-10 inches thick, indicating the use of the caves and the repair of the brickwork from some pre-Christian date down to the tenth or twelfth century A.D. The third copy (No. 1) is in a similar position at a cave to the north of the last. Fragments of brick 3 inches thick lie in this cave, which was therefore also occupied in pre-Christian times.

The inscription which I have numbered (1) was discovered on a visit that I paid to the hills in 1886 with Mr. G. M. Fowler, who was then the Assistant Government Agent of the district ; the other two were found by him on a second examination which he made of the hill in 1887. The hills had been explored some years before our visit by Mr. S. Haughton of the Civil Service, who first drew my attention to the fact that inscriptions were cut at them. He copied a few himself, but was not so fortunate as to discover these earliest ones. I am indebted to him and to Mr. Fowler for copies of all the inscriptions found by them. I have not acknowledged each one separately as I recopied all but one short one myself on subsequent visits.

All the copies of the first inscription made by us were incomplete, owing in two cases to the flaking of the rock, which had destroyed the latter portion of the inscriptions (1)420 ANCIENT CEYLON

and (8), this last being cut in the natural face of the rock ; and in the other case to the rather faint characters, which were at some height from the ground. At a later date, in 1901, I succeeded in copying these by using a rough ladder in order to reach them*

I give facsimiles of all three from my hand copies, arranged one under the other. The inscription No. 1 is twelve feet long to the point where the stone has given way, and the letters are three inches high. No 2 is fifteen feet long, with letters from two to three inches high, cut about a quarter of an inch deep. No. 3 is fourteen feet long as far as portions of the letters remain, and its full length has been about fifteen feet; the letters in it are four inches high and are a quarter of an inch deep.

.{The e of loke is accidentally missing in the copy.)

The complete inscription is as follows *:?

Raja Naga jita Raja Uti jaya Abi Anuradi ca Raja Uti ca karapitase ima lena catu disasa sagaya agatagata na Pasu wiharaye aparim(i)ta loke ditu yasa tana.

Abhi Anuradhi, the wife (of) King Uttiya (and) daughter (of) King Naga, and King Uttiya have caused this cave to be made for the Community of the four quarters, present or future, at the Pasu wihara, an illustrious famous place in the boundless world.

In addition to its age, there are several points of interest in connection with this inscription, the date of which belongs to about the middle of the second half of the third century B.C. In the first place, it confirms the statement of the early annals that King Maha-Naga ruled over southern Ceylon with the

1 In this and other transliterations and names the letter c is pronounced like ckt as in * church * ; the vowels af i% and e have the continental sound ; u is pronounced as in {gun * ; ae represents the sound given to this diphthong in the New Historical Dictionary; *, dt n

are hard as in ' dot * and ' ton * ; I is like // in * full * ; g is always hard as in ' gun * ; t and d are distinctly dental; ^ is very soft, approaching h i the other letters are pronounced as in English.

t

a HAAA

T. w

4- ± J& J 4: ft S K U A 6 X & .ti

^y^aJU

U: bl-UtlClf tbfr-f li^ei Ir

^W -f MO U/ V k

s$

diCr + fl

KA i/

LILU

FIG. 151* Facsimiles of Inscriptions*422 ANCIENT CEYLON

title of king while his brother was supreme monarch at Anura-dJiapura. We learn also that he had a daughter who is not mentioned in the histories, and that she was married to her uncle King Uttiya, an unusual circumstance in Ceylon, although Yatthala-Tissa appears to have married the daughter of his sister, the latter being the Abhi Anuradhi of the inscription. King Wasabha (66-no A.D.) also married his uncle's widow (Mabu, i, p. 140), and other instances of such connections occur in later times.

We may perhaps venture to assume that some idea of the position of women in Ceylon at that early date may be gathered from the fact that her name precedes that of the king. In dealing with the primitive religion I gave another instance of the precedence of a lady, perhaps a century afterwards ; while in the middle of the first century B.C. we find a queen Anula (47-42 B.C.) reigning over the whole country for five years. Also in the inscription numbered 38 it will be seen that the name of a female chieftain, Parumaka Alapusaya, is mentioned. Dr. Davids has drawn attention to the circumstance that women are always placed before men in Buddhist texts.1

It is also clear from the statements in the Mahavansa that from the earliest times women were allowed great freedom and independence in Ceylon. Even if some of the accounts are fabrications of the annalists from whose works Mahanaina compiled his history, the incidents related by them at least prove that they believed such actions of ladies of a high rank to be customary. There is no evidence of the seclusion of women, such as we see in the Ramayana. Thus the Vaedda women are represented in the Jataka story as proceeding to meet shipwrecked traders, who are not reported to evince any surprise at their accosting them without reserve. The Vaedda princess KuwenI is described as marrying Wijaya without waiting to obtain the consent of her parents, who would have refused it in all probability.

In the story of the reception of Mahinda, the first Buddhist apostle, at the royal palace in about 244 B.C., it is stated that

1 The Questions of JKing Milinda, p. 83, note.THE EARLIEST INSCRIPTIONS 423

King Tissa sent for Anula, the wife of his brother, the King Naga of this inscription, and apparently the mother of Queen Abhl Anuradhi, and probably also Tissa's own sister,1 to hear him expound the doctrine of Buddha. ' The said princess Anula proceeding thither, together with five hundred women, and having bowed down and made offerings to the theras [Mahinda and his five companions] placed herself respectfully by the side of them * (Mah,, i, p. 53). In the afternoon when Mahinda was about to preach in the royal garden ' innumerable females of the first rank resorted thither, crowding the royal garden, and ranged themselves near the thera' (p. 54). According to the Dipavansa ' Noble women and maidens, the daughters-in-law and daughters of noble families crowded together in order to see the thera. While he exchanged greetings with them night had fallen' (p. 175).

The name of the place at which the inscription is cut is repeated at a cave lower down the hill in another inscription cut in similar early letters, as follows:?

(4.) Gapati tapasa Sumana kulasa leiie sagasa dine

agata anagata catu disa sagasa Pasu wisaraye. The cave of the family (of) the ascetic Sumana,

the householder; given to the Community, to

the Community of the four quarters, present or

future, at the Pasu tank,

I think that there can be little or no doubt that the monastery was the Paclna, or Eastern, wihara which is recorded (Mah., i, p. 79) to have been established by King Devanam-piya Tissa, the first Buddhist King, and elder brother of Uttiya, who succeeded him. Pasu represents the Pali word Paci, east; several examples of the change of c into s, and i Into u might be quoted.

Tissa ascended the throne in 245 B.C., and is said to have reigned forty years ; but this cannot be trusted, as the reigns of the kings who lived about that time have been extended by the chroniclers in order to make the supposed arrival of the

1 I have already pointed out that the Indian SUkyas from wlaom the royal family were partly descended were accustomed to marry their sisters.424 ANCIENT CEYLON

first Magadhese settlers tinder Wijaya synchronise with the very doubtful date adopted by the Sinhalese historians as the time when Buddha attained Nirvana or died, viz. 543 &.c. The real date was 477 B.C. according to Sir F. Max Miiller; * but doubts have been expressed regarding even this date, and Dr. Fleet has adopted 482 B.C. as a more satisfactory one.

There are no data for fixing the true lengths of the reigns between 245 and 205 B.C., but apparently all have beem doubled in length by the early chroniclers.2 We shall be nearly correct in assuming that the wihara was established in about 235 B.C., and that the inscriptions may have been cut ten or fifteen years later.

The reason why King Uttiya used the term * illustrious famous place' is explained in the Mahavansa (i, p. 75) in the account of the transportation of the celebrated Bd-tree to Anuradhapura. ' On the tenth day of the month, elevating and placing the Bo-branch in a superb car this sovereign [Devanam-piya Tissa] who had by enquiry ascertained the consecrated places, escorting the monarch of the forest, deposited it at the site af the Pacina wihara; and entertained the priesthood [monks], as well as the people, with their morning meal. There (at the spot visited by Buddha's second advent) the chief thera Mahinda narrated, without the slightest omission, to this monarch, the triumph obtained over the Nagas (during the voyage of the Bo-branch) by the deity gifted with the tea powers. Having ascertained from the thera the particular spots on which the divine teacher had rested or taketi refreshment, those several spots he marked with monuments.1

The reference to the action of 'the deity gifted with the ten powers/ that is, Buddha, shows that Mahinda was not relating an incident of the voyage of the Bo-branch, but the manner in which he was supposed to have terrified the Nigas into submission at this place when he came to Ceylon and

1 Dhammapadat p. xxxvi,

?* See my remarks on the chronology of the early kings of Ceylon at the end of this chapter. In the genealogical table I have- allotted those from 245 to 205 B.C. half the time allowed in the Mah£vaasa,THE EARLIEST INSCRIPTIONS 425

visited Nagadipa. When the second note which Tumour inserted in brackets is omitted the meaning is quite cleat.

Thus the words of the inscription confirm the statement of the history that even at that early date the story of Buddha's visits to Ceylon was currently believed. This monastic establishment evidently marks the place at which he was thought to have suppressed the civil war between the Naga kings Culodara and Mahodara, and at which the Rajayatana tree (Kiripalu in Sinhalese, Buchanania angustifolia] of Sakra was planted for the Nagas to worship (Hah., i, p. 6).

There is a discrepancy regarding the site of the Pacina wihara as proved by the inscription and that which is mentioned in the history. According to the Mahavansa, in the quotation just given it would appear to be only half a dayfs journey from the place at which the Bo-tree was landed, but on p. 79 it is said to be at the port itself. I am unable to explain these conflicting remarks; the record left by King Uttiya must outweigh any ideas regarding the site expressed by a monk of Anuradhapura. A similar mistake is made by the annalist regarding the position of the Piyangala wihara, which on p. 113 is represented as being less than two days* march for a monk from Anuradhapura, whereas the actual distance in a straight line is some 63 miles, which the windings of the path would make seventy or more- This wihara was certainly at Kurundan-kulam, and an inscription left there refers to it by name as 'this fearless1 excellent mountain Piyangala * (me abhaya isiri paw Piyangala). Until I studied King Uttiya's inscription I believed that the PScina wifaara was at Piyangala, which is in the midst of wild forest, about 15 south-west of MuUeittfvu.

It is recorded (Mah., ii, p. 58) that Sena, queen of Dappula II (807-812 A.D.), * repaired the terraced house on [at] the Pacina wihara/

It is surprising to read that King (526-539 A.D.)

removed the celebrated * gem-set throne/ over the of which the Naga kings were represented to have

1 The character of the hill that In this

must have been used with the meaning * not causing fear/426 ANCIENT CEYLON ?" .

at the time of Buddha's visit, from the Pacina wihara to a house at the foot of the Bo-tree at Anuradhapura. The throne may have been constructed to suit the stoiy related by Mahinda to the credulous Devanam-piya Tissa, by way of confirming it.

The tank mentioned hi the fourth inscription is a shallow one of eight, or ten acres, with a straight low embankment eight or nine feet high, a typical village tank of the smaller kind, having an inferior water-supply provided merely by rainfall flowing into it from the adjoining jungle for a length of about a mile (see Fig. No. 148).

At the Naval Niravi hill where these inscriptions are cut no remains of a built wihara or a dagaba have been discovered. There is an earthen platform which has a supporting wall of stone, at the western side of the hill. As no traces of a building are to be seen on it it may have been the site of the Kiripalu tree of Sakra, or a Bo-tree. At the southern hill a broken statue of Buddha in a cave proves that a wihara was there at a later time.

As this monastery is of such an early date, and without doubt one of the earliest of which traces have been discovered in Ceylon, I now give the rest of the numerous inscriptions copied by me at the three hills and two others of the neigh* bourhood, some few being nearly as old as those of the king and queen, according to the indication afforded by the shapes of the letters. Unfortunately all are mere dedications of caves to the use of the Buddhist monks.

Other inscriptions at Naval Niravi hill.

(5.) To west of the upper royal cave. Bata Sumanasa lene

sagasa dine.

The cave of the workman Sumana; given to the Community.

(0.) To north of the last. Upasaka Nagaha lene sagasa

dine.

The cave of the lay devotee Naga; given to the Community.1

1 In all cases the words * of - Buddhist Monks * are to be understood as following * Community,'THE EARLIEST INSCRIPTIONS 427

(7.) To west of No. 6. Tisa terasa lene saghasa niyate. The cave of the thera Tissa is assigned to the Community. (8.) To south-east of the upper royal cave. Damarakita

terasa lene ;atu disa sagasa dine. The cave of the thera Dhammarakkhita; given

to the Community of the four quarters. (9.) To north of the last. Damarakita teraha lene sagasa

(letters of first century B.C.). The cave of the thera 'Dhammarakkhita; to the

Community. (10.) To north of NO. 4. Bata Sumanaha lene cadu disa

sagasa.

The cave of the workman Sumana; to the Community of the four quarters.

(11.) To north-west of the upper royal cave. Bata Dama-gutaha Asatisa putaha Asadamarakita lene sagasa agata anagata catu disa.

The cave (of) Asadhammarakkhita, of the son of Asatissa, (son) of .the workman Dhammagutta; to the Community present or future (of) the four quarters.

(12.) Cave full of bats, below royal upper cave, (i) Sagasa; (3) Parumaka Majimaha putasa Paru-maka Sidataha Parumaka Cu4a Sidaha Parumaka Tisaha.

To the Community. (The cave) of the Chief Sidd-. hattha, of the son of the Chief Majjhima; of the

Chief Cucja Siddha; of the Chief Tissa. (13.) Above the last, Bata Budarakitaha matulaniya upa-sika Pusaya le (ne) saghaye niyate (ist cent. A.D.). t The cave (of) the female devotee Pusaya, the aunt

of the workman BuddharakkMta, is assigned for the Community.

The next four are cut over shelters or caves round the overhanging sides of one immense boulder,, each in one line. The inscription No. 4 is also cut at this boulder in a similar position.428 ANCIENT CEYLON

(14.) Matula baginiyana lene agata anagata catu disa

sagasa niyatase.

The cave (of) the sisters (of) Matula; they have assigned (it) to the Community of the four quarters, present or future.

(15.) Barata Mahatisaha lene sagasa niyate; followed by the symbols Fish, Trisula over circle, Swastika,, and Aum monogram! The cave of the royal messenger Mahatissa is assigned

to the Community. (18.) The cave of the royal messenger Mahatissa is assigned

to the Community of the four quarters. (17*) Parumaka Humaneha lene (letters of first cent. A.D.).

The cave of the Chief Sumana.

Inscriptions at Tevandan Puliyankulam rocks, many of them over the shelters formed under large overhanging boulders that lie on the top of the rock (see Fig. No- 149).

(18.) On west side of southern rock. Gapati Vasali puta

Maha Sumanasa,

(The cave) of Maha Sumana,, son (of) the householder Vasali, (19.) On east side of north-west rock. Parumaka Uti

puta Cu$a Nagasa le^te.

The cave of Cu Ja Naga, son (of) the Chief Uttiya. (20,) On south side of north-east rock. Gapati Damasena puta Sumana Malasa ca Gapati Majima Tisa puta Digat(i)sasa ca lene.

The cave of Sumana Malla, son (of) the householder Dhammasena, and of Dlgha-Tissa, son (of) the householder Majjhima-Tissa. Group to the north of these. (JHL) On south side of southern rock. Tebakata Tisa puta

Royogutasa lene.

The cave of Royogutta, son (of) Tebakata Tissa. (22,) On east side of middle rock. Parumaka. Siginika

T(i)saha lene.

The cave of the Chief Singhinika Tissa (Tissa of the Nose!).THE EARLIEST INSCRIPTIONS 429

(23.) Under the last. Barata Utara Kasabaha pati ucaya. The dwelling (?) of the royal messenger Uttara

Kassapa. (24.) On sonth side of north rock. Gapati Pusa.. sa Tisasa

lene.

The cave of the householder Pusa..sa Tissa. (25.) On east side of north rock. Dame davanipi gapatl Visakaha line. The stone-cutter, evidently ignorant of Pali and therefore possibly a Dravidian,, has omitted the lower parts of the letters da and pi, and made mistakes in the vowels. The cave of the devout householder Visakha, beloved

of the Gods. (28.) On south-east side of north rock; Parumaka Asa

Adeka Velasa jaya Tisaya lene. The cave (of) Tissaya, wife of the Chief Asa Adeka

Vela,

(27.) At south-east end of north rock. Parumaka Nuguya Vela putana Sigara Malasava Nuguya Malasava lene.

The cave (of) Sigala Malasava (and) Nuguya (Nud~ guhya) Malasava, sons (of) the Chief Nuguya Vela. (28.) Under-side of east rock. Magasa lene. The cave of

Magha. (89.) Western cave on top of rock. Badira Mahatisa puta

Maha Sumana lene. The cave (of) MaM Sumana, son (of) Mahatissa the

Deaf. (30.) West side of eastern cave, Citagutasa ca Baraxuya

ca lene.

The cave of Cittagutta and Bharaniya. (81.) Southern cave. Ramasi lene.

The cave (of) RSmSsL

(32.) At the side of a flight of steps cut in the rock at the north end of the hill there is an inscriptioB in one line which may be the first instance of what is known in Ceylon as Paeraeli Basa, or transposition of letters in written or spoken words.430 ANCIENT CEYLON

In the facsimile I first give the inscription as it stands, and then a corrected copy, that is, one with the same letters turned round horizontally or vertically or both. It must be read from right to left, and only the consonants in the word savi require transposing, making this word Siva. When thus corrected the inscription is :?

Meka ni salaku savi tipaga pinuvada meda, or when transposed?Dame davanupi Gapati Siva kulasa nikame. The work of the family (of) the devout householder

Siva, beloved of the Gods. Inscriptions at Erupotana hill. *

(38.) South-east cave on south side, Parumaka Ku (4 letters) Siva puta Abayasa lene sagasa niyate. Trisula over circle. The cave of Abhaya son (of) the Chief Ku ....

Siva is assigned to the Community. (34.) South cave on south side. Parumaka Nadika putasa Parumaka Mitasa lene agata anagata catu disa sagasa dine.

The cave of the Chief Mitta, of the son (of) the Chief Nandika; given to the Community of the four quarters, present or future.

(35.) North side of south cave, near the tank. Tisa teraha atevahika Sumana teraha lene agata anagata catu disa sagasa.

The cave of the thera Sumana, pupil of the thera Tissa; to the Community of the four quarters, present or future.

(86.) South side of south cave. Damagutaha lene sagasa.

The cave of Dhammagutta; to the Community.

(37.) South-eastern cave, east side. Parumaka Hadaka

bariya upasika Nagaya ca puta upasaka Tisaha

ca upasaka Deva ca lene agata anagata cattx disa

sagasa niyate.

? .;.. The cave (of) the female devotee Nagaya, wife (of) the Chief Saddhaka, and of the lay devotee Tissa (her) son, and (of) the lay devotee Deva, is assignedTHE EARLIEST INSCRIPTIONS 431

to the Community of the four quarters, present or future.

(38.) Northern cave, containing a broken statue of Buddha. Fragments of bricks in the brick wall of this cave measure 3 inches, 2*30 inches, and 2 inches in thickness.

Parumaka Pita jaya Parumaka Satanasata jita Parumaka Lapusaya lene agata anagata catu sagasa. A symbol follows, apparently a flagstaff surrounded by a fence of four uprights and one cross bar at their top. Possibly it represents the Flag of Victory (of Buddhism), supported by the four great Truths.

The cave (of) the (female) Chief Alapusaya (PAlari-busha), daughter (of) the Chief Santanasatta, wife (of) the Chief Pita; to the Community of the four (quarters), present or future. (39.) At south end of eastern rock. Tisaguta terasa sadi wiharaya barata Majima, . . Tisaya lena sida-sano agata anagata catu disa sagasa neyate. The cave ' Beautiful * (of) the royal messenger Maj-jhima . . , Tissaya, for the excellent wihara of the thera Tissagutta, is assigned to the Community of the four quarters, present or future. (40.) Copied by Mr. Fowler. Barata Tisaha lene.

The cave of the royal messenger Tissa. At Vedikkinari Malei, a hill some miles to the north, near Ariyamacju.

(41.) North care. Parumaka Pusamita puta Ma(jima)ha

lene agata anagata cudi sagaha. The cave of Majfhima, son (of) the Chief Pusamitta ; of the Community (of the four quarters), present or future.

(42.) South cave («). Maha Sarauda puta Gutasa lene sagasa. Parumaka Bamaheta putaha Maha Gutahe l(ene).

The cave of Gutta, son (of) Maha Samudda ; to the Community. The cave of Maha Gutta, son (of) the Chief Brahmahatta.432 ANCIENT CEYLON

(43.) South cave (b). This is another example of ' Paeraeli

Basa/ When the letters are correctly arranged it

becomes Nele hasati dicu taba. It is read from

right to left. The Cave of the workman Cudi Tissa.

At Kaccatkodi, a mile and a half south of Erupotana.

(44.) (i) Senapati puta Parumaka Nadika puta Pama-

tisaha; three dots in a vertical line, forming a

full stop. Parumaka Nataha upasaka, (2) upasaka

Anediya, upasaka Buti Sumanaha (see Fig.

No. 152).

(The cave) of Pamatissa, son (of) the Chief Nandika son (of) Senapati. Of the Chief Nata, the lay devotee ; (of) the lay devotee Anediya ; of the lay devotee Bhuti Sumana.

(45.) Another example of * Paeraeli Basa/ Hagasa nale (Na)la Bati gaba. The inscription is read from the middle outwards, first to the right and then t to the left. The room of Nala Bhatiya, a cave * of the Community. (46.) Asadama Gutaha lene sagasa.

The cave of Asadhamma Gutta ; to the Community.

Some of these inscriptions, especially those at Naval Niravi

Malei, may be as old as the last quarter of the third century

B.C., while the rest with a very few exceptions belong to the

second century and the first half of the first century B.C.

The most interesting inscription after those of the king and queen is No. 34. Strange to say, apparently the same chief caused a similar one to be cut, letter for letter identical throughout the first portion, at the eastern side of a rock termed Kucjirabigala, near Haelawa, in the extreme south-east of Ceylon.1 It runs as follows :?

(47.) Parumaka Nadika putasa 2 Parumaka Mitasa lene Maha Sudasana sagasa dina.3

1 The cave over which it is cut was occupied by a bear at the time of my visit.

2 It is a distinctive feature of this and No. 34 that this word is in

the genitive case.

3 Most probably the right cut at the top of the n was accidental;THE EARLIEST INSCRIPTIONS 433

The cave ' Great Beautiful * of the Chief Mitta, of the son (of) the Chief Nandika ; a gift to the Community. (48.) Another on the west side of the same rock is?

Bata Pusagutasa lene Ma(ha Su)dasana lene sagasa dine. The cave of the workman Pusagutta, the * Great

Beautiful' cave; given to the Community. The bricks in a wall at this cave average 17-20 inches inlength, 8-90 inches in breadth, and 3-16 inches in thickness; Bt. is 28-1 and the contents 484 inches. The size Indicates the second, or early in the first century B.C. as the time when they were burnt.

The inscriptions numbered 84 and 47 are in the earliest characters and appear to date from some time prior to 100 B.C. The most probable explanation of their authorship is that the person who caused them to be cut may be one of the chiefs who accompanied King Duttha-Gamini from southern Ceylon during his war against the Tamils of northern Ceylon. The name of the chief's father renders it extremely likely, or perhaps certain, that the inscription may be attributed to the famous Nandi-Mitta, or Nandika Mitta, the first of the ten celebrated champions or chieftains of King Duttha-Gamini. If so, this would provide a satisfactory explanation of his leaving two inscriptions at places so widely separated.

The fanciful derivations in the histories, out of which some of the champions' names have been evolved, are of course ridiculous. In the case of another of them, Gothayimbara, who is said to have been so called because he was short and was strong enough to uproot * imbara * trees, the writer ignores the fact that Ayimbara was a personal name of the time. An inscription, of perhaps 100 B.C. at Nayindanawa wifeara in the North-western Province runs :?

(49*) Parumaka Mahatisa puta Cu4a Ayimaraha lene , Ayirnare pavatahi.

the interpretation would then become the usual formula * given to the Community/

F F434 ANCIENT CEYLON

The cave of Cuda Ayimbara, son (of) the Chief Maha-tissa, at the Ayimbara mountain.

Thus Gdthayimbara may mean either the e Short Ayimbara/ or Ayimbara son of Gotha. Kota is a nickname now used in Ceylon for a short person. In the same way the story regarding Nandi-Mitta may be put aside as absurd.

We learn from Mah., i, p. 88, that he belonged to a family of high position. His uncle, whose name (Mitta) he bore, was a general (cdmupati) under the Tamil king Elara, and was a native of a village in the north-eastern part of the island, near a hill called Citta, which has not been identified. Nandi-Mitta lived at his uncle's village as a youth, and afterwards with his uncle at Anuradhapura, eventually proceeding to southern Ceylon to join Duttha-Gamini. His residence for some years in the south might enable him to dedicate a cave to the Buddhist monks in that part of the island.

After returning to northern Ceylon as one of the Sinhalese king's leading chieftains, if his native village was in the same district a§ the Pacina wihara, which is equally to the northeast of Anuradhapura, he would be predisposed to do the same for the monks connected with that ' illustrious famous' temple. According to the history he was of a pious disposition and a devoted Buddhist. He is expressly stated to have had the furtherance of that religion in view in joining the Sinhalese prince. " I will bring about the revival of the glory of the religion of Buddha/' he is reported to have said (Mah., i, p. 89). A chieftain of such influence holding these opinions would be certain to make gifts to the monks, and therefore in the absence of any negative evidence there is good foundation for the opinion that it was he 'who caused both the inscriptions to be cut.

In the inscriptions at the Kaccatkodi caves, No. 44 belongs to a Pamatissa who was also the son of a chief called Nandika. The differences between the forms of the letters in this inscription and that of Nandi-Mitta, as seen in the use of the straight r instead of the bent one, and the employment of ha instead of $a for the genitive case, may perhaps point to some other person than a brother of Nandi-Mitta. There still remainsTHE EARLIEST INSCRIPTIONS

435

a possibility that this is one belonging to the same family. The father of Nandika is here termed Senapati, which may be either a personal name, or a title, the General. At this early date one would rather expect it to be the latter, especially as it is not preceded 'by the word Parumaka, Chief, as in the case of that of his son. Thus there is a possibility that he might be the great General of the family, Nandi-Mitta himself, Pamatissa thus being his grandson. Such an identifica-

Fig. 152. Cave Temple, Kaccatkodi.

tion would suit the forms of the letters, and would render it unnecessary to assume that there were two chiefs called Nandi, both closely connected with a Senapati, in the same immediate neighbourhood.

No other names can be identified with those given in the histories. It is surprising to see a female Chieftain mentioned in No. 38; it is the only example of the kind, I believe, but the names of two female Chiefs of the Vaeddas were given in a previous chapter.436 ANCIENT CEYLON

With regard to the characters used, it is interesting to observe in no less than four of these early inscriptions (Nos. 13, 26, 27, and 88) a letter I which in India I believe is only found in southern inscriptions. I am not aware that it occurs in early cave inscriptions in other than the northern parts of the island. It is used in the name of a chief called Palikada, written also Palikada, whose son was the donor of a cave at Wessagiri near Anuradhapura. Dravidian influence appeared to require a letter to represent a cerebral sound of the letter / which is not found in Sanskrit.

lam afraid that it would be unsafe to assume that the names given in Nos. 26, 27, 28 and 31 may be those of Dravidians; there might be such chiefs in northern Ceylon whose families were Buddhists.

Returning to the royal inscription, we already see a great difference between its alphabet and style, and those of Asdka's inscriptions. There are no duplicated consonants, which I think do not make their appearance in Ceylon before the ninth century A.D., while compound letters, excepting in such words as Siddham or Swasti, ' Hail/ or ' May it be well (with you)/ are not found until a still later date.1

The aspirated consonants and, long vowels were already practically abandoned, although an occasional long a and aspirated I, c, dt or g occur in other early inscriptions. The royal grant is, In fact, written in early Elu, or ancient Sinhalese, as much as in the Pali language.

The letter j is already represented by the form employed in India for the aspirated jh; it had nearly disappeared in Ceylon early in the first century B.C. The long initial I is used for the short i, as in the Tdnigala inscription No. 54. A special form of m of a deep cup shape with a central horizontal cross bar, differing from the letter generally used in India, and afterwards abandoned in Ceylon by the end of the second century or early in the first century B.C., had already made its appearance. The trifid s always takes the place

1 Sir A. Cunningham found only three compound letters in the early inscriptions at Sanchi. The Bhilsa Topes, p. 268.THE EARLIEST INSCRIPTIONS 437

of the usual curled letter, which in these forty-nine inscriptions only occurs in one word in No. 20.

These variations in the alphabet prove that writing had already been employed for a considerable period in Ceylon, long enough to allow time for a local development of the letters to take place.

As the bent form of r is alone used in the royal inscription, the presence of the straight form may perhaps elsewhere generally be evidence of a later date than that of inscriptions in which the crooked letter occurs.

With regard to the language used by King Uttiya it is interesting to see the word Una, cave, instead of the usual lene of practically all later inscriptions. It appears to be confirmed by the last word of the inscription, tana. There are only two other special variations from the ordinary language of similar inscriptions found in the island. One is the expression agatagata, na instead of dgata andgata, { coine or come not' in place of * come or not come/ The other is the use of ase f they were/ evidently suffixed to verbs in the sense of * they have/ both in this inscription and in No. 14. The object also is placed after a transitive verb, as we see it in Nos. 53 and 54, below.

In No. 14 the word bhdginiydna is evidently a plural form lil&.putana in No. 27. I have also met with a form ditdna (the last letter being damaged) where the context shows that two daughters are mentioned. The inscription in which it is found is at Kandalawa wihara, in the Kurunaegala district, and is as follows :?

(50.) Parumaka Majimasa gapati Anu(ra)di puta Gapa-tiya dita(na) Tisagutasa Cudasa lene. Tisagutasa Cudasa bata Sumanasa lene saga(sa). The cave of Tissagutta (and) of Cucja, daughters (of) Gapatiya, the son (of) the (female) householder Anuradhi, (daughter) of the Chief Maj jMma. The cave of Sumana, brother of Tissagtitta (and) of Cuda; to the Community.

In the other inscriptions boia appears to represent bhatika, * workman'; it occurs too often in these and many other438 ANCIENT CEYLON

Inscriptions, and almost always before other names, to be a personal name ' Bhatiya/ which in fact is commonly found in the form Bati, as in No. 45, and later examples. Pati ucaya in No. 23 may perhaps be derived from «/vas to dwell. The next inscriptions known are two which are cut at a wihara established under an immense towering rock in the Puttalam district, called Parama-kanda. One of them, No. 51, is cut on the vertical face of a low rock at one side of a small pool of water, termed in Ceylon a pokuna. The other, No. 52, is at a considerable height on the face of the precipice, over the entrance to the wihara. High above it is a nesting place of the Indian Peregrine Falcon (Falco peregrinator) which has doubtless bred there for immemorial ages. The whole site is strikingly picturesque. In the case of both inscriptions a close examination of the letters is not possible on a casual visit. My copies of them are as follows :?

(51.) Two symbols, the second being the fish. Parumaka

Abaya puta Parumaka Tisaha duta kana. The assigned pool of the Chief Tissa, son (of) the

Chief Abhaya.

It has been suggested by Dr. E. Miiller that the letters dutaka may refer to Duttha-Gamini, but it is most unlikely that a sovereign would apply a nickname meaning, 'Angry' to himself in one of his grants. I prefer to assume that the letter pu has been omitted or has been worn away. With it the last word would become pukana, pool, Duta would then be dishta, assigned or ordered.

(52.) A symbol unexplained. Parumaka Abaye puta Parumaka T(isa)ha lene agata anagata caya d(i)sa sagasa.

The cave of the Chief Tissa son (of) the Chief Abhaya. To the Community of the (four) quarters, present or future.

The last part is indistinct; I read it with a field glass in 1876* The word which I copied as caya, six, is most probably catu, four, as usual.

There are two other inscriptions near the same Mil, both on a low rock called Tonigala, the Boat-rock, at the sidejrfTHE EARLIEST INSCRIPTIONS 439

a small tank, Galawaewa, the Rock tank. Their cutting is by far the boldest of any inscriptions in Ceylon. Each is about 100 feet long, with excellently chiselled and quite upright letters a foot high and cut an inch deep in the rock.

(53.) Parumaka Abaya puta Parumaka Tisaha vapi Acagirika Tisa pavatahi agata anagata catu d(i)sa sagasa dine. Two symbols, the first being the fish, followed by three dots in a vertical line as a full stop. Devanapi Maharaja Gamini Abaya niyate Aca nagaraka ca (Tavi) rikiya nagaraka ca Parumaka Abaya puta Parumaka Tise niyata pite raj aha agata anagata catu disa sagasa.

The tank of the Chief Tissa, son (of) the Chief Abhaya, at the Acagirika Tissa mountain ; given to the Community of the four quarters, present or future. (By) the great king Gamini Abhaya, beloved of the Gods, (are re-) assigned * both Aca-nagara and Tavirikiya-nagara (which were) assigned by the Chief Tissa son (of) the Chief Abhaya, father of the king, to the Community of the four quarters, present or future. (54.) Parumaka Abaya puta Parumaka Tise niyate iina vapi Acagirika Tisa pavatahi agata anagata catu d(i)sa sagasa. Emblem and fish, followed by three dots arranged in a vertical line as a full stop. Devanapiya Maharaja Gamini Abaye niyate Aca nagaraka ca Tavirikiya nagaraka ca Acagirika Tisa pavatahi agata anagata catu d(i)sa sagasa Parumaka Abaya puta Parumaka Tisaha visara niyata pite.

By the Chief Tissa son (of) the Chief Abhaya is assigned this tank at the Acagirika Tissa mountain to the Community of the four quarters, present or future. By the great king Gamini Abhaya, beloved of the Gods, (are re-) assigned

1 As the property of the Community of monks.440 . ANCIENT CEYLON

both Aca-nagara and Tavirikiya-nagara at the Acagirika Tissa mountain to the Community of the four quarters, present or future. The tank of the Chief Tissa son (of) the Chief Abhaya assigned by (my) father.

I cannot see any reason to doubt that the inscriptions numbered 53 and 54 belong to the only king of an early date called Gamini Abhaya, who had a father and grandfather named Tissa and Abhaya respectively. They must have been cut by King Duttha-Gamini, who reigned from 161 to 137 B.C. Before he reconquered northern Ceylon, which had been in the hands of Tamil conquerors for some forty-four years, his father and grandfather ruled over southern Ceylon, after Maha-Naga and his son Yatthala-Tissa, as tributary sovereigns under the Tamil king, Elara. The Rajavaliya says (p. 25), * In thosp days King Kawantissa, residing in Magama of Ruhuna, paid tribute to the Tamil king.' This was also the practice while the previous Sinhalese kings held Ceylon, The same work states, * The kings of Magama in Ruhuna and of Kaelaniya used regularly to pay annual tribute to the king of Anuradhapura '? (p. 24).

We now learn from these inscriptions that under the foreign domination they had not even the title of * king/ like Maha-Naga, but were merely termed ' Chief * (Parumaka) like numerous others in the country. Although the title commonly indicated that its bearer was a person of importance in the country, some of these Parumakas occupied subordinate posts, and sometimes were even village headmen. An inscription at Gallaewa wihara in the North-western Province, which having both the bent and straight forms of r and the cup-shaped m, probably belongs to the second half of the second century B.C., runs :? (55.) (i) Barata Maha Tisaye kape (2) Parumaka Naga

gsmiya detake. Cut by the royal messenger Maha-Tissa, the CMef

Naga (being) the village headman.

Tissa, the father of Duttha-Gamini, married the daughter of£another subject king or chief who ruled over the districtTHE EARLIEST INSCRIPTIONS 441

of which Kaelaniya, near Colombo, was the headquarters. It would seem that he acquired or succeeded to his father-in-law's territory, which must have extended far up the west coast, so as to embrace the tract of country in which Parama-kanda is found. At a much later date it is certain that the Kaelaniya kingdom included this district and extended many miles to the north of it, up to the Kala-oya x; and this may have been its limit in earlier times also. This will account for Tissa's being able to make grants to this temple while Elara was ruling at Anuradhapura.

Both the inscriptions at the Parama-kanda wihara purport to have been cut to record grants made by this Chieftain Tissa ; but the difference in the shapes of the letter r in them appears to show that the first is older than the other, which may perhaps have been cut by order of Duttha-Gamini as a record of his father's work at the cave temple. If both were the work of the same stone-cutters it is not likely that such a variation would be made in the forms of the letters. The older one may date from the first quarter of the second century B.C.

In No. 52 and the two following inscriptions we find the straight r always used, and the earliest forms of m and /. The symbol inserted beside the fish does not appear to occur elsewhere in Ceylon, and I offer no explanation of its presence. It is the letter m with a central upright, of the earliest known script, and it occurs in Spain and Egypt (ist Dynasty). I should assign these inscriptions to the middle of the second century B.C.

I place next an inscription over a cave at a large boulder lying on the side of the hill at Dambulla, on the road from Kandy to Anuradhapura. The early shapes of the letters r,

1 This is proved by the list of tanks repaired by Parakrama-Bahu I at the time when he was ruling over only southern Ceylon and Kaelaniya, and Gaja-Bahu was king at Polannaruwa (Mah., ii, p, 149). Those which can be identified extend through the district immediately south and west of the Kala-oya, and include Magalla, Giribawa, Mora-waewa, Maediy&wa, Talagalla, and Siyambalan-gamuwa. On p. 150 the Tabba (Tabbowa) district, which is far north of Parama-kanda, Is referred to as being tinder him.442 ANCIENT CEYLON

m and / render it probable that it also dates from the time of Duttha-Gamini. It is as follows :?

(56.) Damarakita teraha lene agata anagata catu disa

sagasa dine. Gamani Abayasa rajiyahi karite.

The cave of the thera Dhammarakkhita; given to

the Community of the four quarters, present or

future. In the reign of Gamani Abhaya it is made.

In the Dipavansa (p. 209) a thera termed the ' learned

Dhammarakkhita ' is mentioned among those who came from

India at the laying of the foundation bricks of the

Ruwanwaeli dagaba, but there is nothing to prove that he

remained in Ceylon, or that this inscription was cut by his

orders. The name was not uncommon, and is found in the

inscriptions numbered 8 and 9, and elsewhere.

(57.) Of about the same age is one at a dewala, or demon temple, on Dewala-hmna, a hill at Tittawaela, in the Northwestern Province:?

Bata Maha Tisaha lene. Gamani Abayasa rajiya

sika(ka) sagasa. t

The cave of the workman Maha-Tissa. (In) the reign of Gamani AJ haya. To the Community who keep the Precepts (sila).

Possibly the following inscriptions belong to the same period. They are found at Nuwara-kanda in the Kurunaegala district, a hill buried in the jungle, on the bank of the Daeduru-oya. Another inscription of later date informs us that its ancient name was the Tissa mountain. The size of the bricks found there has been given previously.

(58.) Gamika Siva puta Maharajaha ramata Kanatisaha

agata anagata chatu disagasa dine. (The cave) of Kanatissa, devoted to the great King, son (of) the villager (headman) * Siva; given to the Community of the four quarters, present or future. (59*) Gamika Siva puta Garni Kanatisaha lene.

1 Gamika is probably equivalent to the modem word Gamarakt* a village lieadman or elder. Compare No. 55»THE EARLIEST INSCRIPTIONS 443

The cave of the villager (headman) Kanatissa, son

(of) the villager (headman) Siva. (60.) Garni Kanatisaha Badakajaka Anu(rada) ha lene

agata anagata catu disagasaga. The cave of Bhaddakacchaka Anuradha (son) of the villager (headman) Kanatissa ; to the preeminent Community of the four quarters, present or future. (61.) Tisaguta terasa lene.

The cave of the thera Tissagutta.

I omit many cave inscriptions at places where no reference is made to the king of the period, although the forms of the letters indicate that many of them belong to the second century B.C., or earlier.

The next inscription is cut above a cave on the edge of a deep precipice at Mihintale. I examined the letters closely by the aid of a ladder held back by two men and almost overhanging the precipice, so that there should be no uncertainty regarding them. It belongs to Prince Sali, the son of Duttha-Gamini, whose romantic love story is related in the Maha-vansa (i, p. 127), which explains how he abandoned his right to the throne in order to keep his low-caste wife.1

The inscription is preceded by a complex symbol which may represent the Flag of Victory of Buddhism, raised high on a pole which rests on a horizontal base-line. Under the flag, on the same staff, is the trisula resting on the circle, and below this a reversed disk-and-crescent. Four short uprights, two on each side of the pole, which stand on the base-line may indicate the Four Great Truths of Buddhism, or the four-fold

1 * He had a son renowned under the designation of the royal prince

Sali, gifted with good fortune in an eminent degree and incessantly devoted to acts of piety. He became ' enamoured of a lovely female of the Candala caste. Having been, wedded in a former existence also to this maiden,2 whose name was Asokamala, and who wUs endowed

with exquisite beauty, fascinated therewith he relinquished his right to the sovereignty.* She is said -by tradition to have lived at a Duraya village at Hengamuwa, in the North-western Province.

2 His _ grandfather was also believed to have been a pious Candala in his former life.444 ' ANCIENT CEYLON

forces of the sovereign, protecting the symbols. The inscription is very short:?

(62.) Gamini dhama rajasa putasa Aya Asalisa lene.

The cave of the Noble Asali, of the son of the devout

king Gamini.1

Mr. Bell, the Government Archaeologist, met with some cave inscriptions in the North-central Province, left by the sons of Saddhatissa, the brother of Duttha-Gamini One of these, at a hill called Kuda Arambaedda-hinna, which is part of Ritigala, is as follows (Annual Report, 1893, p. 9). The Mng is of course Lajjitissa (119-109 B.C.).

(63.) (La)jaka Tisa maharaje wihara karawaya Abada-

luka wawi saga dinL

The great king Lajjaka Tissa caused the wihara to be made (and) gave the Abadaluka tank (to) the Community. (64.) Another at the same place is?Gamani Abayi kubara

saga dim.

Gamani Abhaya gave the field (to) the Community. Apparently this belongs to Watta-Gamini before he came to the throne in 104 B.C. ; it is noteworthy that he does not give himself the title ' Noble.7

In his Annual Report for 1897, p. 11, Mr. Bell mentions another inscription by Watta-Gainini at Min-vila, and at p, 9 one by Lajjitissa at Duwegala, but gives no copy or transliteration of them.

(65.) He records one of nearly the same period over a cave at Saessaeruwa, in the North-western Province, and gives a transliteration of the first part of it as follows:?

Devanapiya Maharajaha Gamani Abayaha jita

Abi Anuridiya, etc., the rest probably being merely

the usual dedication of the cave to the monks.

Abhi Anuridhiya, daughter of the great king Gamani

Abhaya, beloved of the Gods, etc. Mil Bell attributes it to a daughter of King Watta-Gamini;

1 I have two other inscriptions of one district in the North-western Province, in which a * Parumaka Asaliya' is mentioned, but evidently he is some other person, his father-1 being a chief called Naga.THE EARLIEST INSCRIPTIONS 445

it may be gathered from Hah., i, p. 129, that this king had one whose name is not given by the annalists.

Through the kindness of my friend Mr. F. Lewis of the Forest Department in Ceylon, I am able to add a recently found inscription cut at a cave at Kusalana-kanda, near Ru-gama, in the Eastern Province. It was discovered and copied by his Forest Ranger, and has not been examined by Mr. Lewis ; but it appears to be so important in connection with the identification of the authors of several other inscriptions, that although it may prove to require some correction I now give a facsimile (Fig. No. 151), and a transliteration and tentative translation of it as it stands in the hand-copy sent to me.

(66.) Upaja Naga p(u)te Raja Abaye nama tata-p(u)te Gamani Tisa narnate nakarate sudasane sagasa. Born the son (of) Naga (and) by King Abhaya named (his) ' own son ' (the prince) named Gamani Tissa has prepared the ' Beautiful * (cave) of the Community.

This agrees so accurately with the account in the Maha-vansa (i, p. 129) of Watta-Gamini Abhaya's adoption of the son of his brother, King Khallata-Naga, that it appears to settle the question of the identification of the sovereign called Gamini-Tissa, who is thus Mahacula Maha-Tissa.

It still leaves some difficulties. In the first place, the letters are all of the very earliest shapes, with the bent r, the angular s, and the cup-shaped m; one would not expect to find all these forms still in use during the reign of Watta-Gamini. Secondly, we have the Gal-lena inscriptions which follow, that appear to belong to the same prince, who calls himself in them merely ' the Noble Tissa/ and uses a decidedly later type of letters. The difficulty in connection with the writing may perhaps be explained by assuming that there was still a retention of the old forms of letters in the beginning of Watto-Gamini's reign, while an alphabet more in accordance with that used in India was coming into use by the stone-cutters after he had regained the throne in 88 B.C.

With only two exceptions there is a peculiarity observable in Gamini-Tissa's inscriptions; in six out of the eight that446 ANCIENT CEYLON

are now known he gives a name to the cave that he dedicates to, the monks, while in other cave inscriptions the proportion of the caves so named does not exceed three or four in a hundred. The following inscription was discovered by Mr. H. Nevill at Henannegala, in the south-eastern part of the Eastern Province, and was published by him in the Taprobanian (Vol. i,

p. 38, ff.)-

(67.) Undescribed symbols. Gamini Tisaha pitaha ca Majama Raj aha ca niyata gama nisa paribegani sagasa Giritisa game, Karajinitisa gama, Wila gama, Kasuba nagare Malaga Naka like. The villages assigned by the father of Gamini-Tissa and by the Majjhima Raja (King of the Middle Country) as a resource for the food of the Community (are) Giritissa-gama, Karajinitissa-gama, Wila-gama. Written (by) Malaga Naga of Kas-sapa nagara.

It is rather strange that the name of Gamini-Tissa's father, Khallata-Naga, is not inserted in this inscription. The king of the * Middle' Country, that is, the Malaya or hill district, may have been Watta-Gamini, who took refuge there when the Tamil invaders occupied Northern Ceylon in 104 B.C. Possibly this inscription was cut by Gamini-Tissa during that period.

Following the last we have a series of five inscriptions at Gal-lena wihara, in the North-western Province, all recording the dedication of caves, four of which are distinguished by special names.

(68.) Devanapiya Maha raja Gainarti Abhayasa puta Tisayasa Maha lena l agatanagatasa cat(u) disa sagasa. The ' Great * cave of Tissa the Noble,2 son of the great king Gamani Abhaya, beloved of the Gods; to the Community of the four quarters, present or future.

1 Dr« E. Miller lias lene* Am. Inscriptions^ p. 73. 8 Tfeaya = Tissa + Aya, as at Kota-daemu-hela below, where occur. &e also Dr. Muller's inscription numbered 34 (a) in the son of a King Abliaya is called, Tisaya, ' the Noble Tissa.*«8. ^SivLHOrEAyMrivl/^U^vL^OJlViA^IAAMA*!; I A I

A £ urvdrr mtaflata- Catu, j£ f ? P*

^

T6. d I/ 1 H I- A W ± K A Lr D Y J/ L b # + I B J I v!/ J I A «i.^5^

Ji.f

OambuQa. ^L 0 J ] ? q l7 ttj { f{ ° ^ -1- 31

^ 1/7 X A h ^ ^L/dfc o AJ 6i U xf =L

D

ATI " » e^^aj 8 ir i

X LJDDLlJJJA^r (2) . 03 8 £ 6S a) VI tj 6 J (S) W) 5S t^ a) ZJ 0» 3) 3 (5 SU)

+ 9o ci \Ji. ii b y + Jl o U; i. *o i n 0 J X tiLj 5 i Ji: h tr }xt3

4)ttiT| puwa. ^ j JJ.J E (ti X Ir^X I/ Ir («f ?; O l!rj A UU tw A

rxh (5) Li XX ^ jcJ (61 Ir^'ffl^Clf (ri n }

FIG. .153. Facsimiles of Inscriptions.

447448 ANCIENT CEYLON

Gamani Abayasa puta Tisayasa lene Sihapane1 agata anagata catu disa sagasa. Devanapiya Maha raja Gamini Abayasa puta Tisayasa lene Naga heti agata anagata catu disa sagasa. Under this are two symbols, (i) the Swastika, or magic cross, raised on a pole standing on a horizontal base line from which rise four short upright lines, two on each side, as in No. 62, the tops of the two middle ones being joined by a straight line; and (2) a flag on a standard which rests on an upright cross enclosed in a rectangle. This may be a fence round it. Dr. Miiller erred in placing this inscription at Giribawa ; it is cut over the wihara at Gallena.

The cave ' Sihapane' of Tissa the Noble, son of Gamani Abhaya; to the Community of the four quarters, present or future. The cave ' Under-the-Rock * of Tissa the Noble, son of the great king Gamini Abhaya, beloved of the Gods; to the Community of the four quarters, present or future.

The stone-cutter was an ignorant man who began to cut the word anagata in place of Naga, and then cancelled the initial a. He may have made some mistake in the second word, which I take to be a name of the cave, and possibly intended iGr.Sihapdhane,' the Lion Stone/ Compare Nila/pa-natata, lit. 'the Blue Stone plain/ in an inscription at Ridi wihara, given with others at that place in the account of the Ancient Weapons. Nila pana is equivalent to the modern kalu gala, gneiss.

(70.) Devanapiya Maha rajasa Gamini Abhayasa puta

Tisayasa lene Sita guhe agata anagata catu disa

sagasa. At the end are the same two symbols as

in No. 69,

The cave ' Cool Cave * of Tissa the Noble, son of

1 Dr. Miiller informed me that the initial is 5 and not P as I copied It In 1878.THE EARLIEST INSCRIPTIONS 449

Gamini Abhaya, of the great king beloved of the Gods; to the Community of the four quarters, present or future.

(71.) Devanapiya Maha raja Gamani Abayasa puta Tisayasa lene Cuda Sudasana agata anagata catu disa sagasa.

The cave ' Small Beautiful *? of Tissa the Noble,

son of the great King Gamani Abhaya, beloved

of the Gods; to the Community of the four

quarters, present or future.

(72.) Devanapiya Maha raja Gamini Abayasa puta

Tisayasa lene agata anagata catu disa sagasa. The cave of Tissa the Noble, son of the great king Gamini Abhaya, beloved of the Gods; to the Community of the four quarters, present or future.. This is cut over the wihara. #

With regard to the names of these caves, various titles of such dwellings are sometimes met with elsewhere, as in Nos. 47, 48 and 81. An inscription in letters of the earliest type at Periyaka Juwa wihara, in the North-western Province, runs :? (73.) Symbol, an upright plain cross with wide arms each consisting of two lines joined at the ends by transverse ones. Parumaka Nakatika Tisa puta Pani-maka Sumanasa dane. Five dots in a vertical line, making a full-stop. Maha Sudasane nama lene sagasa,

The gift of the Chief Sumana, son (of) the Chief Nakatika Tissa. The cave 'Great Beautiful* by name ; to the Community. The name of the donor's father may perhaps be Naga Tikkha Tissa, or he may have belonged to a village of the district now called Naekatta,

Another in characters of the first century B.C., at Rankiri-jnacja wihaxa, in the same Province, is:?

(74.) Gamika Wasabayi Parumaka Wasabaya tiba nami

lene.

The villager (headman) Wasabhaya's cave, which has the name * the Chief Wasabhaya.*

G G45o ANCIENT CEYLON '

The shapes of the letters in all the Gal-lena inscriptions are distinctly those of the first century B.C. At that period there was only one king, Watta-Gamini, who was called Gamini Abhaya, and his adopted son -Mahacula Maha-Tissa must have caused the inscriptions to be cut while the king was still reigning, and probably, as he is entitled ' Devanam-piya/ in the latter part of his reign, that is, about 80 B.C. The ornis-^ sion to mark the long a or aspirated b in some of them is not unusual elsewhere; it is, in fact, the general rule in Ceylon. (75.) After Gamini-Tissa succeeded to the throne he made over the great Dambulla cave to the monks, and left there the following inscription :?

Symbol, a Swastika elevated on a pole with two short vertical bars on each side of it rising from the base line on which it stands. Devanapiya Maha rajasa Gamini Tisasa maha lene agata aaagata catu disa sagasa dine. The great cave of the great king Gamini-Tissa, beloved of the Gods ; given to the Community of the four quartets, present or future. (78.) After this we have one at Mihintale cut by his wife. Maha rajaha Gamini Tisaha bariya upasika Ramadaraya l(e)n(e) sagasa. The cave (of) the female devotee Ramadharaya, wife of the great king, of Gamini-Tissa; to the Community.

We learn from the Mahavansa that Mahacula had two wives. One was the notorious Anula, the mother of Kuda-Tissa, whom his brother married after his death; the other, who became a nun, was the mother of Kalakanni-Tissa. Evidently it was she who caused this inscription to be cut.

Prince Gamini-Tissa must have been more than a youth when Ms uncle Watta-Gamini adopted him on his accession, as Ms son; or the succession would not have been secured to Mm in preference to the king's own son. It may be conjectured that it was a politic act of the king to pacify the party who supported Tissa's claim to the sovereignty. As the Tamil invaders afterwards held northern Ceylon for fifteen years,THE EARLIEST INSCRIPTIONS 451

he may have been between thirty and forty years old when Watta-Gamini regained the throne in 88 B.C. He might possibly cut the Rugama inscription immediately after his adopted father began to reign, that is, in 104 B.C., before the latter had built the Abhayagiri dagaba and wihara and thereby .acquired the title Devanam-piya. At the latest, it must have been cut soon after 88 B.C. The inscriptions at Gal-lena may have been cut shortly before Watta-Gamini's death, after he had acquired the title.

Next come the inscriptions cut by the sons of Gamini-Tissa. (77.) Mr. Bell found an inscription at Andiya-kanda,

another part of Ritigala, which runs :? Devanapiya Maha raja Gamani Tisaha puta Devana-piya Tisa A? lene agata anagata chadu disa sagasa lene.

Mr. Bell fills the blank in the second name by making the word Abaha, but the inscriptions which follow indicate the expression Ay aha, and the translation would then be :?

The cave of Tissa the Noble, beloved of the Gods, son of Gamani-Tissa, the great king beloved of the Gods, A cave of the Community of the four quarters, present or future.

This inscription may belong to Prince Kuda-Tissa, and the absence of the royal title in his case shows that he had not succeeded to the throne, that is, it must have been cut before 50 B.C., and most probably during his father's lifetime.

Next comes the inscription at Nuwara-gala in the Eastern Province, which was republished in 1907 by Mr. F. Lewis, in the Journal of the Ceylon Branch of the Royal Asiatic Society. It was discovered and first published by Mr. NevUl in the Taprobanmn, vol. i, p. 150. The forms of the letters /and m prove that it belongs to the first century B.C. (78.) Devanapiya maha rajaha Gamini Tisaha puta

Maha Tisa Ayaha lene sagike. The cave belonging to the Community, of Maha-Tissa the Noble, son of the great king beloved of the Gods, Gamini Tissa. The son who left this inscription may be the one who became452 . ANCIENT CEYLON

King Kalakanni-Tissa (42-20 B.C.). Mahacula's other son being called Kuda-Tissa in the Mahavansa apparently cannot be the prince here termed Maha-Tissa.

Kalakanni-Tissa is mentioned as a devout Buddhist, and for some time he actually became a monk. As he does not term himself king, this record must date from prior to 42 B.C.,, and probably from his father's reign.

After this we have thirteen inscriptions at Kotadaemu-hela and neighbouring rocks in the southern part of the Eastern Province, which were discovered and published in the Tapro-banian (vol. i, p. 150) by Mr. H. Nevill. All are described by him as very nearly identical I therefore give only one transcript, adding the word jaya, as found in some others.

(79.) Dama raja puta Maha Tisa Ayaha jita Aya Abaya puta Aya Tisaha (jaya) Abisawaraya dana sagasa. A gift to the Community by Abhisawara, wife of the Noble Tissa, son (of) the Noble Abhaya, daughter of Maha-Tissa the Noble, son (of) the devout king.

In some of the inscriptions Maha-Tissa Js daughter is termed * Abisawara Ayabaya/ and her son's name is also written Tisa Aya and Tisaya.1

As the name of the ' devout king9 is not given, and Mr. Nevill did not publish a facsimile of the inscriptions, there is some doubt regarding the identification of his son Maha-Tissa. At present I can only assume that he is the same person as the Maha-Tissa, the son of Gamini-Tissa, who cut the inscription numbered 77. The fact that he does not receive the title of king indicates that at the time when these were cut he had not succeeded to the throne. For this to be the case it is very evident that early marriages must have been the custom in the royal family, and even in that case Gamini-Tissa must have been born before 120 B.C. for his great-grandson's wife, whose age would not exceed her husband's, to be

1 It is uncertain if the expression sawara indicates a connection the Vaeddas, Sawam usually stands for so&am, * barbarian/ or in Ceylon, * Vaedda/THE EARLIEST INSCRIPTIONS . 433

old enough to cause these inscriptions to be cut before 42 B.C.; the year when Kalakanni became king.

The inscriptions are interesting as showing that at this period all the. members of the direct line of the royal family had the title Aya (=Ariya), e Noble/ instead of Prince and Princess. This title is applied to the princes in Prince Sali's inscription, those at Gal-lena, probably that at Andiya-kanda, three others at B5wata which follow, and No. 34^ of Dr. Miiller's work.

It would seem that Gamini-Tissa's grand-daughter, Abhaya, had married some local chieftain of south-eastern Ceylon, and that her son, whose wife caused these inscriptions to be cut, continued to reside in that district.

Three inscriptions were cut at some caves at Bowata, in the extreme south-east of Ceylon. These also were found by Mr. Nevill, and published without facsimiles in the Taprobanian (vol. i, p. 52 ff.). They are as follows:?

(80.) Symbols, Stlla and fish. Maraja putha Maha Tisa Ayena karite. (This) is made by Maha-Tissa the Noble, son of the great king.

(81.) Symbols, Fish and sula. Samanaha tedasa Batika Na puta sawa putaha pute dama raja dama raki(ta) ra(ja) Maha Tisa Aye karite (i)ma lena Maha Sudasana sagasa dine.

This cave,f the Great Beautiful/ is made by Maha-

Tissa the Noble, son (of) the samana (monk), the

famous Bhatik a-Naga, the (best) son of all sons,

the devout king, the king who protected the

Dhamma (religion); given to the Community.

Without facsimiles of these inscriptions any identification

of the prince who caused them to be cut must be tentative.

There is only one king called Bhatika in the first century B.C. ;

he began to reign in 20 B.C., and was apparently the brother

of the Princess Abhaya of No. 79. His name was Abhaya,

and as his younger brother was called Naga there could be

no reason for terming him Bhatika (the elder brother) unless

his brother's name was also Abhaya in addition to Naga,

or his own name was Naga in addition to Abhaya. In the454 ANCIENT CEYLON

same way, two kings called Tissa were discriminated in the second century A.D., the elder one being called Bhatika. Provisionally, therefore, I attribute these inscriptions to a son of Bhatikabhaya (Naga) who is not mentioned by the historians ; the king's title ' Maharaja' in No. 80 shows that he was the supreme king of Ceylon, and not merely a subordinate ruler of southern Ceylon. The character of the father agrees with that of Bhatikabhaya, who was a most devout king.

The third inscription at Bowata is:?

(82.) Undescribed symbols, Samanaha Tedapana Tisa raja Uti puta Aya Abayasa jita Abi Anuradiya. (The cave of) Abhi Anuradhiya, daughter of the Noble Abhaya, son (of) King Uttiya, (son of) the samana (monk) Tedapana-Tissa.

This King Uttiya may have been a king of Ruhuna or southern Ceylon, there being no ruler among the kings of Anuradhapura who can be identified by this name, except the first one, whose father died before Buddhism was introduced into the island.

With this, my list of the earliest inscriptions is ended. I believe that it includes all royal inscriptions cut prior to the Christian era, so far as they are known at present,1 unless Mr. Bell has found some that are not yet made public. It is unnecessary to give transcripts of numerous others that merely record the dedication of caves by unknown persons at unknown places. The latest record I have seen, in an inscription, of the grant of caves to the monks is contained in one by King M%hawannabhaya II (304-332 A.D.).2

I add one other rock inscription of the second century A.D., as it was covered up when the embankment of Iratperiya-kulam, a reservoir near Vavuniya in the Northern Province,

* Dr. Muller has one (No. 34a) of a prince who cannot be identified. It is:?Pacina raja puta raja Abayaha puta Tisayaha lene agata, eta 1 The cave of Tissa the Noble, son of King Abhaya, son (of) _ the King

of the East (or Pasu country); to the Community, etc.* It appears to be of early date.

* Mr. Bell found one record of the twelfth century. (Arch. Survey, Annual Report for 1897, P« 9-)THE EARLIEST INSCRIPTIONS 455

was restored, and therefore will not be seen when the inscrip~ tions of that Province come to be copied.

(88.) (i) Sida. Sata nparaja bare Ti(ya)gasala pa (2) rinika parasiha pite Gainini Abha raja (3) ha hamaneka udi Alavicaka haraha tire Tihadaya wiharahi (4) bhiku sagahata dine. Hail! The son-in-law (of) the wise sub-king (and) father (of) Tiyagasala, the pre-emiaent hero, the Crown-prince, King Gamini-Abhaya has gives six amunas of undi (pulse) to the Community of monks at thd Tihadaya wihara, on the shore of the Alaviccha Lake.

This inscription belongs to Gaja-Bahu I (113-135 A.D.) whose father-in-law, MahaJlaka-Naga, succeeded him. The prince who is referred to in such unusual terms is not men' tioned in the histories by this name. He would appear to have greatly distinguished himself in the invasion of Madura,1 the only war in which Gaja-Bahu is known to have been engaged. He may have died while his grandfather or uncles held the throne, as the prince who succeeded them, from 193 to 195 A.D., was called Culanaga. From this record we learn that the ancient name of the tank was the * Scorpion Lake/ and this enables the Gonusu (Scorpion) district which is mentioned in later times to be identified as this part of the island.

I annex a genealogical table of the early kings of Ceylonu The date when Devanam-piya Tissa ascended the throne Is practically certain within a few years* The Dipavansa states that it occurred seventeen and a half years after the accession

of the Indian emperor A£6ka, which may lie within four years of 263 B.C. Dr. Duncker, in his History of Antiquity, vol. ii, p. 525, has fixed it at 263 B.C. Sir F. Max Miiller, in the Introduction to the Dhammapa&a, p. xxxvi, by entirely different reasoning arrived at the year 259 B.C. Professor Rhys Davids

1 He must have been a youth at the time; his brothers died in 195 and 196 A.D. The manuscript * Piradhana Nuwarawal * states Ifcat Gaja-Bahu was only 16 when he became king. I suggest that Mahallaka N§ga was probably a son of Wasabha.456 ANCIENT CEYLON

states in his'Buddhist Suttas, p. xlvi, that it must be within a year or two of 267 B.C. I adopt 263 B.C. as a mean date; * the accession of Devanam-piya Tissa would thus take place in 245 B.C., and in any case not more than ten years earlier than that date.

In the Mahavansa, Tissa's father, Muta-Siva, is stated to have reigned 60 years, and the latter's father, Panqlukabhaya, 70 years. The same work also records that Pandukabhaya was born in the year in which his grandfather, Pancjuwasa-Deva, died. Pan Juwasa-Deva was succeeded by his eldest son Abhaya, who after reigning 20 years was deposed by his brother Tissa.

Before Abhaya's deposition his nephew Pancjukabhaya; had taken the field against him, and had also married his niece Suvanna-PalL It may therefore be assumed that Pancjuka-bhaya's son, Muta-Siva, was born about the time when Abhaya was deposed, which was 17 years before Pancjukabhaya succeeded in acquiring the sovereignty. According to Sinhalese chronology, the age attained by Muta-Siva would thus be the length of his own and his father's reigns, plus this 17 years, or a total life of 147 years. Even if we allow him an extremely long life we cannot accept more than 90 years as his age when he died; but as a more probable lifetime I take 80 years. This would fix the deposition of Abhaya at 325 B.C. as the earliest reasonable date, if the accession of Devanam-piya Tissa took place in 263 B.C.

The lengths of the reigns of the two preceding kings, Wijaya and PancJuwasa-Deva, given in the table, are those of the Mahavansa. As it is unlikely that both Muta-Siva and his father lived exactly 80 years, I have apportioned 70 years to the latter, whose life would then terminate in 275 B.C. This does not affect the total length of their two reigns.

What I wish to emphasise is that if the lengths of the reigns of Wijaya, Pan4uwasa-Deva, and Abhaya are correctly given by the annalists, and no kings are omitted by them, Wijaya

1 In his Early History of India, p. 145, Mr. V. A. Smith makes the date 272 or 273 B.C.THE EARLIEST INSCRIPTIONS 457

cannot possibly have become king more than a few years prior to the date, 414 B.C., given by me as the earliest reasonable one for his accession.

Although it has been suggested that the names of some Sinhalese kings may have been dropped by the historians, it appears to me clear that all probabilities are strongly against the omission of the names of any other early sovereigns. By omitting them the chroniclers would be merely intensifying the difficulty which they experienced in stretching back their chronology so as to make it extend to 543 B.C., the assumed date of the death of Buddha. Being left without other kings to fill up the gap, they were obliged to double the lengths of the reigns between 205 and 245 B.C., and also those of Panduka-bhaya and Muta-Siva, thus making these two last stretch to a ridiculous and impossible extent. As the existence of other kings would have relieved them from this necessity of falsifying the chronology they would be most unlikely to omit their reigns.

It is much more probable that fictitious names would be inserted in order to span the gap up to 543 B.C. than that the names of actual rulers of Ceylon would be struck out of the list.

If there is any additional error, therefore, it must be looked for in the lengths of the first three reigns. But it is evident that in any case these cannot be lengthened more than a very few years. The historians allow a reign of 38 years to Wijaya, 30 to PancJuwasa-Deva, and 20 to Abhaya, who however was alive for more than 17 years later, since it is recorded that his nephew Pandukabhaya appointed him after that period City Conservator of Anuradhapura. As Panduwasa-Deva was married immediately after he came to the throne, we may assume the age of Abhaya, his eldest son, to have been 66 1 when he was appointed to this office. There'is nothing to show that he died immediately afterwards, and he may have survived for several years. Thus it is clear that no addition can be made to the length of reign allotted to him by the

1 Made up by* 29 years of his father's reign, 20 of his own, and 17 years of his life after his. deposition.458 ANCIENT CEYLON

historians/since a service of only four years as Conservator would bring his age to 70 years.

We are therefore left with only Wijaya and Panduwasa-Beva as the sole kings whose reigns might have lasted a little longer than the time stated in the histories. If it be permissible to assume that both were not more than 25 when they became kings, Wijaya's age would become 63 at his death, and Panduwasa-Deva's 55. Even if we extend both up to 70 years it would carry the beginning of Wijaya's reign only 22 years further back. But the probabilities are overwhelmingly against such an addition to their ages. It would show, as a result that in the case of five legitimate consecutive rulers (omitting Tissa, the brother of Abhaya, as a usurper), not one died under the age of 70 years. Such a chain of long-lived monarchs is unheard of, and is manifestly inadmissible.

I am not concerned in attempting to reconcile the date of Gotama Buddha with that of Wijaya, who is stated to have arrived in Ceylon and become the first Sinhalese king in the year when Buddha died.1 According to the genealogical table of Buddha's relatives they appear to have been contemporaries, as the queen of Panduwasa-Deva, the nephew of Wijaya, was the daughter of Buddha's cousin, if the Sinhalese histories are correct. Any error in the chronology is likely to be found among this queen's ancestors ; it is possible that two or three names have been omitted between her and Amitodana, the unde of Buddha. Such an omission would account for the discrepancy in the dates of Buddha and Wijaya, without its being necessary to assume that the list of Sinhalese kings is at fault* I have shown; this in the table, therefore.

1 Mr. V. A. Smith, in hfe Early History of India, 1908, p. 42, states that Dr. Fleet now considers 482 B.C. the most probable and satisfactory date of the death of Buddha.TABLE OF THE EARLY KINGS OF CEYLON.

In italics are tlu-,i« in inscriptions. The dates are those of the respective reigns.

K. of N. K* of ?tti. son ? iSuppa-Dfivl «? Anura Jaya-Sgna. K. of Kapila (Sakya) Devadaha. K. of Koli (Sakya) ..._..... I ; I

Siha-hanu ? Kacchana Yasodharl ? Aujana (Sakya) Kacchana ? (m. Siha-hanu)

r ~T i i 3 sous Pamit.t Amita Amitodana Bi» Suppabuddha) Pancjlu (Sakya) ? K. of M.uM.1 crs } - - da. SuddhOdana -Maya -Mabja Pajapati Dandapani^Suppa-Buddha -Amita

Nanda Rupa-Nanda

Susima Siddhattha =-Yasodhara r\~,rn T\~\.4-~

I I V.? K. of Ma- K, cif I^tii X^ *? ,. i ? *auclle*' JL eva-JDatta (BUBDHA, died if bcmt 480 B.C.) 1 ; Rahula (monk) Probably two j or three names omitted. ! i 6 sons came to Ceylon

1 3° °rt ? I 1 (Vaedi

r^~ i i ~ i r ~~ Pitnijuwris i 1 eva *«Bhadda~ Kacchana Ga 375-345 I (youngest da,) (K. in mini Rama Uruwela Anuradha Wiiita Dirfiayu R5hana India)

? 345-3*5 I I ~ ~1 Fissa 7 sons Girikan da-Siva, Ummada-Citta Df^hi Gamini

Suvarma-Pali =^^^^^^^^_^^= Pandukabhaya

| ! 308-275

Muta-SIva or Mutatissa ' 275-245

Thsa?Anuld X? MaUa-Naga ? Anula

Aht

^^ttabhaya Maha-Siva Mitta Sura-Tissa or tJddhaculabhaya orMahtsena 215-210

(monk) 220-215 killed by Sena

and Guttika

-UUiya (Tamils)

m* 210-205

Asela Kira Anula (nun) Sivali (nun)

205 (m. Mahanaga)

.1 by Elara 205-161

\ Maha Arittha (monk)

^ _-_ _-. . da.

A ti?tA*f£

/

«-Wi lYtt* | or Sanmddanava (nun)

(nun)

». Puttha-Glmloi SaddhS-Tissa Manila (nun) Sai:

Gamini

161-137 *

(nun)

Soma-DevI«=Abh.aya K. of Giri.

j - i '?itn !----------

Thulkthanaka Khtptta-Naga == Anula === \AyaAsSli Lajaka-tisa* 119 JNaga

119-109 109-104

L*»S

1'

I 1

Watta-Gamini =-Somiir-DevI Gamani-Abaya, and Abaya 104, ' 88-76

Ramadarayd (nun) «JWahacula Mahatissa ?AnnJS ?Maha-Naga Abl Anuridiyd

\Gamini-Tisa 47-42 or Cora-naga

Kalakam^Tissa

$ya

or Tisa Ay a 76-^62

62-50

Ay a Aya

Knda-Tissa

Ti'sa Ay a

So-47

B.C.

30 B c.-9 A.D,

Mahadathika Maha-Naga 9-21 A.D. Maha-Naka I

A.D.

Mahalim Aya |\maiida-Gammi 21-30

Kanijanu-Tissa 30-33 or Kaniraja Sinkana

Ila-Naga 38-44-Maha-Matta Alunaka

Cumbhaya, 33^35

Mahallaka«Naga 135^141 I Naka or Naga

SIvaK, 35

Damila-dev! «=Candamukha-S!va 44-52 ?

Datta Subha 60-66

Metta?Son da. ? Yasala'laka-Tissa

I I 52-60 Tisa

T j

Metta - Wasabha66-uo

I Wahaba

Mahamatta«-Wankanasika-Ti33a 110-113

Bhatika-Ti^a 141-165 Batiya-Ttsa Butiyn Kanittha^Tiasi 165-193 AfaJw-Tisa - | da, ««Gajabahuka-*Ganiini 1^ 1 1 3-1 3 5 Gamini A bha Gayabahii Gamini Abaya

Tiyagasala Cula-Naga 193-195 Khudda-Naga i Nafta

X

?]

Sahala kala Ahapiya

iri-Na

Siri-Naga 196-215

To face p. 458XII THE EARLIEST COINS

THERE is nothing to indicate the date when the first coinage was introduced into Ceylon from India; all that can be said regarding it is that coins were in the country in the second half of the third century B.C. I myself saw two silver Puranas or Salakas, nearly square but rather thick coins without any punch-marks, resembling the copper coin numbered 20 on Plate I of Sir A. Cunningham's Coins of Ancient India, which were found in 1884 with the four relic-receptacles that had evidently been deposited in the relic-chamber of the Yatthala dagaba, built by King Maha-Naga or his son in the third century B.C. at Tissa or Magama, the ancient capital of southern Ceylon. I have already described the relic-cases in the chapter which deals with the ancient dagabas.

The Buddhist monk who was in charge of the largest dagaba at Tissa, which was undoubtedly built by Maha-Naga, informed me in 1884 ^at some similar coins made of copper, with small punch-marks on their surface, the shapes of which he could not describe, were also found in the debris thrown round its base by its despoilers. They were all replaced in the relic-chamber when it was closed during the restoration of the structure, but the description that was given of them leaves no doubt as to their presence at that work also.

The histories of Ceylon contain no statement that invaders held the southern part of the island before the eleventh century, in the early half of which it is recorded in the Mahavansa (ii, p. 90) that the forces of the King of Sola occupied that part of the country and despoiled many wiharas. Even if the relic-charnbers of these two dSgabas had been broken into at that time (of which, however, there is no record) it is improbable that any Sinhalese king who restored them460 ANCIENT CEYLON

afterwards would think it necessary to place such early coins in the new relic-rooms. The presence of the coins therefore is very strong evidence that it was only at the spoliation in the time of the Pandian king Magha (1215-1236 A.D.) that the relic-chambers were rifled. The finding of the car-nelian gem belonging to the royal seal, to which allusion has been made in a former chapter, is a further proof that this was the case. At the YatthSJa dagaba, which I often saw before its restoration, I could not observe the slightest evidence of any restoration or rebuilding of the dome, though I looked carefully for it; only one size of bricks was used in it, "and those in the dome were all unbroken and evidently undisturbed ones. It may be concluded, therefore, that the relic-chambers remained intact until the thirteenth century ; and that the Puranas were placed in these dagabas in the third century B.C., and were lost or thrown away by the persons who broke into the structures in the time of the Tamil king Magha,

In 1885, several of these coins were discovered at Mulleittivu under circumstances that gave them a special interest. A man who was in charge of a small coconut garden on the north side of the town, where the soil all around is full of fragments of a rough type of pottery, as the result of an unusual fit of energy determined to level a mound of sandy material, and to utilise the soil for filling up some hollows near it. When he reached the level of the adjoining ground he was surprised to meet with the top of a large ring of coarse earthenware such as is used in Ceylon by some of the Kandian Sinhalese for lining wells at their houses. On clearing the sandy soil out of the inside of this ring he found others below it, and discovered that he had unearthed an ancient shallow well at the bottom of which there was fresh water. The rings were 3 feet in internal diameter; the top one was 6| inches deep and the others about 8 inches. At the present day such well-linings are from 2 feet to 2| feet in diameter, and about an inch thick.

At a short distance above the water-level, and embedded in the sand, he obtained a number of silver Puranas, andTHE EARLIEST COINS 461

some thin oblong copper plaques which proved to be an entirely new type of money1 described by me in the previous year from specimens obtained at Tissa. The total number of Puranas was 51, and of the plaques 16. The late Mr. R. Massie, the Assistant Government Agent of the district, obtained nine of the plaques, out of which he presented four to the Colombo-Museum, two to me, and at a later date two to the British Museum. I got the other seven when I visited the place shortly afterwards.

It would appear that the original owner of this money, possibly fearing the result of some disturbance or war, first threw some sand into his well and placed his small stock of cash on it; he then filled up the well and to mark the spot raised over it the mound to which its modern discovery was due, little expecting that more than two thousand years would elapse before it would be disinterred.

In addition to these coins four other oblong plaques were found by me at Tissa in 1883, in excavating channels from two sluices at the Tissa tank, and I obtained two halves of others from a neighbouring village. The position of these coins enables us to fix the date of the earliest type of this money as yet discovered.

At the end of the embankment of the Tissa tank, on the high side of a hollow or small water-course, there had been a village of potters and other artizans who were accustomed to throw the ashes and rubbish from their houses and furnaces into the hollow, which thus became,a kind of * Kjokken-modding/ Afterwards, soil carried down by rains covered up this deposit, and eventually filled all the hollow to the depth of eighteen feet at the deepest part. By a lucky accident, a channel from a new sluice was cut by me through this very site, and numerous articles belonging to the ancient workpeople were met with, including thousands of fragments of pottery, some few of which were inscribed with letters of the

1 Doubts have been expressed as to whether the plaques were coins or votive' offerings, but I was led to understand that the authorities of the British Museum do not share them. I have shown below that all the early Indian and Ceylon coins were amulets as well 'as money. ' ' ' ? . *462 ANCIENT CEYLON

earliest angular type which is certainly of pre-Christian date. In two instances there are words on the upturned sides of rice-plates, which appear to be the names of the persons for whom they were made. One was inscribed Gapati Sivasa, * the householder Siva's *; the last letter is near the broken edge of the fragment of "earthenware, and possibly the name of this person's son followed it. The other, which is also

incomplete, is ?-----ke Dayapusaha Aba, f Abhaya, (son) of------

Dayapusa'; the missing word may have been gamike, ' the villager/ or bojike, 'the headman/ On all other fragments only one or two letters were found.

From the primitive forms of the letters, which do not include a single round s, or the rounded vowels or lengthened k or r, which stamp the date on post-Christian writing, it appears certain that the letters on the pottery in the upper part of this stratum, which were all written on the earthenware before it was baked, were inscribed at the latest three-quarters of a century before the Christian era, while those in the lowest part most probably date from the second and third centuries B.C., when the construction of the large dagabas and other important monastic edifices must have necessitated the presence of a large force of workmen. On many of the bricks laid In the Yatthala and Mahanaga dagabas similar letters were written or stamped before they were burnt.

The lowest stratuirf of remains was four feet thick in its lowest part, which was eighteen feet below the surface of the ground. In the very bottom of this layer one of the oblong coins (No. 1) was unearthed in 1883 in my presence, and it must, I believe, belong to the third century B.C. A second (No. 2) was taken out of another part of the same stratum, and therefore probably belongs to either the second or third century B.C. A third (No. 3) was found slightly above the pottery layer, and may be a century or perhaps two later than these. A fourth (No. 4) was met with at the remains of some early dwellings that were cut through in opening a channel from another sluice at this tank. The fragments of pottery which were found there were similar to those at the former cutting, and some bricks were of the same dimensionsTHE EARLIEST COINS 463

as those used in the great dagaba of Maha-Naga and appeared to have been made in the very same moulds. This coin, therefore, may also possibly date from the first, second, or third century B.C.

At the excavations made subsequently at Anuradhapura, twelve specimens of the oblong coins were found by Mr. S. M. Burrows, late of the Ceylon Civil Service, behind the northern wahalkada at the Abhayagiri dagaba,1 and others were discovered by Mr. Bell, near the Jetawanarama,2 but so far as I am aware not under circumstances which afford a clue to their age. It may be assumed that those having the simpler designs on them are of pre-Christian manufacture, while others are of later date, and one at least is proved by the Aum monogram on it to belong to about the third or fourth century A.D. I am indebted to the courtesy of Mr. Bell for permission to include descriptions of these coins with the rest.

THE PURANAS, OR DHARANAS, OR SALAKAS

Many of these coins have been discovered in India, and Sir Alexander Cunningham estimated that he had seen between four thousand and five thousand specimens.3 They have been figured and described on several occasions, but I think that no account has been given of any from Ceylon, except nine much worn examples found by Mr. Bell among the debris at an early monastic site in Anuradhapura, which was surrounded by a ' Buddhist Railing/ These were described and figured by him in i8g2.4

All the Puranas found in the island have been imported from India. The punch-marks on them, each impressed by a separate small punch, and almost all near the sides of the coins, are, with perhaps two or three exceptions, identical with those on Indian coins; and silver and copper, the materials of which they seem to be composed, are not products of Ceylon.

1 levers, Manual of the North-Central Province, p. 234.

2 Arch. Survey of Ceylon. Fourth Progress Report, p. 13.

3 Corns of Ancient India, p. 42.

4 Arch. Survey of Ceylon. Fourth Progress Report, pp. 4 and 12.464 ANCIENT CEYLON

These coins are thus a proof of the early trade with India. The majority are so much worn that any symbols that may have been impressed on them have almost disappeared.

The common designs that can be recognised on the obverse of this money are the rayed sun-symbol, a circle with six emblems round it, the dog, the elephant, the bull, fishes (some of,them in tanks), the turtle, forms of trees, and a three-arched structure, surmounted in one case by a crescent. All the mammals face towards the right. The usual emblems that are absent from the coins which I have seen are the human figure, the bow and arrow, the caduceus, the Swastika, and birds.

On the reverse side some coins have several symbols which are generally nearly worn away, but as a rule there are few marks on that face, among which are the rayed sun, the tree, and the structure with three arches. In one case a person has engraved a design which perhaps was intended for the early cup-shaped letter m, with a cross-bar in the middle, as it occurs in local inscriptions of earlier date than 100 B.C. Another has a punched symbol which resembles an early letter, but may be part of an animal design.

The shapes vary as in India, about one-third of the coins being round in outline, while the others are more or less rectangular, and occasionally have one corner or two adjoining ones cut away. According to Indian authorities this indicates that when first cut off the strip of beaten silver hardened by an alloy of copper, from which it was taken, the coin was found to be too heavy and was therefore reduced in weight in this manner. It is obvious that it might still vary from the correct weight to the extent of some grains.

The full weight of such coins as these has been shown by Sir A. Cunningham to be about 57*6 grains.1 If this was the original weight of those found at Mulleittivu all must have been subjected to wear for an extended period, since the average of thirty-three2 is only 33*8 grains, and runs from 28 to

1 Is it more than a coincidence that an early silver coin found in Crete weighs 56*4 grains ?

2 The rest, which were extremely worn and unfortunately were unweighedy were * acquired * by an inquisitive servant.THE PURANAS 465

39 grains; the heaviest weigh 38^ and 39 grains, and the lighter ones from 28 to 30 grains. That these very low figures are due chiefly to wear and not merely to original short weight appears to be confirmed by the well-rounded corners of all the rectangular coins, and the indistinct or fragmentary state of the punch marks on all but four or five.

There are no sharp angles like those in the examples illustrated in the Coins of Ancient India. Some from which one or two corners have been cut are now among the lightest; it is to be presumed that these have lost half their original weight while in circulation, as they are too much worn to be coins that were at first of half the full weight. In Mr. V. A. Smith's Catalogue of the Coins in the Indian Museum, Calcutta, the weights of 108 selected out of about 300 range from 35-7 grains to 55-6 grains.

The late Sir Alexander Cunningham, who was for many years the Director-General of the Archaeological Survey of India, and was the greatest authority on these coins, stated the mean weight of 800 to be 47 grains,1 and the average loss to have been one and a half grains per century of their age. He characterised a loss of 19 grains in a presumed age of 600 years as ' very exceptional' (C. A. /., p. 55), and he referred to a coin weighing 34 grains as an example of long wear.

If the loss of weight of the lighter coins found at Mulleittivu has been at this average Indian rate it would remove the date of issue of several of them to about 2000 B.C. At the * very exceptional * rate the date would still be carried back to the eleventh century B.C. ; and in the case of even the heavier ones it would extend to the eighth century.

I express no opinion on these ages; I merely point out the times to which Sir A. Cunningham's data would remove their origin if my belief regarding the date of their deposition is correct, leaving those with a knowledge of the subject to draw their own conclusions. Sir A. Cunningham thought that Puranas were issued by 1000 B.C., but Mr. V. A, Smith, in the Introduction to his Catalogue (p. 135), remarks that this estimate ' almost certainly is much in excess of the truth/ * ?. A. /., p. 55.

HH466 r ANCIENT CEYLON

As he states that it is well established that the full normal weight was 'about 58 grains/ it is to be presumed that he considers the rate of their erosion which was accepted by that distinguished archaeologist to be too low, although he states (p. i) that ' Sir Alexander's unique experience extending over considerably more than half a century enabled him to accumulate a mass of knowledge, both general and special, concerning all classes of Indian coins, which nobody can hope to rival/

The date when the Puranas were buried at Mulleittivu is approximately fixed, as will be shown below, by the type of the oblong coins found with them as probably in the first half of the second century B.C., and it is extremely unlikely to be later than the first years of the first century B.C.

Since the amount of the loss of weight of these Puranas must be accepted as proof that they were in circulation for a period amounting to at least several centuries, it is apparent, if that date be correct, that they all belong to a time prior to the introduction of Greek coinage into India.

They are of two general types, with some intermediate gradations, a larger thin coin of which both nearly square and rounded specimens occur as in India, and a much sirikHer but thicker coin which is usually oblong in shape, although both rounded and square examples occur. The larger coins have numerous punch marks on them, in several" cases bn both sides; the smaller specimens have few marks and those almost worn away. On some faces no marks are visible. The general appearance of the small coins, the surface of which is of a rougher texture than that of the others, and the extremely worn condition of their marks, a few of which can only be faintly seen in a strong light, lead me to suppose that they are considerably older than the larger coins ; but there is not much difference in the weights of the two varieties, since although several of the lightest coins are of the smaller type other small ones are as heavy as many of the larger variety. It is important to note, however, that in the smaller coins the surface exposed to wear is little more than half that of the larger ones, 'and therefore an equal loss of weight in their case must indicate a far greater age.THE PURANAS 467

The different shapes are all visible in the illustration in C. A. I., from a photograph, of the early carving (B.C. 250) found at Gaya, which represents the purchase of the garden for the celebrated Jetavana monastery at Sravasti. The small coins are few in number in that relief. In the later one at Bharhut (150 B.C.) only the larger coins, both rectangular and round, appear to be shown.

Copper, which was used in the coins, is found sparingly in Northern India, but I think is not mentioned in the Vedas. Silver certainly was known in India at that early period, the moon being described as ' Silvery ' (R.V., ix, 79, 9).

Although wealth is everywhere defined in the Vedic prayers as consisting of cattle, horses, and gold, it is to be remembered that the authors of the hymns were priests who always demanded the most valuable things, and that even if there was a gold coinage of some kind there would be few gold coins to pray for. As Mr. Del Mar remarks in his History of Money, they could only be required as multiples of a coinage of lower values. The omission to mention gold coinage uj the hymns is therefore not a definite proof of its non-existence.

Mr. Del Mar has pointed out that the state of society and civilisation in India in the [later] Vedic age was one that apparently necessitated the use of some kind of money ; and if the reference to a gift of the value of a thousand or ten thousand pieces in the Sama Veda (Prapathaka, iii, 10, 9) is correctly translated by Stevenson it is clear that coins were numerous in the second miUennium before Christ. The extract is as follows:?'O Wielder of the Thunderbolt, thou art not impoverished by a noble and surpassingly splendid gift, not by one of a thousand pieces' value, no, nor by one of ten thousand, not even, Possessor of Wealth, by such a gift a hundred times repeated/

In the Rig Veda (Griffiths) viii, i, 5, the words are, ' O Caster of the Stone, I would not sell thee for a mighty price, not for a thousand, Thunderer! not for ten thousand, nor a hundred, Lord of Countless Wealth/

As in early times there was usually only one coin which468 . ANCIENT CEYLON

was found in great numbers in a country, it was a common practice to omit specifying any special coin, or even money, when mentioning large sums; only the number of the coins was given. Thus the pre-Christian annalists from whom the author of the first part of the Mahavansa borrowed his historical facts stated that at the building of the so-called 'Brazen Palace' at Anuradhapura King Duttha-Gamini, in order to provide for the wages of the workpeople, deposited 'eight hundred thousand' at each of the four entrances. It is also recorded that in offerings at a festival at the Bo-tree he expended ' one hundred thousand/ and that he rewarded the architect of the Ruwanwaeli dagaba, for his lucid explanation of his design, with a suit of clothes?a Robe of Honour? worth 'a thousand'; and other similar examples might be quoted.

Of such statements there are several instances in the Rig Veda. In Book x, 17, 9, the line occurs, 'Give food and wealth to present sacrificers, a portion, worth a thousand, of refreshment/ In x, 102, 2 we find ' Loose in the wind the woman's robe was streaming what time she won a car load worth a thousand'; and in verse 9 of the same hymn, * Therewith hath Mudgala in ordered contest won for cattle for himself a hundred thousand/

I have not searched for earlier example^. Those which I have quoted appear to be quite as unmistakable references to money as the instances from the Mahavansa. Since gold and silver money, which must have been preceded by a currency of lower value, is mentioned in the early part of the Ramayana as being well known (Book i, 13; ii, 32), I accept these references in the Rig Veda as dear proofs of the existence of some form of money that was in extensive use in later Vedic times.

We cannot expect ever to see many examples dating from, such a fat-distant period. Although, thanks to the early annalists, it is certain that numberless coins, which in some instances were stated to be termed Kahapana, were in use in Ceylon in early times, not one specimen of them had been seen twenty-seven years ago; and even now few have beeniyv, t g ^

^/w i* f

\ Jv r r

'^.4f C__yi t_J:

FIG. 154. Mulleittivu and Tissa Coins.

To jace p. 469THE PURANAS 469

discovered, notwithstanding Mr. Bell's excavations throughout the ruins of the ancient capital, Amiradhapura. When that is the case with coins of such comparatively recent date, little surprise should be felt that the money of Vedic times has remained so long unknown. Its absence is not a proof that such a coinage did not exist.

I mention this because it appears to be probable that the oldest examples of the Mulleittlvu money, and also some of the Puranas found by Mr. Bell, may date from an extremely early time, and though later than the Vedic age may have preserved the type of a coinage which may have been current in that period, or shortly afterwards.

I give illustrations of the best specimens found at MuUeittivu, together with typical examples of the smaller variety (Fig. No. 154). Some of the symbols on the former coins are clearly defined; it must be presumed that these were impressed long after the money was issued, the reduced weights of the coins on which they occur plainly showing that they have been in circulation for a period long enough to have nearly or totally worn them away had they been stamped soon after the coins were made.

Beginning at the top and proceeding down the left side in the direction taken by the quadrupeds, the emblems on the coins which are illustrated are as follows :?

(a.) Rectangular, with two cottiers cut off; -72 in. by -68 in.; weight 37 grains.

Obverse. Standing Humped Bull, wearing a collar, which is indicated by a projection on the nape and throat; Wheel or Sun symbol; Dog; Symbol composed of two concentric circles?that is, a disk with a circular band round it?from the outer circumference of which project six emblems (parts of three only can be seen); Tree.

Reverse. A straight leafy Branch in centre, in a very narrow ellipse; and remains of other symbols, among them apparently a Fish, a Structure of five or six arches, a * Taurine ' emblem, and possibly another form of Tree.

(b.) Rectangular, with comers rounded by wear; -66 in, by -61 in.; weight 39 grains.470 ANCIENT CEYLON

O. Dog with raised tail and forequarters lowered as though about to spring forward, standing on two arches, the tops of which are visible; Tusk Elephant; Circle with emblems ; two Arches of a structure of which probably a third one has been destroyed by being over-stamped with the Elephant; Humped Bull with collar; Sun emblem.

R. Broad plain cross in a circular punch-mark; Tree (inverted) and remains of other symbols among which may be the three-arched Structure.

(c.) Rectangular, with corners rounded by wear; -68 in. by -62 in; weight 38 grains.

O. Bull; Circle with six emblems ; a form of Tree punched over one of the last emblems; three arches of a Structure which probably had five, three in the lower row and two above them; these last are separated by a central space over which stands a Dog, with its hind feet on one arch and forefeet on the other; Sun emblem, the inner disk of which is connected with the outer ring by tiny spokes.

R. (not illustrated). Upright Axe with handle, or part of a Steel-yard, the whole punch-mark being a half ellipse; three Beads attached to the sides of a sector-shaped punch-mark; Tree, and fragment of a symbol.

(d.) Irregular oval in shape; -87in. by 72 in,; weight 30 grains.

O. Dog ; an uncertain symbol; Sun emblem ; uncertain Quadruped behind with thick legs; a figure from the five angles of which rise twigs with three leaves which form crosses, a flower or fruit on a short stem being in each intermediate space (only half the figure is on this coin); concentric Circles with six emblems. Quadruped wearing a broad collar and having two cuts above the tail.

R. Symbol which is possibly a Yak-tail Fly-whisk; Sun emblem; Fish; above these the Structure with three arches, surmounted by a crescent; above this a fragmentary symbol; part of uncertain symbol; a long punch-mark in which are a ' Taurine' symbol and! two concentric circles without surrounding emblems; two uncertain symbols; an oblong punch-mark, inside which is a transverse bar near each end and a minute emblem between these.THE PURANAS 471

(e.) Irregular oblong ; 1-02 in. by -44 in.; weight 37^ grains.

0. Symbol resembling three or four leaves (not beads) projecting from the sides of a hollow, towards its centre ; traces of a symbol above it; a long elliptical punch-mark, the symbol in which is partly destroyed by the next one; the upper part of it may be a form of Trisiila and when the coin is turned round the other part resembles a Bull; Sun emblem; large ?' Taurine' emblem; remains of another symbol above it.

R. Circles with surrounding emblems ; Sun emblem, about 16 rays.

(f) Rounded; 76 in. by -63 in; weight 37 grains.

0. Structure of three arches resting on vertical walls, the central arch rising half as high again as the others. Over it a snake formed by a deeply-waved line; it resembles the conventional clouds of temple artists, but possibly was not impressed by the same punch as the arched structure; worn symbol, apparently an animal and possibly a lion ; Sun emblem, 16 rays ; a design like a Palmira tree, but apparently a Flower with a circular centre and five petals, on a stem, perhaps stamped over the head of an Elephant.

R. Symbol like an upright Axe with short handle, punched over part of another emblem with a straight bar beneath it and possibly an arched Structure ; traces of two other emblems.

(g.) Nearly circular; -58 in. by -58 in.; weight 34^- grains.

O. Dog (?) ; Circle surrounded by emblems ; Fish, partly over-stamped by another design; Sun emblem, and others.

R. Traces of emblems.

(h.) Oblong, one corner cut off, -55 in. by -34 in.; weight 295- grains.

O. Rude punch-mark, shaped like a Quadruped but turned to left.

R. Blank.

(i.) Oblong, one corner cut off; -56 in. by -45 in; weight 34f grains.

O. Punch-mark, resembling a letter, but possibly part of a symbol R. Traces of symbols/47a ANCIENT CEYLON

(k.) Oblong, two corners cut off; -56 in. by -44 in ; weight 29 grains.

O. Trace of part of concentric Circles with surrounding emblems ; trace of Sun emblem.

R. Blank.

Three other small coins are as follows :?

(1.) oblong; -57 in. by -44 in.; weight 28^ grains.

0. Fragments of Circle with surrounding emblems, and traces of three other symbols. R. Faint traces of two symbols.

(m.) short oblong, one corner cut off; -57 in. by -50 in.; weight 29^ grains.

Only indistinct crescentic symbol with straight back.

(n.) oblong; -55 in. by -44 in.; weight 30^ grains.

Traces of two symbols on one side, and of one on the other.

The average weight of thirteen coins of the smaller variety is 32*9 grains, the heaviest weighing 38^ grains.

(o.) The lightest coin weighs 28 grains, and is a little larger than these. It measures -59 in. by -56 in., and is thus almost square, with rounded comers; one corner is cut off.

There are four or five very faint punch-marks on O. One is a Structure of three arches with a channel below it; another appears to be the fore-quarters of an Elephant; between these is part of a Sun-emblem ; there are also faint traces like part of the Circle with six emblems, and of two other symbols. On R. all that can be seen is a worn punch-mark of a narrow Leaf-shape, and perhaps the trace of another symbol.

In the case of the nine Puranas discovered at Anuradhapura by Mr. Bell, in 1891,l the antiquity of the building at which they were disinterred is proved by the size of the bricks employed in the lower part of the walls. These measured 18 inches by 9 inches by 3 inches, Bt. being thus 27 square inches and the contents 486 cubic inches. My * Amended List * of bricks shows that these dimensions probably belong to the latter part of the second century or the early part of the first century B.C. Unfortunately there is nothing to indicate the period when the coins were deposited there; judging by their extremely worn state it may have occurred at some date con-

1 Arch. Survey of Ceylon. Fourth Progress Report, pp. 4 and 13.":?"- THE PURANAS 473

?siderably after Christ, as other money, both South Indian (Kurumbar and Pallava) and Roman (Theodosius), was also found at the site.

Three of the Puranas were apparently spurious imitations of silver coins, being made of copper and having still traces of the silver coating which had been applied to them. Two of these were square, with half inch sides and weighed 24 and 29 grains. The third was an oblong, -7 in. by -37 in., weighing 40 grains. On its obverse was a standing figure unlike those of the true Puranas, and perhaps copied from an oblong coin like those described below, with a length amounting to three-quarters of the coin ; on the reverse appeared a chequer pattern.

The other six were silver coins, three being more or less oblong, and three nearly square. O£ the former variety, one coin measured ^56 in. by -44 in., and weighed only 16 grains. It had the Sun emblem, and the remains of another. Each of the others weighed 19 grains, and was a broad oblong, with one corner cut off. Both had traces of symbols.

Of the square coins, two had sides of half an inch; one with ;a corner cut off weighed 30 grains, and the other 42 grains ; these also showed traces of symbols. The third one, of an irregular shape, had a length of -75 in., and weighed only .17 grains. On its obverse was a Sun symbol, a Tree, and other worn punch-marks, one of them appearing in the figure to be the three beads projecting into a hollow. It may represent pearls in the shell, a powerful amulet in Vedic times (Ath. Veda, iv, 10).

It is noticeable that on nearly all the Puranas of apparently the earliest date traces are visible of some of the same punch-marks as those on the coins of the latest type. Among these the wheel-like Sun symbol was the emblem most generally ?employed.

The symbol consisting of two concentric circles from which ?project six emblems is common on Indian coins. In the Mulleit-tivu coins the emblems usually are a * Taurine * or bull's-'.head symbol enclosed in a semi-ellipse* and another which .has been termed a Chatta or umbrella, but which may be a iorm of Axe-head with a narrow stem and a semi-circular474 ANCIENT CEYLON

cutting edge. These occur alternately, being three times repeated. On another Mulleittivu coin a Fish is placed between them, the three emblems being twice repeated.

The meanings of the symbols will be considered after the other coins have been described.

THE OBLONG COPPER COINS

THE TISSA COINS

These coins are all thin copper oblongs of one type, cut off strips of beaten copper which were themselves cut off a larger sheet. The designs on both faces were impressed simultaneously by means of two dies that were nearly as large as the coins. On the obverse, which has a flat border, a person is represented, standing facing front, and holding an upright object, apparently a flower stem, at each side of the coin. On the reverse the Swastika appears, exactly resembling those cut at some pre-Christian inscriptions, being raised high on a central pole which rests on a transverse base line from which rise two short upright bars at each side of the central post. This symbol is found on all the oblong copper coins of the island, and also on a large circular copper coin which supplanted them. As a typical emblem of ancient Ceylon it is stamped on the cover of this work.

In the following descriptions of the coins the letter f. is * front/ and r. and 1. are 'right' and 'left" respectively.

1. 1-14 in. by -46 in.; weight 52! grains.

O. A figure of a deity, facing L Over and round the head runs a slightly waved line or circlet. The 1. hand appears to rest on something represented by three upright lines; the r. fore-arm apparently turned upwards. Legs slightly apart and feet turned outwards. There may be a tunic or cloth extending onto the thighs. The figure is well proportioned and even somewhat graceful.

R. Indistinct. Part of the Swastika symbol visible with its arms turned r. To I under arm of cross there are indistinct marks in relief. Colombo Museum.THE OBLONG COPPER COINS 475

2. 1-18 in. by -46 in. ; weight 44 grains.

O. Within a flat border, above the level of which the design does not rise,1 a full-length figure of deity, wider at the hips than shoulders, facing 1 A thin oval circlet round and over head, springing from the shoulders. Feet turned half outwards. The arms hang down, 1. hand appearing to grasp an upright bar at border, and r. holding a curved stem, at the top of which is a trumpet-shaped flower or a cornucopia with a circular flower over it, and leaves or buds above. At each side, near border, is a thin upright line. One anklet on each leg and a bangle above r. wrist.

R. In the upper third a Swastika, turned r. with stem prolonged downwards to middle of lowest third of coin, where it springs from a slightly curved horizontal line. Two short equidistant vertical lines of the length of the arms of the Swastika rise from the base line, on each side of central stem. Below base line and separated from it by a well-marked channel there is a parallel line, beneath which is a broad flat border line. Between the Swastika and basal uprights there are raised marks on each side of the central stem, but what they represent.is uncertain. Col. Mus.

3. 1-22 in. by -50 in. ; weight 41 grains.

0. Below a flat border line outside which is a sunk channel, a standing deity, facing f. Thick circlet from shoulders round head. Four anklets on each leg and perhaps bangle on r. wrist. Arms hang down, hands grasping on each side a curved line which may be a flower-stem, that on r. having a side view of a trumpet-mouth at level of shoulder, above which may be an open flower and buds, as in No. 2. Stem on 1. may also end in flowers and leaves. Below the feet a horizontal row of beads consisting of three thin upright ones in the centre and two larger round ones at each side.

R. The same Swastika, with thick lines, turned r.; base

1 This style of false relief is a distinct characteristic of Sinhalese art, whether of early or later date, and is also seen in Southern India.

Dr. Dresser, unaware of this, stated his belief that it was ' not practised by any other peoples than the Egyptians and the Japanese.* Japan,, its Architecture, Art, and Art Manufactures, p, 341.476 ANCIENT CEYLON

line nearly straight. Below it two slightly bent flat parallel T)ars separated from each other and the base by deep wide flat-bottomed channels. At sides of central pole are raised marks of uncertain character. Col. Mus.

4. 1-20 in. by -55 in.; weight 35 grains.

O. Broad flat border along top and upper part of r. side. Standing deity, facing 1, with irregular circlet over head and resting on shoulders; above the head a flat horizontal bar separated by a channel from the border. Feet turned half-'Outwards, and clearly and accurately represented, with heel, instep, and side view of toes. A large anklet on each leg. Arms hang down, and hold at each side near border a thin curved line which may be a flower stem, ending at shoulder level in indistinct flowers and leaves. Above them two beads on each side, near circlet.

R. Swastika as before, turned 1., on thick winding base, "below which and separated from it by a wide flat channel are two similar raised parallel bars, with narrow channel between them. On both sides of central pole are faint raised designs, that on 1. resembling an early letter Ke. Col. Mus.

5. About half only, from Sittrawila near Tissa, -54 in. by -49 in. wide.

0. Lower half of standing deity. An upright line to r. R. In opposite direction to O. Swastika, turned r.

H. Parker.

6. Over half, from Sittrawila, -66 in. by -50 in. wide.

O. Flat border at bottom. Lower half of standing deity with hanging arms ; 1. hand grasps an upright line.

R. Indistinct. Lower part of Swastika symbol, with raised marks in side spaces. H. P.

THE MuLLEirrlvu COINS

These coins are also oblongs of beaten copper, wider than those of Tissa ; they are all of one type which differs in some respects from that of the Tissa coins. Each has the standing deity on O., and the peculiar Swastika symbol on R.; but all are characterised by having in the space under the arm of the Swastika and usually on the right side of the central pole, aTHE MULLEITTIVU COINS 477

recumbent humped bull, sometimes kneeling on one knee, and on the opposite side a vase out of which grows a plant. The flat border of the Tissa coins is absent, as well as the line under the base of the Swastika, and sometimes the circlet over the head of the deity. These coins cannot have been long in circulation, the edges being sharp as though freshly cut; but on several of them the design is quite faint.

7. 1-16 in. by -64 in.; weight 49 grains.

O. Standing deity, facing 1, with legs slightly apart. Arms hang down, and hands grasp two upright lines at level of hips. That on 1. has a boss on it below shoulder level; opposite the neck it divides into two arms which are at first horizontal and then vertical, forming a bident. The other seems to be similar. From top of inner prongs of bident an arched band rises over the head. Beginning above the shoulders and extending round the head are seven clearly defined beads, and another is on each side close to the hips. Bangles on both wrists and an anklet on 1. leg ; the other ankle and both feet cannot be distinguished.

R. Opposed to O. Swastika symbol, turned r., on straight base. In space to r. of central pole and facing Swastika a recumbent humped bull. In space to 1. a round-bodied object, a vase with plant; one stem has three leaves. Mouth of vase faces Swastika. H. P.

8. 1*03 in. by -61 in.; w%ight -445- grains.

O. Narrow-waisted standing deity, facing 1, with legs well apart and arms hanging down. The 1. one grasps a strong upright pole, which at level of neck becomes a bident or trident, only one prong being recognisable. A bead above each shoulder and traces of two others above them. If a circlet of beads passed round head they may have been only five in number. Feet and r, hand indistinguishable.

R. Opposed to O. Broad-stemmed Swastika symbol, turned r., with pointed arms. Recumbent humped bull facing it on r.; a vase with indistinct plant on L, turned towards Swastika. H. R

9. 1-12 in. by -72 in.; weight yfe grains.

O. Standing deity, facing 1, with two large anklets on each478 ANCIENT GEYLON

leg. Arms hang down and hands grasp two upright poles near sides of coin, of which that on 1. becomes a wide bident above level of neck, 1. prong ending in a barbed point. A bead above each shoulder, and some kind of covering or headdress on head, extending beyond it laterally.

R. Opposed to 0, Raised Swastika, turned 1. Facing it on r., the recumbent humped bull. Vase on 1. with plant, facing Swastika. H. P.

10. 1-20 in. by 70 in.; weight 38\ grains.

0. Standing deity, facing i, very narrow-waisted, with bangle on 1. wrist. Side uprights barely distinguishable; that on r. appears to end in bident opposite neck. A circlet of beads extends from shoulders and round head, on each side of which three are visible.

R. Opposed to 0. Raised Swastika, turned L, with arms and basal uprights pointed. Indistinct symbols below it; that on r. may be the recumbent bull facing it. H. P.

11. 1-14 in. by -66 in,; weight 46J- grains. About one-tenth of the coin has been cut off r. upper corner; allowing for this, the full weight would be about 50J grains. Design is in higher relief.

0. Standing deity, facing f., with wider waist. Hands hold upright poles or stems near sides of coin. Arched line over head with lower edge scalloped. A bead above each shoulder, and another over it oil I; there may have been seven round the head. There are indications of fiowers on stems on each side of the legs.

R. Raised Swastika, turned r. Base straight; short side uprights resting on it are pointed at ends. In space to r. a kneeling humped bull, with r. fore-foot on ground. Tail is curled back on 1. quarter, its end being on the ground. The bull is tied to the Swastika by a halter. Two points on its neck may indicate two collars. On L a wide vase on a flat base; tree grows out of it, an upright stem, from wMch branch four side shoots, ending in leaves or flowers. H. P.

12. i-12 in, by -65 in.; weight 51 grains.

O. Standing deity, facing f., indistinct. Hands hold uprights near edge ; that on r. may end in bident. A beadTHE MULLEITTIVU COINS 479

on each side above shoulders, and anklet on r. leg. Transverse bar below feet bent downward in middle.

R. Opposed to O. Raised Swastika on strong stem, turned r. Recumbent humped bull on 1. facing Swastika, with hanging ears and dewlap. On r. a round-bodied vase out of which grow probably three stems, two only being visible, each ending in a leaf or flower. H. P.

18. 1-04 in. by 60 in.; weight 2gf grains.

O. Standing deity, facing f. Anklet on left leg ; lower part of other leg indistinguishable. Upright pole on r. appears above shoulder to curl round towards head in a rough circle, which may possibly be a large trumpet-mouthed flower.

R. Opposed to O. Raised Swastika, turned I. The only two basal uprights which are distinguishable end in points. To r., recumbent humped bull facing Swastika. To 1., vase with wide mouth facing Swastika ; r. shoot from it ends in a "broad pointed leaf. No transverse bar under base line of Swastika, this being the only coin which shows clearly that this special feature of the Tissa coins is omitted from those found at Mulleittivu. H. P.

14. 1-18 in. by -66 in.; weight 56 grains.

O. Standing deity in relief, facing f., feet turned half ?outwards ; holds upright pole at each side. That on r. becomes a bident at level of chin, with horizontal base and prongs sloping slightly outwards. A broad circlet passes from inner prongs over the head, and on the under side has a scalloped edge. There are seven beads round the head, the two lowest "being close to shoulders. On r. wrist two bangles, and anklet on r. leg. To 1. between leg and pole, a raised symbol, the -side view of a lotus leaf on an upright stem. To r. a thin curved stem ending in large flower. A fold of the dress hangs down between the legs. Three cuts across r. foot indicate a shoe or slipper.

R. Opposed to O. Raised Swastika, with thick arms turned r., two short thin bars cross the ends of its arms and of the basal uprights. To r. recumbent humped bull facing Swastika, to which it is tied by a halter. To I, a wide heart-shaped vase, pointed at bottom and having a high neck and480 ANCIENT CEYLON

lip, with upright plant ending in one leaf; it has a r. and L branch, former ending in a large leaf. H. P.

15. 1-04 in. by -66 in.; weight 51 grains.

0. Standing deity, facing 1, holding upright poles at sides ; bangles on wrists. A flattened circlet passes over head, from uprights. Traces of beads round head.

R. Opposed to 0. Raised Swastika, appearing to turn r, with pointed basal uprights. End of central arm on r., the-only one distinguishable, seems to be bent horizontally towards the stem and to end in a fork. To r., recumbent humped, bull, facing Swastika. To L, a vase out of which grows an. upright tree, perhaps ending in three large leaves. H. P.

18. 1-20 in. by -66 in.; weight 54 grains.

0. Standing deity (goddess) with small waist, facing f., and:, grasping a thick pole at each side of coin. Two bangles on. r. wrist, and one or more on 1.; anklets on legs. A broad, bead on each side of neck.

R. Raised Swastika, the side uprights end in points ; ends-of arms are not visible. Bull and vase as usual.

Col. Mus.

17. i-2i in. by -74 in.; weight 52 grains.

O. Vigorous figure of deity (goddess) with very narrow waist, facing f.; apparently an anklet on each leg. A raised bead on each side of neck. Hands seem to hold usual upright poles.

R. Opposed to 0. Raised Swastika, turned r. Basal uprights end in points ; ends of arms not distinguishable. Usual bull and vase, transposed to L and r. respectively. Vase has flat base, very narrow mouth, out of *which grows a plant of three shoots each ending in one leaf. Col. Mus.

18. 1*20 in. by -68 in.; weight 51 grains.

O. Standing deity, facing f.; one or two anklets on each leg. Upright pole on r., at edge of coin.

R. Raised Swastika, turned 1, on wide stem, with very short basal uprights. To 1, recumbent humped bull facing base line; the only instance in which it does not face the Swastika. To r. the vase also inverted, with plant of three shoots each ending in a leaf. Col. Mus.THE MULLEITTIVU COINS 481

19. 1-20 in. by 71 in.; weight 29 grains.

O. Standing deity with wider waist, facing t; two bangles on r. wrist, and a bead above each shoulder. Feet and other hand not visible. Upright pole to r. appears to end in bident with prongs Inclined outwards. Portion of arched line over head is visible. On each side of legs are reliefs that I cannot identify.

R. Opposed to O. Swastika turned r. Bull with curled tail; two shoots of plant visible. Col. Mus.

20. A portion only. -97 in. by 76 in.; weight 19 J grains. 0. Standing deity, facing f.

R. Recumbent bull, repunched from O. R. Massie.

21. A portion only. -92 in. by -56 to -64 in.; weight 42 grains.

O. Standing deity, perhaps facing half 1. Bangles and anklets as usual, and upright poles at sides ; the pointed head of one, at shoulder-level, appears to show that they are javelins. A bead on each side of neck.

R. Opposed to O. Swastika, turned r. Design to r. indistinct; to 1. a full-bodied vase in good relief, with small mouth and distinct lip. Parts of three shoots each terminated by a leaf. ? Brit. Mus.

22. i-io in. by -68 to 73 in.; weight 47^ grains.

O. Standing deity with narrow waist, facing 1 Each hand holds upright pole at side of coin, that to r. having a square knob at its base. Bangles and anklets as usual, and a bead on each side of neck. Circlet passes over head, appearing to rest at each end on the side poles.

R. Indistinct; part of Swastika, Usual bull and vase; plant of two shoots each branching into two. ? Brit. Mus.

23. To these I add one coin purchased for me at Anura-dhapura by my friend Mr. Balfour of the Irrigation Department, as it is of the same type as the foregoing, and unlike others dug up at that town which are described below.

1-09 in. by -62 in.; weight 41 grains.

O. Standing deity with narrow waist, facing 1 Anklet on 1. leg. Arms as usual; L one holds an upright pole which appears to become a bident at level of neck.

11482 ANCIENT CEYLON

R. Raised Swastika turned r. Bull as usual; one stem of plant on L, ending in a leaf. H. P.

24. Another of this type found by Mr. Bell at Anura-dhapura,1 at the structure surrounded by a ' Buddhist railing/ measured 1-12 in. by -7 in., and weighed 5i|- grains.

O. Standing deity, facing f. holding the shaft of a trident in each hand.

The raised Swastika, turned r. Recumbent bull on r., and plant of three stems on L, each ending in a leaf, and springing from a cross-bar. Col. Mus.

THE ANURADHAPURA COINS

These coins, with the exceptions of the two last described, represent a type which in some instances differs in important respects from either the Tissa or Mulleittivu coins. In the case of at least seven of them the figure is seated, "with the leg doubled up so that the heel approaches the body; the right leg hangs down. In each of the seven the L hand rests upon the L thigh, and r. arm is bent upward, the hand being raised to the level of the shoulder, and probably holding a flower. The other coins have standing figures on O.

On R., the humped bull does not appear to occur on the coins I have seen, except on 23 and 24, its place being taken by other symbols and marks which are usually indistinct; in one instance the Aum monogram is present, and probably in another the Swastika. The Vase is visible on some coins, and may have been on all originally. In all the specimens the workmanship differs from that of the Tissa coins, and the die for the 0. was often not much larger than the figure. I believe that all these coins are now in the Colombo Museum, with the exception of one which is in my own collection.

25. i-06 in. by -56 in.; weight 49! grains.

O. Standing goddess with small waist, facing 1, holding upright pole at each side of coin, that on r. possibly ending in trident at level of chin. A bead above each shoulder, and apparently anklets. To r, of legs a symbol resembling an

1 Arch. Survey of Ceylon. Fourth Progress Report, pp. 4 and 13.THE ANURADHAPURA COINS 483

upright post with its top turned towards legs. To L a symbol like tke supposed lotus leaf of No. 14, but smaller.

R. Opposed to O. Raised Swastika turned r., with very broad arms, the ends of which, as well as those of the basal uprights, are pointed. Indistinct symbols to r. and L, that on 1. being forked and somewhat like a * taurine' emblem.

26. 1-02 in. by -56 in.; weight 13 grains.

0. Standing deity, facing f., with arms as usual. Two bangles on 1. wrist, and anklets high on legs.

R. Opposed to O. Raised Swastika turned L A round elevation to r. may be a vase.

27- 1*06 in. by -56 in. ; weight 24 grains; lower edge irregular. *

0. Standing figure of deity with very narrow waist, facing f. Arms as usual, two bangles being on 1. one. Two anklets on 1. leg; perhaps a third above calf. To r. of legs a raised bead opposite the knee, resting on an upright post, the top of which is pointed.

R. Raised Swastika. Little to be seen.

28. 1*01 in. by -52 in.; weight unknown.

0. Standing deity as usual; two bangles on L wrist* A cloth from waist nearly to knees. Upright pole on r. may end in bident at shoulder level. A bead above each shoulder.

R. Raised Swastika. Indistinct.

29. 1-04 in. by -54 in.; weight unknown.

O. Standing deity, stouter than usual, facing 1, with two bangles on L wrist; hand holds upright pole which may end in bident at level of neck.

R. Indistinct.

?t-

30. 1-07 in. by -43 in.; weight unknown.

Both faces indistinct, but part of standing figure visible.

31. 1*01 in. by -53 in.; weight 17 grains.

O. Standing deity facing 1, and holding poles as usual Two anklets on each leg and a bangle on each wrist. Circle of beads may have passed over head, two being visible on each side. Upright poles may have ended in bidents.

R. Very faint, but part of stem of Swastika can be seen ; perhaps part of vase to r.484 ANCIENT CEYLON

32. 1-17 in. by -51; weight 22 grains.

O. The die has moved and rendered design blurred. A goddess standing facing f., holds on r. a bident. Bead on each side of neck. To L of r. ankle a large bead rests upon a thick upright stem which is widened out at the top; this resembles the 'disk on the altar/ To r. of legs, an upright bar with upper part turned r.

R. Opposed to 0. Raised Swastika, turned r. On 1. a raised oval may be the vase.

33. 1-04 in. by -43 in.; weight 17 J- grains.

O. Standing goddess (? MuktakesI, a form of Durga, with loose hair) thin, with wasp-like waist, facing f. Short skirt from waist to mid-thigh. Long loose hair incised after stamping, in fine lines, hangs down on L, reaching below waist. A bead above 1. shoulder. Side poles not visible. Mark of top of die extends in an arc over the head.

R. Opposed to O. Part of raised Swastika.

34. 1-17 in. by -62 in.; weight 45£ grains.

O. Standing deity, facing 1, with cloth apparently folded round the hips. Arms hang down as usual. Trace of top of bident on L

R. Raised Swastika, turned 1. To r., large round-bodied vase, with lip. To L, a symbol resembling an early letter n.

35. 1*10 in. by -46 in.; weight 30 grains. Purchased for me in Anuradhapura.

O. Standing goddess, facing f., and holding poles as usual. The head is gone; navel distinctly marked. Two bangles and an armlet on r. arm ; anklets on both legs. A row of five or seven beads round head, and perhaps a horizontal row above them. Under the feet three short vertical raised lines. Upright on 1. widens out at shoulder level; both are too much worn for tops to be discriminated. A line on L between the leg and side pole may be the edge of her dress. . R, Raised Swastika, turned r,; the lower arm appears to be straight and extends horizontally to r. and L of centre line. Two narrow parallel waved lines below base line, . .- . HP.THE ANURADHAPURA COINS 485

86. 1-25 in. by -76 in. ; weight unknown.

O. Seated figure, turned f., with face to 1. Two anklets on r. leg, and one or two on the other. The 1. leg doubled under body ; r. leg hangs down from knee. The 1. arm hangs down and hand rests on 1. thigh. The r. upper arm hangs down and fore-arm bends out horizontally. Hair-knot, or head-dress, or helmet extends at back of head. Raised work in front of face, some of it being part of an arched band that passes round and over head.

R. Swastika, turned 1. No bar below its base line. To L, a symbol which may be the vase ; to r., a small symbol indistinct.

37. 1*20 in. by -82 in.; weight 74 grains.

O. The die merely includes a sitting figure (? female) in the same posture as last. Two armlets on L upper arm, and an anklet on 1. leg. Hand on r. holds a flower on its stem before the face. One bead above 1. shoulder. A head-dress, or crown, or helmet on head, above which the die ended in a point.

R. Opposed to O. Indistinct. Swastika, and symbol on r.

38. 1*09 in. by '62 in.; weight 40 grains.

O. Lower third of coin untouched by die. Seated figure (? female) in the same attitude, with one bangle on 1. wrist, and two anklets on 1. leg. The arm bent up near shoulder on r., and flower held at shoulder level. A bead on r. of head.

R. Opposed to O. Swastika turned L, with thick base line. To L, vase with three shoots ; to r., a symbol indefinable.

39. 1*21 in. by -71 in.; weight unknown,

O. Seated figure (? female) in same attitude, bangles on wrist, two armlets on L upper arm, two anklets on 1. leg, perhaps only one on r. leg. Fore-arm on r. raised to shoulder level. The die did not include more.

R. Very indistinct.

40. 1-24 in. by -75 in.; weight unknown.

O. Seated figure in same attitude ; bead over L shoulder. Five nearly 'upright short lines below L foot and calf, rising from a horizontal one which turns down on L close to r. foot, may indicate a seat or throne. Hand raised to level of. head on'r, A bead between 1. arm and waist.486 ANCIENT CEYLON

R. Raised Swastika. Vase to r., with shoots each ending in a flower or leaf.

41. 1-31 in. by -74 in.; weight unknown.

O. Seated figure in same attitude, but r. leg doubled under body, and 1. leg doubled under r. one. Two bangles on 1. wrist, and one or two on r. one. A bead on each side of waist. Hand raised to shoulder level on r. A curved band seems to pass round head. Upright pole near 1. edge of coin, ends invisible.

42. 1-50 in. by -79 in.; weight 74 grains.

0. Seated figure with wider waist, in usual attitude. Hand on r. holds flower at shoulder level. A bar passes down from 1. knee and is then turned horizontally to r. toes ; it may represent the outer side of a throne, or the side of the dress.

R. Opposed to 0. Raised Swastika, turning L, rising from a waved base line, below which, and separated from it by a wide channel, is a straight horizontal band. To L, the vase with two round flowers or leaves above it. An emblem to r., like a bull's head looking downwards.

43. 1-43 in. by -72 in.; weight 75 grains.

O. Standing deity, facing 1, as usual; one or two bangles on r. wrist, and anklets on r. leg; other hand and foot not visible. Upright pole on r. Lines are incised, curling from the top of the head upward and outward, evidently to indicate loose hair.

R. Raised Swastika, turned L, on strong stem which springs from a wide straight base line, below which is a flat channel. The short basal uprights are pointed. To L the vase. To r. no emblem visible.

44. 1*42 in. by -72 in.; weight 64 grains. A hole is drilled near the top for suspending the coin on a string.

O. Standing deity, with very long narrow waist, facing f. An armlet on r. upper arm, bangles on each wrist, and anklets on legs, two being on L leg. Opposite the hips the hands grasp two upright poles; that on 1. ends at level of head, apparently in a trident. Top of the other uncertain. A symbol on each side of legs; that on 1. may have been a bead or disk on a post but the bead is not now visible.THE ANURADHAPURA COINS 487

R. Raised Swastika, turned r. The short basal uprights on 1. are pointed ; those on 1. have square ends. Vase to r., out of which grow incised leaves on stems, one having shape of a Bo-leal Under the vase is a wide bar curved into an arch. To I, an indistinct symbol, resembling a bird with raised wing,1 facing Swastika.

45. i-2i in. by -70 in. ; weight 58 grains.

O. Middle part of design gone, and rest fragmentary. The details appear to differ from those on other coins, and it is doubtful if a deity was represented. Two curved bands ending in curls are on 1., and perhaps a vase below them.

R. Raised Swastika, turned 1.

46. i -45 in. by -76 in. ; weight 82^ grains. A hole is drilled near the top for suspending the coin on a thread or fine string.

O. Standing deity, facing f. There may be a helmet, or crown, or raised cover on head. A wide broken line passes over head from shoulders. On each side the hands hold an upright pole which appears to end in a bident or trident. On L of legs an upright design.

R. A raised border, excepting on r. Swastika turned r. A horizontal bar separated from base line of Swastika and border by two channels. To r., probably a smaller Swastika turned r.; to L, a rectangular raised line like early letter u.

47. 1*53 in. by -65 in.; weight 126 grains. A hole is drilled near the top for suspending the coin on a thin string.

O. Standing deity (? female) with wasp-like waist and hanging arms ; hands hold upright pole at each side of coin. That on 1. ends in cross-bar at level of shoulder, above which are four flowers, lower two being circular and upper two trumpet-mouthed. The pole on r. winds slightly and ends in a thick curl at level of neck. Above this a relief like a standing deer ; but lower part maybe intended for a flower and the rest part of a band passing over head. Thin transverse bar above it, separated from border by a channel.

R. Large raised Swastika, turned r. Under its base line a transverse band separated from border by a narrow channel.

1 A bird with raised' wings is carved in false relief on a pillar at the Abhayagiri dagaba, and on one at the Dakiinu dSgaba.ANCIENT CEYLON

To r. a clear Aum monogram, with straight sides. To 1. side view of flower resting on a thick cross-bar, bud on r. of it.

48 to 51. Four other thin coins, averaging 13^- grains in weight, have the following dimensions :?i-oi in. by -45 in. ; 1-05 in. by -50 in.; 1-05 in. by -50 in. ; and i-io in. by -50 in. All much worn on both sides.

52. An additional coin of this type found by Mr. Bell at Anuradhapura,1 at the site with the * Buddhist railing/ measured *8o in. by '62 in., and weighed 44 grains.

O. Standing deity, facing f., holding shaft of trident in his r. hand, and perhaps another in 1. hand, which is indistinct. Mr. Bell thought that irregular upright lines near these were the edges of his dress.

R. Raised Swastika, turned r.; plant on L, of three stems, springing from a cross bar. Indistinct marks on r.

The mean dimensions of the Tissa and heavier Mulleittlvu coins are :

Tissa coins, 1-18 in. by -49 in.; weight 46 grains (mean of 3). Mulleittivu coins. 1-13 in. by -67 in.; weight 50 grains (mean of n).

The Anuradhapura coins differ greatly in weight, which varies from 13 grains to 126 grains.

Sir A. Cunningham has given two scales of the weights of ancient Indian money, one for copper coins and the other for silver coins.2 These are as follows :?

SCALE OF COPPER COINS. i Kahapana . . 144 grains. I ? ? ? 108 ?

I"-,, - - 72 .,

3- ? - - 36 ?

J- ,, - - 18 ?

SCALE OF SILVER COINS. 2j Kahapanas . . 144 grains. i£ ? . . 86-4 ?

i ? . . 57-5 ,,

1 ? . . 28-8 ?

I ? - - J4'4 »

In the later Sinhalese coins we find both the copper and silver coinage following the same copper scale. This is seen in the following table, which gives the mean weights of some of the ordinary * Massa' coins in my possession, taken

1 Arch. Survey of Ceylon. Fourth Progress Report, pp. 4 and 13.

2 Coins of Ancient India, pp. 46 and 47.THE EARLIEST COINS

489

without selection. I annex for comparison the weight of some coins of the south Indian king Raja-Raja, purchased by me in Madura, from which I have excluded only coins that are evidently cut away at the edges.

Date. Name. Number of Coins. Mean Weight. Grains. Material.

1153-1186 Parakrama-Bahu 25 62-5 Copper

1186-1187 Wijaya-Bahu .... 14 55'6 ,,

1197-1200 JLilawati ...... CO 65-4

J

I2OO-I2O2 Sahasa-Malla 5° 63-1 »

I2O8-I209 Dharmmasoka-Deva . *4 &4"3 ,,

1296- Blmvanaika-Bahu . 25 653 »»

Il86-Il87 Wijaya-Bahu . . . IO 64*1 Silver

985?1011 Raja-Raja (Indian) 25 63*4 Copper

The oblong money can be divided only into a larger and a smaller type, as shown in Fig. 155. The former includes the Mulleittivu and the larger Anuradhapura coins, and has a mean width of -70 in. ; the rest of the coins average -50 in. in width. Whether these sizes indicated different values is doubtful.

The great variation in the weights proves that no special scale was followed in them; the plaques were tokens rather than money. Yet they may have answered all the purposes of money in being used as mediums of exchange which probably liad fixed values in the country.

Histories and inscriptions alike prove that coins called kahapanas existed in countless numbers in Ceylon in very early times ; yet no other coin which could possibly represent this money has been discovered. That such coins were made of copper is rendered certain by the discovery of the circular coin described below (p. 503), which appears to be a double kahapana, as Mr. Still stated. Necessarily, this must have been preceded by the single kahapana and its subdivisions, which could not be formed of a more valuable metal than the money of higher value, and therefore must have been copper coins. Thus, until some other form of copper money of suitable weights has been found it appears to me that490 ANCIENT CEYLON

these oblong plaques must be accepted as partly filling the gap. Mr. Bell's spurious oblong Purana, with a figure on the obverse like those on the plaques, strongly supports this view.

Cast coins of the same size and shape occur in southern India (see p. 506). The slight amount of wear in most of the plaques may be due to their being hoarded as amulets; some are considerably eroded on their faces. In the irregularity of the weights the coins only followed the example of the Puranas found in Ceylon, the weights of which show that while all probably had the same value as mediums of exchange they were in reality tokens, that is, they did not circulate in Ceylon at their intrinsic value. The surprise which the Sinhalese king expressed to the freedman of Annius Plocamus at the exact weights of the Roman coinage is a proof that all the local money varied greatly in this respect.

That the oblong type of coin continued to be issued up to the third or fourth century A.D. is clearly proved by the form of the ' Aum' monogram on the coin nuipbered 47, the m of which is of a type which is found in some inscriptions of that period. I met with a similar letter cut on the faces of two stones inside the valve-pit or ' bisdkotuwa' of a sluice at Hurulla, a tank constructed by King Maha-Sena (277-304 A.D.). Large coins of a circular shape made their appearance at about this time, having a similar ' Aum * monogram on them, and it may be assumed that the issue of the oblong money then either ceased or was of less importance than before.

As all probably had a two-fold value as coins and also as protective amulets the discovery of a few isolated specimens about religious edifices of a later date does not quite prove that they continued to be issued up to that time.

Two years ago Mr. Still mentioned that he had examined some 200 specimens, among which were three cast ones with outward-curving sides, found near the Thiiparama, Another cast one was found ip the excavation inside the Kiribat dagaba, and a fourth near the Thuparama. (Journal, R,A.S., Ceylon, 1907, p. 199 ff.).THE EARLIEST COINS 491

The special Swastika symbol of all the early Sinhalese coins, including also the large circular coin just mentioned, which will be described later on, is cut at the beginning or end of three pre-Christian inscriptions in Ceylon, and it was also discovered by me engraved on the outside of pottery taken out of the lowest stratum of the remains at Tissa. Its occurrence there proves that it had been adopted in Ceylon as early as the second or third century B.C. It is cut at the beginning or end of the inscriptions numbered 69,70, and 75, which belong to the first century B.C. The central bar and four side uprights are found in the symbol which precedes the inscription numbered 62, by Prince Sali, which dates from about the middle of the second century B.C. Although I believe it does not occur at any inscription of post-Christian date, its presence on the oblong coin No. 47 and the large circular coins shows that it continued to be employed as a local symbol until the fourth century A.D., or later. It appears to be unknown in India.

The Indian meaning of the Swastika, the cross with bent arms, is Su + asti, * it is well/ that is, * may it be wel!7 It indicates its luck-bringing power as an auspicious wish, and the words themselves in the form Swasti are cut at the commencement of numerous later inscriptions in Ceylon. But the symbol goes back to a date that is far anterior to any such interpretation. Its earliest occurrence is, I believe, at the first city on the site of Troy, the inhabitants of which are considered by Mr. R. H. Hall to have been * just on the border between the Age of Stone and the Age of Metal' * ; and their latest date is stated by this authority to be about 2500 B.C. (op. cit. p. 49). As the Swastika was found by Dr. Schliemann on pottery at the bottom of the stratum belonging to this early race it may belong to the fourth millennium B.C. It also occurs in Egypt as a decorative motive in the ceilings of the Theban tombs of the eighteenth dynasty (1700-1400, B.C.).2 Its

1 The Oldest Civilisation of Greece, p. 23.

2 Perrot and Chipiez, Hist, of Ari in'Ancient Egypt. Vol. ii, p.

359 (from Prisse) ; Prof. Maspero, Egyptian Archaeology, p. 16 ; Errnan, Life in Ancient Egypt, pp. 102, 397,, and 479,492 ANCIENT CEYLON

highly developed form in that country proves that it was known there long prior to its use in these tombs.

It may have been carved at the inscriptions, and may also be placed on the coins, as a special emblem of Good Luck or Prosperity, which acts as a protection from evil influences.

In describing the inscriptions I have already suggested that the four short basal uprights may typify the Four Great Buddhist Truths, as supporters, or more probably, especially on the coins, the four-fold forces?chariots, elephants, cavalry, and foot-soldiers?of the sovereign protecting the emblem, the prosperity of the country being supposed to depend largely on its ruler. In that case the central pole on which the Swastika is elevated might represent the sovereign as upholder of the prosperity of the country.

In other countries the Cross is sometimes drawn with a short bar across or near the end of each arm, and it is of interest to observe that in the case of the Swastika on coin No. 14 two thin bars are thus shown across the terminal parts of each of the two arms the ends of which are visible, as well as across the ends of the short uprights. A Swastika with one bar of this kind is also represented on coin No. n of Plate X in Cunningham's Coins of Ancient India. As every line in ancient symbolism has its own meaning there must be a special reason for inserting these peculiar cross-bars.

The only explanation with which I am acquainted, of this barred, or as he terms it * guarded' Swastika, is that given by Mr. J. M. Campbell, of the Indian Civil Service, in Vol. 24 of the Indian Antiquary (1895)l?that such lines are due to a belief that any cross, or, in its usual Indian form, the Swastika, is a favourite house for spirits. He supposed that the crossbars at the ends of the arms were intended to prevent the ready egress of good spirits who might have been induced to reside in it, and thus to ensure its beneficial or protective action. It is evident that, as he also remarked, they might equally be drawn to prevent the entry of evil spirits who might desire to take up an unauthorised abode in it, and this is the more

1 On the Spirit Basis of'Belief and Custom, p. 164.THE EARLIEST COINS 493

probable explanation of the cross-bars, as I shall endeavour to show in a subsequent chapter.

The guarding power of labyrinthine and spiral and meander designs, or zig-zag or crossed lines is, as Mr. Campbell pointed out (loc. cit p. 161), the cause of their constant employment in charms against evil spirits at the present day, both in Ceylon and elsewhere.1 Thus the partiality which the people of the East as well as those of the West and America, have exhibited for the Swastika is doubtless largely based on the belief in its defensive properties against these malignant beings.

In addition, therefore, to its symbolic aspect as an emblem of Prosperity, these cross-bars prove that the Swastika was placed on the coins to fulfil another function, that is, to be a protective charm against the malevolent actions of evil spirits. The pointed ends of some of the arms and short basal uprights may be also due to a similar idea?that of closing them against the ingress of hurtful spirits who might neutralise the omen. A fuller elucidation of the probable origin of the Swastika will be found in a later chapter.

If this was the ancient notion regarding the powers of the Swastika, it will be understood that apart from the general belief in the luck-bringing properties of everything that turned to the right and followed the course of the sun, it would be a matter of comparative indifference, as regards its spiritual aspect, whether its arms turned to the right or the left. In either direction they would equally act as a check to spirit progress. Thus, out of the 52 coins described above, in 18 cases the symbol is indistinct.; on the remaining 34 coins the arms turn to the right in 22 instances, and to the left .in 12.

The line or two lines, which are sometimes waved, below the base line of the Swastika may represent a snake or snakes, which also have guardian powers against evil spirits, especially in the East.

The other designs on the reverse of the Mulleittivu coins admit of simple explanations. The plant growing out of a

1 I have a Sinhalese MS. book of charms and spells against sickness and evil spirits and planets, in which designs of crossed and complicated lines constantly occur.494 ANCIENT CEYLON

vase indicates that the latter is filled with water; and the full vase is well known to be a general emblem of Good-luck which is much employed in the East, the reason being, as may be gathered from the Vidhura-panclita Jataka,1 that if the vase be full it cannot be imperfect. It is thus an emblem of Perfection, and therefore most auspicious. It is not a special symbol of Buddhism.

The recumbent humped bull is the special emblem of the Solian kings of southern India, and its appearance on these coins of Ceylon must point to Solian influence in the country. The coins which have this symbol may thus have been issued in the first half of the second century B.C., by the only Solian King who reigned in Ceylon for a considerable period at an early date, that is, Elara, whose rule is alluded to in very favourable terms by the pre-Christian Buddhist annalists, and who occupied the throne from 205 to 161 B.C. There is no probability that an early Sinhalese king would insert this South-Indian symbol on his coinage, and it is not found on the Tissa nor, with two exceptions,2 the Anuradhapura money which I have seen, that must have been issued by native rulers.

Even if the coins of this type were issued by the Indian usurpers who ruled the country from 104 to 88 B.C., those found at Mulleittivu appear to have been buried iii the first century B.C.

If it be held, however, that the mark on the Purana (i) is the letter hu its shape must prove that the Mulleittivu coins were buried in post-Christian times; but the good state of many of the symbols on the Puranas does not support this conclusion.

On the Obverse I take first the seated figure on the Anuradhapura coins, which can be explained without difficulty.

Among the articles found in removing the debris left round the Yatthala dagaba at Tissa, which, it may be repeated, dates from the third century B.C., there was, by extreme good-luck, a little more than the half of an admirably cut and polished

1 The Jdtaka. No. 545. Translation, p. 152.

2 If Elara issued this coinage some examples of his coins would occur among later hoards, of course.THE EARLIEST COINS 495

thin carnelian of an elliptical shape and perfect colour, which had evidently been the stone set in a seal-ring. The persons who rifled the relic-chamber of the dagaba apparently wanted only the gold setting, and broke and rejected the stone, which remained buried among the brick rubbish thrown out of their cutting. It was discovered when the recent restoration was begun in 1884, and the Buddhist Committee who supervised the work were so good as to present it to me. The other Jialf of the stone was not found. The portion in my possession is a regular ellipse, measuring -80 in. in width, and probably 1-20 ins. in length when perfect. Its present length is -64 in., and the middle thickness is -13 in. An impression of it is shown, considerably enlarged, in Fig. No. 156. I am indebted to the skill of Mr. Norman May, of Malvern, for this admirable reproduction of this interesting seal, in the exact state in which it was left by the camera.

On this portion there is excellently engraved in intaglio the figure of a person sitting upon an ornamental chair, which can be no other than a royal throne. In the impression taken from it the face is turned to the right and the body half right. The king is leaning slightly backward in an easy attitude with his right foot hanging down from the throne and his left leg doubled so that the foot is placed on the chair. His left arm rests above the elbow on the raised left knee, and the fore-arm and hand are elevated, and hold a flat object, at which he is looking, in front of his left shoulder. His right arm hangs down and grasps near his hip a thin sash which passes over the right shoulder and back round his left side, the two ends, which appear to be fringed, standing out at the back of the chair.

He is very simply dressed in a cloth from the waist downward ; the top of it is shown passing round the waist, and its edge hangs down from the left knee, while its folds are clearly seen on both thighs. Round the base of his neck is a thin necklet, and a plain armlet passes once round the arm above each elbow. No bangles are on the wrists ; his ankles are not visible, having been on the missing portion of the stone. No hair is represented on the face; that on his head is cut short, and simply thrown back from the face in loose masses,496 ANCIENT CEYLON

without reaching the neck. There is no hair-knot. His nose is prominent and quite straight, and his forehead rather high.

The throne is of a very interesting shape. The side is an oblong; enclosed in a plain frame there are four horizontal rows of square hollows, each row now consisting of seven, but apparently nine on the full design, separated by raised bars; this represents very open basket-work. The right corner rests upon two feet,1 which are formed of round balls placed upon flat bases. The whole back of the chair winds backward, and the end of the upright .bar at the side curls over above the transverse bar, which passes quite through this upright and across to the other rear upright, immediately below the level of the shoulder.

The engraver has taken great pains to make it perfectly clear that this side upright of the back is a rustic one, and he has shown five short branches projecting from it and cut off at a distance from it equal to about its thickness. This rustic post passes down to the feet of the throne, and into the lower horizontal bar of the frame of the basket-work. From the points where the branches unite with the stem three curled ornaments spring upward on the outer side, the two lower ones ending in a curl which turns inward to the upright, and the top one curling outward below the level of the cross-bar at the back of the throne, and terminating in two tassels which hang from its end.

At the level of the king's face the tip of another design appears at the fractured edge of the stone ; it consists of four leaf-like projections in close contact.

There can be no reasonable doubt that this gem was deposited in the relic-chamber of the dagaba along with the relic-receptacles which have been described in a previous chapter, and it may be assumed that it dates from some time prior to the original construction of the dagaba. When it was submitted twenty-four years ago for the inspection of the

1 We learn from the Maha Hansa Jataka, "No. 534, that one royal throne ha l eight feet.THE EARLIEST COINS 497

authorities of the British Museum, the opinion expressed regarding it was that it is of Indian origin and workmanship, and that it might perhaps belong to the seventh century A.D. ; but on its being re-examined in 1903 in the light which increased knowledge of early Indian art throws upon such designs, it was considered to be of pre-Christian date, and perhaps to go back to the third century B.C. but to no earlier period. This authoritative opinion is therefore entirely in favour of the arguments previously advanced regarding the age of the gem and relic receptacles, since all probabilities forbid the assumption that the dagaba was re-opened, and these articles and especially the two Puranas also found with them were afterwards placed in it in either pre-Christian or early post-Christian times.

In the Mahavansa we read of numerous presents passing between the great Indian Emperor Asoka and the Sinhalese monarch Devanam-piya Tissa, the brother of Maha-Naga; and there are accounts of at least two embassies that Tissa sent to As5ka's capital, Pataliputta, on both occasions the king's nephew, Maha Arittha, being the ambassador. This prince afterwards became a monk, and according to the Dhatuvansa resided at Tissa. Thus we get a direct communication between Tissa and Asoka's capital.

It may be surmised that either the Prince-monk, or much more probably King Maha-Naga or his son Yatthala-Tissa, deposited this finger-ring in the relic-chamber on the occasion of the festival that would be held at the time when it was closed. In the next century, at the closing of the relic-room in the Ruwanwaeli dagaba at Anuradhapura we read (Mah., i, p. 122) of King Duttha-Gamini that * while [he was] within the [relic] receptacle he made an offering of all the regal ornaments he had on his person/ The Dhatuvansa, in relating the account of the deposition of relics in the Seruvil or Serawavila dagaba by King Kakavanna-Tissa, doubtless describes what usually occurred at important structures of the kind. It says, * All the dancing women offered the ornaments that each one was wearing. Then the king and the great ministers, etc., having taken off the ornaments that each one was wearing offered

K K498 ANCIENT CEYLON

them in the relic chamber/ Maha-Naga or his son may have acted in a similar manner at the Yatthala dagaba.

As the gem is an early Indian work, exhibiting strong Greek influence and therefore probably not of south Indian origin, and as it seems certain that it represents a king on his throne, it is quite possible?one might even say probable?that the figure is that of Asdka himself, or is copied from representations of him.

The nearest approach to the attitude of the king which I have found on early Indian coins is that of the sitting Herakles on the coins of Euthydemos, King of Baktria (circa 230-200 B.C.), as he appears in Plate I, Nos. 3 and 5, of Mr. V. A. Smith's 'Catalogue of the Coins in the Indian Museum, Calcutta/ The Indian engraver took a nearly similar design, and adapted it to Indian requirements by raising the bent leg till the foot rested on the throne, and giving the raised hand a small object, possibly a flower but not recognisable as such in this example, to hold in place of a club.

The sitting figure on several of the oblong coins is in the same attitude as the king in the gem, with the exception that one hand rests on the thigh instead of holding a scarf or sash. On both gem and coins one leg hangs down while the other is doubled up; and one hand holds a flower or other object near the level of the shoulder, while the other hangs down to the level of the thigh. In the figure on this gem, therefore, we have the original Indian design of the sitting figure on the oblong coins, as well as the original type of the sitting king on the later coinage of Parakrama-Bahu I and his successors. On these last the throne has degenerated into one or two horizontal lines with short vertical lines crossing them? the basket-work on the gem.

The similarity of the design on the gem and the later coins, where the monarch's name at his side leaves no doubt that the figure is intended for him, renders it most probable that the sitting figure on the oblong coins is also intended for a representation of the ruler of the time. The monarch is placed in nearly the same position on many Indian coins; it was the conventional attitude in delineations of the seated kings.157. liiianl-stone Anuradhajmra .THE EARLIEST COINS 499

In the case of the standing personage on the oblong coins the identification is less obvious. Notwithstanding the fact that the upright figure on the later Sinhalese coins has always been termed a king in the descriptions of them, all the available evidence shows that the design is intended to represent a guardian deity, and not always the same one.

It is certain from the distinctly marked breasts that the standing figures in the coins numbered 25,33, and 35 are those of females; and not unlikely that the figure is intended for that of a female on several other coins on which the wide bust and hips and the extremely narrow waist are the characteristics of female rather than male forms.

The winding stems ending in flowers which some of the more masculine figures hold are special characteristics of the Guardian Deities or Dwarpal who are carved on slabs erected on each side of the steps at the entrances of numerous Buddhist buildings at Anuradhapura and elsewhere. In their case, in one hand a curved stem is held which ends in a high cluster of flowers, while the other supports a vase out of which several flowers rise *; but there is (or was in 1873) one example in which a curved stem ending in a flower (lotus) or bud is held in each hand. Thus, on the coins on which this special design is found the figure seems to be that of a Guardian Deity rather than the Monarch. If so, it is most probably that of a deity on the other oblong coins, and also on the later coins.

This view is strongly supported by the design on the coins which I have numbered 24, 44, and 52, where the article held is a trident, the symbol of Siva and his wife, and of his son Ganesa,2 and perhaps also of his other sons, Skanda, the God of Kataragama, who is now considered to be one of the Four Guardian Deities of Ceylon, and Ayiyanar, the Guardian

1 The illustration (Fig. No. 157) shows one of these Dwarpal of an early type, near the Thiiparama dagafoa. The animal on the side pilaster is a horse, for which, curiously enough, a heap of provender is provided ; he appears to be eating it.

2 In an ancient temple of this deity at Omantan in the Northern Province, a stone trident stands on an altar as the God's emblem, the central prong representing a lingam.5oo ANCIENT GEYLON

Forest Deity of Ceylon. The javelin and the apparent bident which appear on several coins also point to the latter gods or Durga as being the deity who is commonly represented. In some unworn coins of Wijaya-BaJm, also, a weapon with a long sharp-pointed head is distinctly shown at the side of the article held in the right hand of the standing figure ; it resembles the weapon at the side of Skanda on coin No. 9 of Plate VI, C.-A. L, and No. 15 of Plate XXI, Ind. Mus. Cat. In the Pandiyan coin No. 143, Plate IV, of Elliot's Coins of S. India a similar figure who has the trident at his side must be Siva or one of his sons.

In the same manner as in the later Sinhalese coinage, the king is delineated on one face of many Indian coins, and a deity on the other. In the Gupta coinage the latter is often Siva or a goddess; but Skanda also appears in other coins, and he would "be specially appropriate for the Ceylon money on account of Ms local connection with the island. As for the bangles and anklets, all the Dwarpal in Ceylon have them.

That the figure is a deity is also indicated by the presence of the arched line or circlet of 5, 7, or 9 beads which in some cases passes round and over the head of the standing figure, but not over the head of the seated person. Each of the Dwarpal in Ceylon, with the exception of figures of Bhairava, is protected by the expanded hoods of a Cobra which has 5, 7, 9, ir, or in one instance 13 heads; and in several of these carvings which are somewhat worn the heads stand out from the arched line of the hoods like large beads. Thus it is possible that the beads round the head of the standing figure symbolise, if they do not actually represent, the many-headed cobra guarding or sheltering him.

Where one bead is shown on each side of the neck it is

merely the ear-pendant. When near the waist it is the fold

of the sash wMcfa holds up the cloth, The arched line which

overhead in some coins may be a * chatta ' or umbrella,

with a scalloped fringe in some instances.

I therefore, that in all cases the standing figure

on the Sinhalese coinage, whether ancient or more is a guardian deity and not the Icing.FIG. 158. Durga, as Kali, destroying the Asuras (Tanjore Temple).

? P- 501THE EARLIEST COINS 501

In the later coinage the peculiar article held by him, which some have supposed to be a weapon, is a double ' trisula' resting on a circle or lotus that is represented by the bead under it, exactly as it is seen in the post-Christian Ajodya coin No. 15, on Plate XIX of the Indian Museum Catalogue. The double trisiila is also found on the early Yaudeya coin No. i of Plate VI of C. A. /., which is said by Sir A. Cunningham (p. 76) to date from about the first century B.C. ; and on the Eran coin No. 19 of Plate XI of that work. It also appears on the Andra coin numbered 14, in Plate II of Sir W. Elliott's Coins of Southern India. This design is not recognisable on the oblong coins that I have seen.

Whether it was developed from the Greek caduceus, which occurs (or a symbol like it) on some Indian punch-marked coins, is uncertain; whatever its origin, it may have been perpetuated in its present form not only as a lucky emblem, a form of trident, a weapon greatly feared by demons, but also as a monogram that might be interpreted jay a, ' victory/ if the lower part be read as the letter ja and the upper part as ya?as its shape on the Ajodhya coin seems to indicate. In the latter meaning it would be a particularly appropriate emblem for any guardian deity. The word jay a itself is found on coin No. 14 of Plate XX of the Ind. Mus. Cat.

With respect to the female deities who appear on the oblong coins, the weapons which some hold must identify them with some form of Durga, as the slayer of the Asuras or demons. Skanda was also the later champion and leader of the Gods against the demons.

The standing figure, whether male or female, would thus, like the Swastika, be thought to have special protective power against all classes of evil spirits; and that the oblong coins were credited with the possession of beneficial qualities is proved by finding some that were drilled for suspension on the neck as amulets.

The only other distinct symbol on the obverse of these corns is that on Nos. 27 and 32, and perhaps 44, the bead on the post, which has been sometimes termed the disk on the altar. It is found in the reliefs carved on a pillar at the side of one502 ANCIENT CEYLON

of the wahalkadas at the Jetawana dagaba (of the Nandana garden) at Anuradhapura. The other designs included with it there in the spaces of a leafy meander pattern are all emblems that are not exclusively Buddhist, such as the Trisula, the Swastika, the Chank, the Five-headed Cobra, and the Yak-tail Fly-whisk; on other pillars the Elephant, Lion, Bull, the Structure with three arches, and Nondescript animals are carved. Considering the unimportant position which it holds on the pillar, and its small size, it cannot be a Dhamma-Chakka, or ' Wheel of the Law/ such as is worshipped in the Amaravati carvings, and it is not a fan, the circle being little wider than the post in one instance.

A circular fan, with a straight handle, is often carved after pillar inscriptions of the tenth century A.D. in Ceylon, when they contain grants of privileges in connection with monasteries, as one of the common emblems of the Community of Buddhist monks. In the case of the oblong coins, however, it is not probable that this meaning can be attached to a symbol at the side of an Indian deity, where it is much more likely to have some protective function, or to be an emblem of the god. It may be the sun-emblem or discus of Vishnu ; if so, the person at whose side it stands may be that god or his ' sakti ' or female manifestation, Lakshmi, the Goddess of Prosperity.

This symbol appears to be a relic of the early Indian Sun-worship ; it represents the sun as it would appear when it rose due east of the pointer-stone of a sun-temple, on which occasion it would be visible for a moment from the centre of the circle, as a full disk resting on the summit of the stone. In the case of perhaps the earliest existing representation of a pointer-stone, the sun, as an eight-pointed star (with eight intermediate rays of light radiating from a central ball), is delineated as resting on the rounded apex of a tall cone which is carved in relief on the e Stele of Victory ' of Naramsin, King of Agade in the Euphrates valley (3750 B.C.).1

In the coin No. 27 it is clear that the post or column at

1 See the Plate facing p. 160 in .Messrs. 'King and Hail's Egypt and Western Asia, 1907.THE LARGE CIRCULAR COINS 503

the top of which the disk is placed is terminated in a blunt point, like a pointer-stone. There is a round column of nearly the same shape, with a rounded apex, but without the disk, at the side of a three-arched structure surmounted by a crescent, on the Taxila coin No. 6 of Plate II of Cunningham's Coins of Ancient India ; and I have seen quite similar cut pointer-stones, like circular obelisks, on the eastern side of stone circles in the Gambia valley in West Africa.

The learned authors who have described the coins termed Puranas agree that the wheel with straight spokes is a sun emblem and not a Dhamma-chakka; and we know that on each of the faces of the ' tees ' of the early dagabas of Anura-dhapura there was a representation of the sun in relief, which is still to be seen on one of them. A disk with a central flat boss and a circle round it, similarly raised on a pillar which has a base and capital, is carved in relief at the top of the face of each engaged pillar at the sides of the wahalkadas at the Miriswaeti dagaba at Anuradhapura. It has a chatta above it. It appears to be the same sun-emblem, perhaps converted into a Dhamma-chakka (Fig. No. 84).

In these notes on the symbols I have referred to a large circular coin of Ceylon. The first specimen was discovered by me at the Tissa excavations, in digging a channel; it is in the Colombo Museum. Several others have been obtained at Anuradhapura and one at Mihintale. In his Annual Report for 1900, p. 5, Mr. Bell records his finding one in a peculiar brick-lined pit at Anuradhapura, and mentions that about fifty of these coins were discovered at one site on private land at that town ; of these some selected examples were sent to the Colombo Museum. The three which I have seen appeared to belong to the third or fourth century A.D. I append descriptions of the Tissa coin and two others kindly submitted to me by the late Mr. levers when Government Agent of Anuradhapura.

53. A roughly circular copper coin with a mean diameter of 1-27 inches ; weight 220 grains. Found in digging a channel at Tissa. The designs on it and the others were impressed504 ANCIENT CEYLON

by two dies, the marks of which are visible ; they do not rise above the level of the border. Those on the reverse side were afterwards cut more deeply by hand on this coin.

O. The design is surrounded by two parallel circular lines, ?10 in. apart, having between them an intermediate line, broken in one part into a series of dots, and perhaps similarly broken on the opposite side. Owing to the erroneous position of the die only three-fourths of the design is on this face.

In the right lower corner is a well-shaped elephant, facing L, with extended tail. Above it, but to L, a tree standing on a cross enclosed in a square, or surrounded by a fence. On each of the upper corners of the enclosure is a bead or disk surmounted by a crescent, like some so-called ' Taurine ' symbols on Indian coins. The tree has an upright stem from which grow two alternate lateral branches, each, as well as the stem, ending in three leaves, one terminal and the others lateral At the top of the coin and to r. of the tree, the Swastika symbol raised as before and turned r., with four basal supporters. Between it and the tree are three beads, and another is near the rim at the r. lower corner. Between the base of the Swastika and the back of the Elephant is an isosceles triangle lying on its side and pointing L, with a cross-bar at the apex ; to the r. a structure of three arches.

R. A single flat rim. There are three symbols in the upper half of this face and one in the lower hall In the middle of the upper half the Swastika as before, of broad lines, turned r.; near its r. upper corner three beads arranged in a triangle. To L, an indistinct symbol. To r., an Aum monogram of two triangles meeting at their apices, with a cross-bar there and a shorter one projecting on r. of lower triangle. In the middle of the lower half a structure of three arches on each side of which are three beads arranged in a triangle.

54. A roughly circular coin, 1*47 inches in diameter; weight 223 grains. It was found on the bank of the Malwatta-oya at Anuradhapura.

0. Two raised circular bands enclose the design, with a third between them broken into three beads near the top and on the 1, side. In the middle, at the base, a tusk elephantTHE LARGE CIRCULAR COINS 505

with raised trunk and extended tail which branches into three at the end. Below its mouth are three beads arranged in a triangle. Above its tail a structure of three arches under the base line of which is a bead. Above the elephant's back is the isosceles triangle, pointing 1., with an upright cross-bar below its apex. To 1. of this a tree fenced by or standing on an enclosed cross as before, with opposed branches. There are no symbols on the corners of the enclosure. To 1. of this, near the border, three beads arranged in a triangle. Above the arched structure and the triangle the raised Swastika turned r., with one bead near r. end of its base, and three arranged triangularly between its upper part and the top of the tree. Eleven beads in all.

R. opposed to O. Emblems larger and formed with bolder lines. In the middle, at the bottom, the three-arched structure, below the base line of which is a straight raised line. To r., three beads arranged triangularly. Above the arches the raised Swastika, turned r., with three beads on each side of the upper part. In the space to 1. of its basal uprights the Aum monogram. To r. of Swastika and arches, a symbol, part of which only is visible, consisting of a circular band with central bead. Mr. Still has pointed out that when seen in its complete form on other coins this is a trisiila resting on a disk or bead.1

65. A roughly circular coin, 1-27 inches by 1-31 inches; weight 264 grains. Found at Mihintale.

O. One circular band encloses the design; in one part an outer one is visible. Designs are like No. 64, but elephant's tail has only one end. The beads below its head are absent, but there are two to r. of arches and two to 1. of the fence, a total of ten.

R. Opposed to O., and indistinct on r. The design re-

1 Journal R.A.S., Ceylon, 1907, p. 20iff. Mr. Still stated that the weights of twenty examples varied from 197 to 275 grains, the average being 242*75 grains. He considered that they represent a double copper kahapana of 288 grains. All the specimens had the same symbols, arranged In the same manner, on the two faces, the only variation being the transposition of the double-triangle or Aum monogram and the Trisiila on the reverse of a few coins.506 ANCIENT CEYLON

sembles No. 54. Three beads to 1. of arches, instead of r., and one above them.

Although the Elephant, the Tree, and the Structure with three arches might be thought to be connected with Buddhism, it is extremely doubtful if they have such a signification on these coins. All three emblems occur on the Puranas, which date from an age anterior to Buddhism. They may have been merely copied from the earlier coinage, seeing that there is not another exclusively Buddhist emblem on either the earlier or later coinage of Ceylon. The probability of such borrowing of the symbols will appear more evident after the following remarks have been read.

The isosceles triangle appears on several early Indian coins reproduced in Cunningham's Coins of Ancient India, especially those of Eran, where in two instances it is elevated on a pole at the base of which in one case there is a cross enclosed in a square (Plate XI). It is also found on a coin of Ujjain (No. 14, Plate X), where Cunningham calls it a' sun-standard * ; on a Yaudeya coin (No. 5, Plate VI) ; and on several Kuninda coins in Plate V. These examples show that in its correct position the apex of the triangle is at the bottom.

I suggest that the middle cross-bar, which is sometimes on one side of the triangle and sometimes on the other, indicates that it symbolises a sistrum, an identification that is strongly supported by the form illustrated in Plate XXXIX, Fig. No. 14, of General Maisey's Sanchi and its Remains, in which the side bar ends in a hook. The sistrum is not found in the carvings in Ceylon. This instrument is clearly and unmistakably pour-tray ed on an oblong cast coin which Dr. J. R. Henderson of Madras was good enough to forward for my examination. It was found in the bed of the Vaigei river at Madura, and has the elephant in high relief on the obverse, with the sistrum and several other symbols, such as the vase, trisula, crescent, and double trident in a line near the upper edge, The sistrum is a well-known demon-frightener, and therefore would increase the protective power of the coins on which it occurs./ THE SYMBOLS 507

The line below the arched structure may represent a snake, as a guardian deity.

It would appear that at the time when these large circular coins were issued the same confidence was still reposed in the protective powers of the emblems. The sistrum, if it is one, takes its place among them for the first time in Ceylon. We still find the same raised Swastika symbol repeated exactly as in the oblong coins, a proof of the firm belief in its luck-bringing virtues. It is strange that it is now unknown in the island; it is perhaps impossible to meet with ten persons there who are acquainted with either the name or shape of any form of Swastika.

The meaning of the numerous beads on these coins is unknown. The five beads on the later Sinhalese coins afford no assistance in elucidating it, their own meaning being equally unknown. Probably the latter have some reference to the guardian deity at whose side they are represented; on some specimens the uppermost of the five is a lotus bud.

With regard to the structures of three or more arches commonly, when shown on the Puranas, termed * Chaityas/ that is, dagabas, I am not satisfied that this title furnishes a correct interpretation of their meaning. In fact, I can see little reason to apply this term to them. The designs with three and five arches appear to be representations of the domed roofs of buildings which originally may have been Hindu temples as in the Kosambi coin No. n of Plate V of Coins of Ancient India, where the nature of the edifice is indicated by the bull standing at its side. The character of some is also clearly expressed on several coins described on pp. 137 and 138 of the Indian Museum Catalogue, by the peacock on the summit of the central arch, which denotes that the building is a temple devoted to Skanda, or is under his protection. Figures of peacocks are still placed on the ontsides of his temples, and the bird itself and its feathers are considered to be emblems of good luck in India * and Ceylon. If in later instances in Ceylon the arched structures were intended for BuddMst

1 Crookes. Popular Religion and Folklore of Northern India, Vol. il pp. 233, 250.ANCIENT CEYLON

wiharas, it was probably as places to be avoided by evil spirits that they were delineated among other demon-frightening emblems. That such buildings sometimes had domed roofs is proved by the names Ganthakara (Bell) Wihara and Piriwena which occur in the Mahavansa.

Somewhat similar domed buildings are illustrated among the Amaravati reliefs.1 The central roof, like that of the structures on the coins, is at a much higher level than the two

L_\L

\T\

FIG. 159. Relief near Ruwanwaeli Dagaba.

lateral ones. There are also two reliefs of unknown age at Anuradhapura which show domed roofs of wiharas rising one behind another in nearly the same way, in one instance five roofs being visible (Fig. No. 159), and in the other three roofs, as on most of the coins.

On one of the Puranas from MuUeittivu the structure is evidently a temple or palace, the central arch rising from the

1 Archaeological Survey of Southern India, Vol. i, Plate XLII, Fig. 9.THE SYMBOLS 509

ground-level, with high vertical sides. A similar design occurs on Indian coins, and is illustrated by Mr. Theobald in his essay on the symbols. In the Taxila coin No. 5 of Plate III, C.A. I. (No. 34, p. 158, Ind. Mus. Cat.) one of these structures is shown with only two arches side by side; this cannot be a dagaba. In the third Purana illustrated by me the building is also not a dagaba.

An undoubted dagaba appears on the Andhra coins 41 and 42 of Plate II of Sir Walter Elliott's Coins of Southern India. It is a building of a different type, having a single dome, with .two rows of niches for lamps below it, just as they are to be seen in walls about some of these buildings in Ceylon. There must have been few dagabas, and those only small ones, in India before the middle of the third century B.C., whereas the arched symbol on the Puranas appears to be of much earlier date. It is interesting as being probably the first illustration of an Indian roofed building.

The crescent which often crowns the top of the uppermost dome of these arched edifices has not been satisfactorily explained. It is not a Buddhist emblem, and is never seen as an independent emblem on the coins or early sculptures of Ceylon, but it is on a punch-marked Purana from Mulleittivu.

In the Taxila coin No. 17, of Plate II, C.A.I. (No. 13, p. 157, Ind. Mus. Cat.) a person is paying reverence to this symbol fixed on the summit of a three-domed building, below which is a mound of seven beads, which may be a tumulus or a sacred hill. On coin No. 14 of the same Plate (No. 9, p. 157, Ind. Mus. Cat.) the worshipper has turned Ms back on the so-called * Chaitya' and its crescent finial, in order to adore a similar mound of seven beads, which in this case is evidently a sacred hill or tumulus. It is significant that the crescent is never seen on these mounds, but only on the arched buildings.

On the coins of Rudradaman (150 A.D.) and his successors, the crescent appears on the top of the three-domed building; while a symbol of the sun, a bead with six, seven, or eight rays, is on the right of the arches, and one of the moon, in the form of a crescent, on the left of them (Plate XVIII, Ind. Mus. Cat.). It is clear that the tipper crescent in this and other5io ANCIENT CEYLON

instances has some symbolical meaning which the lower one, treated simply as the partner of the sun, does not express.

What this is, may be learnt from the Atharva Veda (ix, 6), which mentions ' Soma, the God who is called Chandramas * [the moon]. Soma is still one of the synonyms meaning the moon. The Rig Veda is in agreement with this, and also refers to Soma as the moon. It says of Soma,c He follows the Wide-strider's [the sun's] rapid movement. ... He with the sharpened horns brings forth abundance ; the Silvery shines by night, by day the Golden ' (ix, 79, 9). Soma is also referred to as ' Subduing our assailants, chasing the demons hard to be encountered' (ix, no, 12). We also expressly learn of him that * The mighty takes his seat, and Soma, ever watchful, guards from fiend and evil sprite. Gold-hued he makes the cloud his diadem, the milk his carpet in both worlds, and prayer his robe of state' (ix, 71, i). One hymn which is addressed to Soma ends with the words ' Those awful weapons that thou hast, sharpened at point to strike men down?guard us therewith from every foe* (ix, 61, 30). Soma is also identified with the great demon-slayer Indra:?'India's self is Pava-mana [Sdma], yea, the BulP (ix, 5, 7). * Indu [Soma] is Indra' (ix, 5, 9).

It is most probably in this aspect, as Sdma, the * ever-watchful' protector from demoniacal interference, that the crescent is so often placed on the arched buildings represented on the coins, whether they are temples or palaces.

Thus it is seen that in the case of most of the early coins of the East, with its elaborate symbolism (excluding those which were mere imitations of Greek models) care was taken to insert on them emblems, or figures of deities, which were believed to have protective powers against evil spirits, as well as others that were thought to be especially luck-bringing.

This may furnish the explanation of the other strange punch-marks of the Puranas, the early signification of many of /which is known, while that of some is difficult to understand. For instance, there can be little doubt as to the purpose of the following figures on the coins.

The Elephant is at once recognised both as the * Vahana/THE SYMBOLS 511

? or riding-animal, of Indra, a persistent enemy of the demons, and as a lucky emblem.1 It is also the Vahana of Ayiyanar, who in India protects villages from nocturnal spirits. The dream of a white elephant was the omen of the birth of the Buddha, Gotama ; and in Ceylon it is still thought to prognosticate the birth of a son, which in India is one of the most fortunate of all occurrences. Miniature elephants of ivory are still sold largely in Ceylon as lucky charms. As I have already mentioned, the elephants' heads projecting from the walling and wahalkadas at the Anuradhapura dagabas were most probably placed there as a protection against evil spiritual influences, and not as mere ornaments.

Indra, the * terrific wielder of the ancient thunder ' 2 which was a favourite weapon of the Gods in their wars with evil spirits, whether Titans or others ; and Agni, ' the master of all wealth '; and the Sun?all, according to the Vedas, noted slayers of demons, and those who practised evil magic ?were all, but especially the first one, termed * Bulls ' in Vedic times, perhaps because of their irresistible power, which the Bull also symbolised in the Euphrates valley and Egypt. This animal afterwards became the Vahana of Siva, who through his Sakti, or female manifestation, slew the demons called Asuras.

The ' Taurine ' symbol, which is in the form of the skull of a bull, perhaps also signifies these Bulls. Such skulls are everywhere employed in Ceylon as potent guards against the Evil Eye, that bug-bear of all people, and the Bull's head or skull was an amulet in Egypt from prehistoric times, as also in early Greece. It is extremely doubtful if this design has, as some have supposed, an astrological signification ; when placed on the corners of the fence or enclosure at the tree its position proves that it was thought to be an additional protection.

The Sun was the luminary whose rays, shown on most of the early Indian coins as straight lines or arrow-heads radiating

1 The Jat^ka * (Translation), Vol. vi, p. 251.

s Rig Veda, iv, 20, 6 ; in x, 92, S, it is stated of him s Unhindered, from the air's vault thunders day by day the loud triumphant breathing of the fearful Bull.*5i2 ANCIENT CEYLON

outwards from a central circle,?the ' arrowy beams' (ix, 76, 4), or the ' long loose locks ' (x, 136, i) of the Rig Veda? every day discomfited the demons, and dispelled the darkness under cover of which they exerted their powers. Naturally, according to this view of the purpose of the emblems, this is the commonest of all designs on the earliest money. The circle with many internal radial lines proceeding from a central ring represents its single wheel, with which Indra destroyed the Asuras.1

The Sacred Tree, of whatever kind, owes its position to its guardian properties against demons ; and according to the Atharva Veda amulets against them were made from many different species. From some kinds the wood of the fire-drill was taken, by means of which the presence of Agni, the Fire Deity, the chief demon-slayer of Vedic times, was secured. In the Atharva Veda, the Bo-tree, the Pipal of India, is called * the Seat of the Gods/ * and thus was a place to be avoided by the demons. It will be seen, therefore, that the Bo-tree or a Bo-branch is not necessarily a Buddhist emblem when it appears on these early coins.

The defensive value of the Cross is explained in a later chapter on the Swastika.

The Snake is a well-known protector against demons. In the Rig Veda (vii, 104, 9) Sdma is prayed to hand over the evil demons to the Serpent. In the Atharva Veda (xii, 3, 55-60) Serpents are mentioned as Guardians of the Four Quarters and the Zenith. Representations of five-headed or seven-headed Cobras carved in high relief are placed at the sides of some of the dagabas at Anuradhapura and elsewhere, as guardians of the relics deposited in them. Similar carvings are also fixed as defenders at the outlets of the sluices and sometimes on the embankments, at the larger reservoirs in Ceylon. In the manuscript which I possess, containing magical formulae and diagrams, the Snake is included as a protector against illness caused by demons (Yakshas). The Snake is also everywhere believed to guard hidden treasures, and even to be

1 Rig Veda, i, 130; g ; iv, 30, 4.

z Sacred Books of the East, Vol. xlii, p. 4.THE SYMBOLS 513

a manifestation of the household guardian spirit. In China it is an emblem of the God who controls thunderstorms, rain, wind, and fire, all powerful weapons against demons, and used by Buddha against the Yakshas of Ceylon. In China its figure is also employed as a charm against evil influences.1

The Dog is also a demon-frightener. According to Sinhalese beliefs he howls at night when he sees them, and in the jungle dialect he is called Jidura, ' the demon-expeller/ In some parts of India he is a sacred animal; and he still protects the household from evil spirits.2 In the Atharva Veda (xlii, 13) the Sun is termed the ' Heavenly Dog/ probably because he was constantly acting as a guardian against the demons, A * Heavenly Dog ' is an evil deity in China ; but other dogs are worshipped as beneficent deities, while a dog's head drawn upon yellow paper is a protective charm.3

It is among the Aryans of Persia that we find the most decided evidence of the power of the Dog over demons. In the Vendidad (Fargard xiii) Ahura-Mazda says of one sp