HISTORY

OF

INDIAN ARCHITECTURE

PLATE X.

OLD DRAVIDIAN TEMPLE AT MAM ALLAPUKAM (page 362).

[Frontispiece to Volume I.

HISTORY OF INDIAN AND
EASTERN ARCHITECTURE

BY THE LATE JAMES FERGUSSON

C.I.E., D.C.L., LL.D., F.R.S., F.R.I. B.A.

Member of the Society of Dilettanti, etc., etc.
REVISED AND EDITED, WITH ADDITIONS

INDIAN ARCHITECTURE

BY JAMES BURGESS, C.I.E., LL.D., F.R.?

Hon. A.R.I. K. A.; Hon. Member of the Imperial Russian Archaeological Society ^ponding

Member Batavian Society ; Late Director of the Archaeological Survey of India, etc., etc.

AND

EASTERN ARCHITECTURE

BY R. PHENE SPIERS, F.S.A., F.R.I.B.A.

Honorary Member of the American Institute of Architects ; Correspondent of the Institute of France

WITH NUMEROUS ILLUSTRATIONS

VOL. I.

Nfl

\1to

v-l

9«537l

AUTHOR'S PREFACE TO THE FIRST
EDITION.

DURING the nine years that have elapsed since I last wrote
on this subject,1 very considerable progress has been made in
the elucidation of many of the problems that still perplex the
student of the History of Indian Architecture. The publication
of the five volumes of General Cunningham's ' Archaeological
Reports' has thrown new light on many obscure points, but
generally from an archaeological rather than from an architectural
point of view ; and Mr Burgess's researches among the western
caves and the structural temples of the Bombay presidency
have added greatly not only to our stores of information, but
to the precision of our knowledge regarding them.

For the purpose of such a work as this, however, photo-
graphy has probably done more than anything that has been
written. There are now very few buildings in India — of any
importance at least — which have not been photographed with
more or less completeness ; and for purposes of comparison
such collections of photographs as are now available are simply
invaluable. For detecting similarities, or distinguishing differ-
ences between specimens situated at distances from one
another, photographs are almost equal to actual personal
inspection, and, when sufficiently numerous, afford a picture
of Indian art of the utmost importance to any one attempting
to describe it.

1 'History of Architecture in all Countries.' 2nd ed. Murray, 1867. [Now
'History of Ancient and Medieval Architecture.' 3rd ed. 2 vols. Murray, 1893.]

viii PREFACE.

These new aids, added to our previous stock of knowledge,
are probably sufficient to justify us in treating the architecture
of India Proper in the quasi-exhaustive manner in which it
is attempted, in the first 600 pages of this work. Its description
might, of course, be easily extended even beyond these limits,
but without plans and more accurate architectural details than
we at present possess, any such additions would practically
contribute very little that was valuable to the information
the work already contains.

The case is different when we turn to Further India. Instead
of only 100 pages and 50 illustrations, both these figures ought
at least to be doubled to bring that branch of the subject up
to the same stage of completeness as that describing the archi-
tecture of India Proper. For this, however, the materials do
not at present exist. Of Japan we know almost nothing except
from photographs, without plans, dimensions, or dates ; and,
except as regards Pekin and the Treaty Ports, we know almost
as little of China. We know a great deal about one or two
buildings in Cambodia and Java, but our information regarding
all the rest is so fragmentary and incomplete, that it is hardly
available for the purposes of a general history, and the same
may be said of Burma and Sia'm. Ten years hence this
deficiency may be supplied, and it may then be possible to bring
the whole into harmony. At present a slight sketch indicating
the relative position of each, and their relation to the styles
of India Proper, is all that can be well accomplished.

Although appearing as the third volume of the second edition
of the ' General History of Architecture,' the present may be
considered as an independent and original work. In the last
edition the Indian chapters extended only to about 300 pages,
with zoo illustrations,1 and though most of the woodcuts
reappear in the present volume, more than half the original
text has been cancelled, and consequently at least 600 pages
of the present work are original matter, and 200 illustrations —

' History of Architecture ' (1867), vol. ii. pp. 445-756, Woodcuts 966-1163.

PREFACE. ix

and these by far the most important — have been added. These,
with the new chronological and topographical details, present
the subject to the English reader, in a more compact and
complete form than has been attempted in any work on Indian
architecture hitherto published. It does not, as I feel only
too keenly, contain all the information that could be desired,
but I am afraid it contains nearly all that the materials at
present available will admit of being utilised, in a general
history of the style.

When I published my first work on Indian architecture
thirty years ago, I was reproached for making dogmatic asser-
tions, and propounding theories which I did not even attempt
to sustain. The defect was, I am afraid, inevitable. My
conclusions were based upon the examination of the actual
buildings throughout the three Presidencies of India and in
China during ten years' residence in the East, and to have
placed before the world the multitudinous details which were
the ground of my generalisations, would have required an
additional amount of description and engravings which was
not warranted by the interest felt in the subject at that time.
The numerous engravings in the present volume, the extended
letterpress, and the references to works of later labourers in
the wide domain of Indian architecture, will greatly diminish,
but cannot entirely remove, the old objection. No man can
direct his mind for forty years to the earnest investigation
of any department of knowledge, and not become acquainted
with a host of particulars, and acquire a species of insight which
neither time, nor space, nor perhaps the resources of language
will permit him to reproduce in their fulness. I possess, to
give a single instance, more than 3,000 photographs of Indian
buildings, with which constant use has made me as familiar
as with any other object that is perpetually before my eyes,
and to recapitulate all the information they convey to long-
continued scrutiny, would be an endless, if not indeed an
impossible undertaking. The necessities of the case demand

x PREFACE.

that broad results should often be given when the evidence
for the statements must be merely indicated or greatly abridged,
and if the conclusions sometimes go beyond the appended
proofs, I can only ask my readers to believe that the assertions
are not speculative fancies, but deductions from facts. My
endeavour from the first has been to present a distinct view
of the general principles which have governed the historical
development of Indian architecture, and my hope is that those
who pursue the subject beyond the pages of the present work,
will find that the principles I have enunciated will reduce to
order the multifarious details, and that the details in turn will
confirm the principles. Though the vast amount of fresh
knowledge which has gone on accumulating since I commenced
my investigations has enabled me to correct, modify, and
enlarge my views, yet the classification I adopted, and the
historical sequences I pointed out thirty years since, have in
their essential outlines been confirmed, and will continue, I
trust, to stand good. Many subsidiary questions remain
unsettled, but my impression is, that not a few of the dis-
cordant opinions that may be observed arise principally from
the different courses which enquirers have pursued in their
investigations. Some men of great eminence and learning,
more conversant with books than buildings, have naturally
drawn their knowledge and inferences from written authorities,
none of which are contemporaneous with the events they relate,
and all of which have been avowedly altered and falsified in
later times. My authorities, on the contrary, have been mainly
the imperishable records in the rocks, or on sculptures and
carvings, which necessarily represented at the time the faith
and feelings of those who executed them, and which retain
their original impress to this day. In such a country as India,
the chisels of her sculptors are, so far as I can judge,
immeasurably more to be trusted than the pens of her authors.
These secondary points, however, may well await the solution
which time and further study will doubtless supply. In the
meanwhile, I shall have realised a long-cherished dream if I

PREFACE.

XI

have succeeded in popularising the subject by rendering its
principles generally intelligible, and can thus give an impulse
to its study, and assist in establishing Indian architecture on
a stable basis, so that it may take its true position among
the other great styles which have ennobled the arts of
mankind.

The publication of this volume completes the history of
the 'Architecture in all countries, from the earliest times to
the present day, in four volumes,' and there it must at present
rest. As originally projected, it was intended to have added
a fifth volume on ' Rude Stone Monuments/ which is still
wanted to make the series quite complete ; but, as explained
in the preface to my work bearing that title, the subject
was not, when it was written, ripe for a historical treatment,
and the materials collected were consequently used in an
argumentative essay. Since that work was published, in 1872,
no serious examination of its arguments has been undertaken
by any competent authority, while every new fact that has
come to light — especially in India — has served to confirm me
more and more in the correctness of the principles I then
tried to establish.1 Unless, however, the matter is taken up
seriously, and re-examined by those who, from their position,
have the ear of the public in these matters, no such progress
will be made as would justify the publication of a second
work on the same subject. I consequently see no chance of my
ever having an opportunity of taking up the subject again,
so as to be able to describe its objects in a more consecutive
or more exhaustive manner than was done in the work just
alluded to.

1 A distinguished German professor,
Herr Kinkel of Zurich, in his ' Mosaik
zur Kunstgeschichte, Berlin, 1876,' has
lately adopted my views with regard to

the age of Stonehenge without any
reservation, though arriving at that
conclusion by a very different chain of
reasoning from that I was led to adopt.

PREFACE TO SECOND EDITION

THE late Mr. Fergusson's ' History of Indian and Eastern
Architecture' has now been before the public for more than
thirty years, and was reprinted (without his consent) in America,
before his death in 1886, and the publishers issued a reprint
in 1891. His method of treating the subject he has thus
described : — " What I have attempted to do during the last
forty years has been to apply to Indian Architecture the same
principles of archaeological science which are universally adopted
not only in England, but in every country in Europe. Since
the publication of Rickman's ' Attempt to discriminate the
Styles of Architecture in England' in 1817, style has been
allowed to supersede all other evidences for the age of any
building, not only in Mediaeval, but in Byzantine, Classical,
and, in fact, all other true styles. Any accomplished antiquary,
looking at any archway or any moulding, can say at once,
this is Norman, or Early English, or Decorated, or Tudor ;
and if familiar with the style, tell the date within a few years,
whether it belongs to a cathedral or a parish church, a dwelling
house or a grange, ... is not of the smallest consequence,
nor whether it belongs to the marvellously elaborate quasi-
Byzantine style of the age of the Conqueror, or to the prosaic
tameness of that of the age of Elizabeth. Owing to its perfect
originality and freedom from all foreign admixture or influence,
I believe these principles, so universally adopted in this country,
are even more applicable to the Indian styles than to the
European."

The successful application of these principles to Indian
architecture was entirely his own : no one had dreamed of it

xii

PREFACE TO SECOND EDITION. xiii

before. It was a stroke of genius to trace out logically the
historical sequences of the Hindu monuments and make them
tell their own story by means of those guiding principles which
he was the first to apply to them, and to elucidate their
applicability in a manner that has been borne out since with-
out exception wherever they have been intelligently applied.
Though descriptions of Indian monuments may be written in
various ways, no one could pretend to take up the systematic
study of Indian Architecture without the aid of this work,
and no history of the architecture can be scientifically written
without appropriating the principles Mr. Fergusson showed how
to apply.

My close intimacy with Mr. Fergusson for twenty years,
and knowledge of his opinions, may have suggested that I
might undertake the revisal of his work ; but, when it was
first proposed, I was engaged on the preparation of certain
volumes of the Archaeological Reports of the Indian Survey
that had been entrusted to me and I could not then undertake it.
On the appointment of a new director for the Surveys, at the
close of 1901, the materials were taken out of my hands and
my engagement terminated. I was then at liberty to undertake
the revision of the work, and in doing so I naturally depended
on the like help that had been afforded to Mr. Fergusson him-
self in 1875, when the resources of the Surveys were at his
disposal. But obstruction was raised where it ought hardly
to have been expected, and it was due to the good offices of the
Right Honourable Lord Morley, Secretary of State for India,
that this was largely overcome. The materials in the India
Office were at once liberally placed at my disposal, and the
Government of India requested to favour the work. This,
however, caused delay, and subsequent severe illness has pro-
tracted the preparation of the work.

It would have been easy to expand this history, but, if it
was to answer its purpose as a handbook, it must obviously
be restricted within moderate dimensions. My aim has been
to condense where practicable and, whilst revising, to make

xiv PREFACE TO SECOND EDITION.

only such additions from the accessible materials accumulated
since 1876, as seemed requisite. The Archaeological Surveys
have collected vast stores of drawings only a fraction of which
has yet been published. Travellers too, influenced partly perhaps
by the interest that Mr. Fergusson's volume had created, have
published works that have added to our information.

The great advances made in Indian Epigraphy and Palaeo-
graphy during the same period have further enabled us to revise
and fix more accurately the dates in the earlier chronology of
India ; but this has not materially affected the author's chrono-
metric scale of arrangement of the monuments, for where the
dates have been somewhat altered, the relative places of the
monuments have not required to be changed, — only they have
been better adjusted ; and in many cases Mr. Fergusson, in his
later years, had accepted these corrections.

For much valued aid and information my thanks are due
to Mr. Henry Cousens, Superintendent of the Western India
Archaeological Circle ; and to Mr. Alexander Rea of the Madras
Circle, from both of whom I have received ungrudging assist-
ance, relative to the districts under their charge.

For Ceylon I am greatly indebted to Lord Stanmore
and the Colonial Office, whilst Mr. J. G. Smither, late Govern-
ment architect, and Mr. H. C. P. Bell, C.C.S., Archaeological
Commissioner, very kindly have read the proofs and supplied
important advice and material for the chapter on the archi-
tecture of the Island.

I owe thanks also for valued help to Babu Monmohan
Chakravarti, M.A., relative to Orissa ; and among others to
Mr. R. F. Chisholm, F.R.I.B.A. ; Mr. H. C. Fanshawe, C.S.I,;
Dr. J. F. Fleet, C.I.E.; Professor Dr. H. Kern, Utrecht; the
Right Honble. Ameer 'AH ; Mr. G. F. Williams, State Engineer,
Udaypur ; Lieut. Fred. M. Bailey, Indian Army ; Mr. F. H.
Andrews ; Dr. L. D. Barnett, and the Rev. Dr. Wm. Millar,
C.I.E. To Messrs. Kegan Paul, Trench, Triibner and Co., I
am indebted for the use of a number of woodcuts.

The history of Indian Architecture has been extended from

PREFACE TO SECOND EDITION. xv

610 to 785 pages, and the illustrations in the text increased
by 98, besides the addition of 34 plates from photographs.

The chapters on Further India, Java and China have been
edited and partly rewritten by Mr. R. Phene Spiers, the editor
of Mr. Fergusson's larger work, the ' History of Ancient and
Medieval Architecture,' published in 1893. Mr. Spiers has
recast these chapters, adding much fresh and important in-
formation to each, whilst he has also added a new chapter
on the Architecture of Japan. For Burma, Mr. Spiers has
had to depend largely upon the few works published during
the last thirty years describing the buildings there found, on
the photographs in the India Office and on the somewhat
meagre notes contained in the ' Progress Reports ' of the
Archaeological Survey.

For Cambodia, Siam and Java, on the other hand, were
available the excellent publications of the French Archaeological
Surveys carried out at first under the supervision of the Ecole
Fran9aise d'Extreme Orient, and now under the skilled direction
of the Archaeological Commission of Indo - China, and of the
Java Surveys under the direction of the Dutch Government
Archaeological Commission.

This section occupied 100 pages with 49 woodcuts in the
former edition ; now, with the addition of Japan, it has been
extended to 163 pages, with 67 woodcuts and 31 plates.

J. BURGESS.
EDINBURGH,

February 1910.

NOTE.

ONE of the great difficulties that met every one attempting to write on Indian subjects
forty years ago was to know how to spell Indian proper names. The Gilchristian
mode of using double vowels, which was fashionable early last century, had then
been done away with, as contrary to the spirit of Indian orthography, though it left
a plentiful crop of discordant spellings. On the other hand, Sir William Jones and
most scholars, by marking the long vowels and by dots to distinguish the palatal
from the dental consonants, had formed from the Roman alphabet definite equivalents
for each letter in the Indian alphabets — both Sanskritic and Persian. Lepsius,
Lassen, and Max Muller in turn proposed various other systems, which ha^e not
found much acceptance ; and of late continental scholars have put forward still
another scheme, quite unsuited for English use. In this system such names as
" Krishna," " Chach," " Rishi," are to be represented byKrsna, Cac, Rsi — so pedantic
a system is, impossible both for cartographer and ordinary reader and, like others,
it may well cease to be.

Meanwhile a notable advance towards official uniformity has been made in the
spelling of Indian place-names. When the 'Imperial Gazetteer of India" was
projected, Government judiciously instructed the editor to adopt the Jonesian system
of transliteration as slightly modified by Professor H. H. Wilson, but devoid of
the diacritical dots attached to certain consonants. The authorisation of this
system in the new maps and Gazetteer, and its use in published works since,
has established its claim to acceptance in a work intended for the general reader.

In the following pages, consequently, this system has been used, as nearly as
may be, avoiding diacritical marks on consonants, but indicating the long vowel
sounds ti, i, ft, as in Lzit, Halebid, Stupa, etc., whilst e and o, being almost always
long, hardly require indication.

Thus a l sounds as in " rural " ; a as in " tar " ;

i ,, ,, "fill"; i ,, "police";

« ,, ,, "full"; ft ,, "rude";

e ,, ,, " there"; and o ,, "stone".

Only the palatal 5, as in "sure," is distinguished from the dental, as in "hiss," by
the italic form among Roman letters, as in "jikhara," "Aroka." A hundred years
hence, when Sanskrit and Indian alphabets are taught in all schools in England,
it may be otherwise, but in the present state of knowledge on the subject it seems
expedient to use some such simple method of indicating, at least approximately, the
Indian sounds. Strictly accurate transcription in all cases and of well-known names,
however, has not been followed.

In Burmese,— which lisps sounds like s and ch, — the spellings used in the Gazetteers
of Burma have been generally adopted.

1 The shut vowel, inherent in all consonants of the proper Indian alphabets, was formerly
transliterated by almost any English vowel : in " Benares" (for " Banaras"). t is used twice for it.

CONTENTS.

INTRODUCTION

• Page 3

BOOK I.

BUDDHIST ARCHITECTURE.

CHAP. PAGE

I. INTRODUCTION AND CLASSIFI-
CATION . . . " . 51
II. STAMBHAS OR LATS . . 56

III. STUPAS — Relic Worship —

Bhilsa Topes — Topes at Sar-
nath and in Bihar — Amaravati
Stupa — Gandhara Topes —
Jalalabad Topes — Manikyala
Stupa 62

IV. RAILS— Rails at Bharaut, Math-

ura1, Sanchi, and Amaravati . 102
V. CHAITYA HALLS — Structural
Chaityas — Bihar Caves —
Western Chaitya Halls, etc. . 125

CHAI'. PAGE

VI. VIHARAS, OR MONASTERIES —
Structural Viharas — Bengal
Caves — Western Vihara Caves
— Nasik, Ajanta, Bagh, Dham-
nar and Kholvi, Elvira, Aurang-
abad and Kuda Viharas . 170

VII. GANDHARA MONASTERIES —
Monasteries at Jamalgarhi —
Takht-i-Bahaiand Shah-Dheri
— Greek influence '. . 209

VIII. CEYLON — Introductory —
Anuradhapura — Polonnaruwa

224

BOOK II.

ARCHITECTURE IN THE HIMALAYAS.

I. KASHMIR — Temples — M^rtand
— Avantipur — Buniar — Pan-
drethan — Malot

251

II. NEPAL AND TIBET — Stupas or
Chaityas — Wooden Temples
— Tibet — Temples in KangrS

XVll

273

XV111

CONTENTS

BOOK III.
DRAVIDIAN STYLE.

CHAP. PAGE

I. INTRODUCTORY . . .302
II. HINDI) CONSTRUCTION —
Arches — Domes — Plans —
.Sikharas .... 310
III. DRAVIDIAN ROCK-CUT TEMPLES
— Mamallapuram — Kailas,

Elura 327

IV. DRAVIDIAN TEMPLES — Patta-
dakal and Dharwar Temples

CHAP.

PAGE

— Conjivaram — Tanjor — Tiru-
valiir — 5rirangam — Chidamba-
ram— Ramervaram— Madurl —
Tinnevelly — Kumbakonam —
Vellor and Perur — Vijayanagar 350
V. CIVIL ARCHITECTURE — Palaces
at Madura and Tanjor —
Garden Pavilion at Vijaya-
nagar— Palace at Chandragiri . 411

BOOK IV.

CHALUKYAN STYLE.

I. INTRODUCTORY — Chalukyan
Architecture — Dharwar tem-
ples — Ittagi — Gadag — Kur u -
vatti — Dambal — Hanamkonda

— Kirtti-StambhasatWorangal
— Mysore — Temples at Som-
nathpur and Beliir— Temples
at Halebid .... 420

DIRECTIONS TO BINDER

MAP OF INDIA, SHOWING THE PRINCIPAL BUDDHIST LOCALITIES To face page 51

MAP OF THE PRINCIPAL INDO- ARYAN, CHALUKYAN, AND

DRAVIDIAN LOCALITIES ,, 251

LIST OF ILLUSTRATIONS IN VOLUME I.

No.

1. Hindu Temple, at Bahulara, near

Bankura ....

2. Naga people worshipping the

Triratna emblem of Buddha,
on a fiery pillar

3. Sri or Gaja Lakshmi seated on a

Lotus, with two Elephants
pouring water over her .

4. Lat at Allahabad

5. Assyrian honeysuckle ornament

from capital of Lat, at Allaha-
bad

6. Capital at Sankisa

7. Capital of Lat in Tirhut

8. Capital of the Lion-pillar at Karle

9. Minar Chakri, Kabul

10. Relic Casket of Moggalana

11. Relic Casket of .Sariputra .

12. View of the great Tope at Sanchi

13. Plan of great Tope at Sanchi

14. Section of great Tope at Sanchi

15. ' Tee ' (Hti) cut in the rock on a

Dagaba at Ajanta .

1 6. Tope at Sarnath, near Benares .

17. Panel on the Tope at Sarnath

18. View and -Plan of Jarasandha-ka-

baithak ....

19. Temple at Bodh-Gaya with Bo-

tree .....

20. Representation of a Stupa from

the Rail at Amaravati .

21. Tope at Bimaran

22. Tope at Sultanpur

23. Stupa at Chakpat

24. Relic Casket from a Tope at

Manikyala ....

25. View of Manikyala Tope .

26. Restored Elevation of the Tope

at Manikyala

PAGE , No.

27. I

'5

28. I

49 29. I

50 30. 1

57 3I- Y:

32. t

57

33- 1

58

58

34- F

60

35- *

61

36. F

68
68

37- I

69

38. IS

69

69 39. E

70 40. C

72

74 41- E

76 42. A

i 43. S

78

44- E

81

45- T

91

46. T

92

47- P
48. A

95

49- P

96

50. P

97

PAGE
Elevation and Section of portion

of Basement of Stupa at Mani-
kyala 98

Relic Casket, Manikyala . . 99
Parinirvana of Buddha, from

Cave No. 26 at Ajanta . . 101
Tree Worship : Bodh-Gaya Rail 105
Relic Casket : Bodh-Gaya Rail 105
Portion of Rail at Bharaut, as

first uncovered . . .106
Tree and Serpent Worship at

Bharaut .... 108
Rail at Sanchi . . .- . 1 1 1
Rail, of No. 2 Tope, Sanchi . 112
Representation of Rail, from

Amaravati . . . .112
ail in Gautamiputra Cave,

Nasik 113

Northern Gateway of Tope at

Sanchi . . . . 115

Bas-relief on left-hand Pillar,

Northern Gateway, Sanchi . 117
Ornament on right-hand Pillar,

Northern Gateway, Sanchi . 117
External Elevation of Great Rail

at Amaravati . . 1 20

Angle pillar at Amaravati . .121
Slab from Base of the Stupa,

Amaravati . . . .121
Dagaba (from a Slab), Amaravati 122
Triratna Emblem. (From a

sculpture at Amaravati) . .124
Triratna Symbol from Sanchi . 124
Plan of Chaitya Hall, Sanchi . 126
Ancient Buddhist Chaitya at Ter 126
Plan of Ancient Buddhist Chaitya

at Ter . . . . .126

Plan of an Ancient Chaitya at
Chezarla . . . .127

XX

LIST OF ILLUSTRATIONS IN VOLUME I.

No.

Si-

52.

S3-

54-
55-
56.
57-

58.

59-
60.

61.
62.

63
64.

65-

66.
67.
68.
69.
70.
71-

72.
73-

74-

75-
76.

77-
78.

79-
80.
81.

82.

PAGE
Elevation of Chezarla Chaitya

Temple .... 127
Section of Chezarla Chaitya

Temple . . . .127
Sudama Cave, Plan and Section,

Sat-Garbha group . . .130
Kondivte Cave, Salsette . . 131
Facade of Lomas Rishi Cave . 131
Lomas Rishi Cave, Plan . .132
Plan and Section of Sita-marhi

Cave 133

Chaitya and Vihara Caves, at

Bhaja 134

Front of a Chaitya Hall . .134
Fa9ade of the Chaitya Cave at

Bhaja 135

Triratna. Shield. Chakra. Tri-

ratna, etc 136

Capital of a rock-cut Dagaba at

Bhaja 137

Plan of Caves at Bedsa . . 1 38
Capital of Pillar in front of Cave

at Bedsa . . . 139

View in Verandah of Chaitya at

Bedsa .

140
141
143
143
144
146

View of Chaitya Cave at Nasik .
Section of Chaitya Cave at KSrle
Plan of Chaitya Cave at Karle .
View of Chaitya Cave at Karle .
View of Interior of Cave at K^rle
Interior of Chaitya Cave No. 10

at Ajanta . . . .149
Cross-section of Cave No. 10 at

Ajanrl 149

Plan of Chaitya Cave No. 19 at

Ajanta 151

View of Facade, Chaitya Cave

No. 19 at Ajanta . . .152
Rock-cut Dagaba at Ajanta . 153
Small Model found in the Tope

at Sultanpur . . . ' • 153
Pillars on the left sideof the Nave,

in Cave No. 26 at Ajanta . 154
View of Manmoda Chaitya Cave

at Junnar .... 157

Plan of Circular Cave, Junnar . 158
Section of Circular Cave, Junnar 158
Round Temple and part of
Palace or Monastery. ( From a
bas-relief at Bharaut) . 159

Interior of VLrwakarma Buddhist

Cave at Elura . . 160

No. PAGE

83. Fa$ade of the Vi^wakarma Cave

at EliirS .... 161

84. Rail in front of the Chaitya

Cave at Kanheri . . .163

85. Capital of a Pillar from the

Chaitya Cave at Kanheri . 164

86. Caves at Dhamnar . . .165

87. Fa9ade of Chaitya Cave at

Guntupalle . . . .168

88. Sphinxes from Buddhist Vihara

at Pitalkhora . . .169

89. Dharmaraja Rath at Mamalla-

puram 172

90. Diagram explanatory of the

arrangement of a Buddhist
Vihara of Four Storeys in
height 172

91. Square Cell from a bas-relief

at Bharaut . . . -173

92. Oblong Cell from a bas-relief

at Bharaut .... 173

93. Plan of Son-bhandar Caves . 176

94. Section of Son-bhandar Cave . 176

95. Front of Son-bhandar Cave . 176

96. Plan of small Vihara at Bhaja . 177

97. Capital of Pilaster at Bhaja . 178

98. Plan of Cave No. n at Ajanta. 181

99. Plan of Cave No. 2 at AjantS . 181

100. Plan of Cave No. 3 at Bagh . 182

101. Plan of Darbar Cave, Kanheri . 182

102. Plan of Nahapana Vihara, Nasik 184

103. Pillar in Nahapana Cave, Nasik 185

104. Pillar in Gautamiputra Cave,

Nasik 185

105. Plan of Sri Yajna Cave, No. 15,

at Nasik . . . .187

1 06. Pillar in Sri Yajna Cave . .188

107. Plan of Cave No. 16, at Ajanta 189

108. View of Interior of Vihara, No.

1 6, at Ajanta . . . 190

109. View in Cave No. 17, at Ajanta 191
no. Pillar in Vihara Cave No. 17 at

Ajanta ..... 192
in. Capital from Verandah of Cave

24, Ajanta . . . .194

112. Pillar in the Verandah of Cave

I, Ajanta .... 195

113. Plan of Great Vihara Cave at Bagh 198

114. Buddhist Viharas at the south

end of the Elura group . . 201

115. Plan of Maharwara Cave, Elura 202

LIST OF ILLUSTRATIONS IN VOLUME I.

xxi

No. PAGE

116. Ancient Buddhist Tower at

Negapattam .... 206

117. Ancient capital found at Patna 207

1 1 8. Capital in Side Chapel of Cave

No. 19, Ajanta . . .208

119. Plan of Monastery at Jamalgarhi 212

1 20. Plan of Monastery at Takht-i-

Bahai .... 212

121. Corinthian Capital from Jamal-

garhi . . . . .214

122. Corinthian Capital from Jamal-

garhi ..... 214

123. Conventional Elevation of the

Fa9ade of a Cell from Jamal-
garhi ..... 216

124. Plan of Ionic Monastery, Shah-

Dheri

125. Ionic Pillar, Shah-Dheri .

126. Footprints of Buddha (From a

bas-relief at Amaravati) .

127. View of the north side of west

chapel, Ruwanveli Dagaba .

128. Part Elevation (restored) of

front of south chapel, Ruwan-
veli Dagaba .

1 29. Stelce at east end of north chapel ,

Abhayagiri Dagaba

130. Thuparama Dagaba, Anuradha-

pura .....

131. Capital from outer circle at

Thuparama Dagaba

132. Lankarama Dagaba, Anuradha-

pura .....

133. Capital of Lankarama Dagaba

Pillars, inner circles

134. Pavilion with Steps west of

Ruwanveli Dagaba

135. Moonstone at the Steps of

the Bo-tree platform

136. View of the Sacred Bo-tree

vihara, Anuradhapura .

137. Sat Mahal Prasada and Galpota

138. Capital of a Pilaster, Pitalkhora
139- Tomb of Zainu-l-'Abidin. Ele-
vation of Arches .

140. Takht-i-Sulaiman — Elevation

of Arches ....

141. Model of Temple in Kashmir .

142. Pillar at Srinagar

143. Capital from Shadipur

144. Restoration of Vihara Cells, at

Takht-i-Bahai

No.
MS-
146.

147.

148.
149.
150.

IS*-

152-
153.
154.

155-

218 '57'

223

158-

231

159-

1 60.

232

161.

233

162.

163.

234

164.

235

165.

236

1 66.

236

J67

240

168.

240

169.

243

170.

246

171.

250

172.

173-

253

175-

176.

254

177.

256

257

178.

257

179.

1 80.

258 181.

PAGE

Temple of Martand, Plan . 259

View of Temple at Martand . 260
View of Central Cell of Court at

Martand . . . .261

Niche with Figure at Martand . 263

Soffit of Arch at Martand . 264

Pillar at Avantipur . . . 265
View of Court of Temple at

Buniar ..... 266

View of Temple at Pandrethan 268

View of Temple at Payer . 269
View of Temple at Malot, in the

Salt Range . . . .271
View of Temple of Swayam-

bhimath, Nepal . . . 278

Nepalese Kosthakar . . 280
View of Devi Bhawani Temple,

Bhatgaon . . . .281
View of Temples of Mahadeva

and Krishna, Patan . '. 283
Pajupati — General view of the

temples and burning ghat . 284

Doorway of Darbar, Bhatgaon 285
View of Hindu Temple at Cher-

gion in ChambS . . . 287

Monoliths at Dimapur . . 288
Doorway in the Temple at

Tashiding .... 295
Interior of Temple at Pemi-

ongchi ..... 296
View of Temples at Kiragrama,

Kingra District . . . 299
Pillar in Porch of a Vaishnava

Temple at Eran . . .301
Capital of Half Column from a

Temple in Orissa . . . 301
View of City Gateway, Vijaya-

nagar . . . . . 3 1 1

Gateway Jhinjhuwada . . 312

Radiating Arch . . . 313
Horizontal Arch . . -3^3

Diagram of Roofing . . 314

174. Diagrams of Roofing . 314

Diagram of Roofing . . 315

Diagram of Indian Construction 315
Diagram of the arrangement of

the pillars of a Jaina Dome . 317

Diagram Plan of Jaina Porch . 317

Diagram of Jaina Porch . . 318

Old Temple at Aihole, Plan . 320

View of Old Temple at Aihole . 321

XX11

LIST OF ILLUSTRATIONS IN VOLUME 1.

No.

182. Plan of Papanatha Temple

at Pattadakal

183. Restored Elevation of the Sun

Temple at Kanarak

184. Diagram Plan and Section of

the Temple at Kanarak

185. View of the Raths, Mamalla-

puram

1 86. View of Draupadi's Rath

187. Plan of Bhima's Rath

1 88. Pillar in Bhima's Rath .

189. Plan of Dharmaraja Rath

190. Elevation of Dharmaraja Rath .

191. Section of Dharmaraja Rath,

with suggested internal ar-
rangements ....

192. Plan of Sahadeva's Rath .

193. View of Sahadeva's Rath

194. View of Gane^a Rath

195. View of Perumal Temple at

Madura ....

196. Entrance to a Hindu Temple,

Colombo ....

197. Head of the Naga figure,

at Mamallapuram .

198. View of Cave Temple, Saluvan-

kuppam

199. Plan of KailSs Temple, at Elura

200. View of Kailas Elura

201. Shrine of the River Goddesses,

Elura

202. Dhwajastambha at Kailas, Elura

203. Dipdan in Dharw&r

204. Plan of Great Temple at Patta-

dakal

205. South elevation of Virupaksha

Temple at Pattadakal .

206. Plan of Sangamejvar Temple

at Pattadakal

207. Plan of Malegitti Temple,

at Badami ....

208. Plan of Meguti Jaina Temple

at Aihole ....

209. Plan of Kailasanatha Temple,

Conjivaram ....

210. Section of Vaikuntha Perumal

Temple, Conjivaram

211. Plan of the Shore Temple at

Mamallapuram

212. Diagram Plan of Tanjor Temple

PAGE

No.

213.

322

214.

323

215-

324

216.

329

217.

330

331

218.

332
333

219.

334

220.

221.

222.

335

336

223.

337

338

224.

339

225.

226.

34°

227.

341 I

. 228.

341

343

229.

344

230.

345

231.

346

232.

347

233-

234-

353 235.

354 236.

355 237.

356 238.

356 239'

240.

358

241.

359 242.

243.

36i 244.

View of the Great Temple at
Tanjor .....

View of Temple of Subrah-

manya, Tanjor . . .
Pier in Subrahmanya Temple .
Plan of Inner Temple at Tiru-

valur .....
Bird's-eye view of Temple at

Tiruvalur ....
Plan of .Srirangam Temple — the

four inner courts . . .

View of the eastern half of the
Great Temple at .Srirangam .

Plan of Temple of Chidambaram
View of Porch of Chidambaram
Section of Porch of Temple at

Chidambaram . . -
View of Ruined Temple at

Chidambaram . . .
Plan of Great Temple at Ram-

e^varam, before 1905 . -
Central Corridor, Rame^varam
Plan of Tirumalai Nayyak's

Chaultrt, Madura . . .
Pillar in Tirumalai Nayyak's

Chaultri . . . .
View in Tirumalai Nayyak's

Chaultri ....

Plan of Madura Temple .
Half-plan of Temple at Tinne-

63

Gopuram at Kumbakonam .
Portico of Temple at Vellor .
Compound Pillar at Vellor .
Compound Pillar at Periir .
Plan of Vitthalaswamin Temple

at Vijayanagar . . .
View of Porch of Temple of

Vitthalasvamin . . .
Entrance through Gopuram at

Tadpatri ....
Portion of Gopuram at Tadpatri
Plan of Temples at 6"ri-6ailam .
Plan of Tirumalai Nayyak's

Palace at Madura . . -
Hall in Palace, Madura . -
Court in Palace, Tanjor . -
Garden Pavilion at Vijayanagar
South Elevation of Chandragiri

Palace . . . . .

PAGE
364

365
366

367
367
369

371
375
376

377
378

381
383

387
387

389
391

393
395
397
399
399

402
403

405
406

409

413
414
415
417

418

LIST OF PLATES TO VOLUME I. xxiii

No. PAGE . No. PAGE

245. Plan of the ground floor of j 255. Plan of Ke.ra.va Temple at Som-

Chandragiri Palace . .418 nathpur . . . . 437

246. Plan of Ittagi Temples . . 425 256. Temple at Somnathpur . . 438

247. Plan of Temples at Kukkanur . 426 I 257. Plan of Chenna Kej-ava Temple

248. Plan of Some^var Temple, at Belur .... 439

Gadag .... 427 358. View of part of Porch at Belur 440

249. Plan of Kuruvatti Temple . 430 259. Pavilion at Belur . . .441

250. Kuruvatti Temple — south eleva-

tion ..... 430

251. Plan of Dambal Temple of

260. View of Kedaresvara Temple,

Halebid .... 443

261. Plan of Hoysale-rvara Temple

Dodda Basavanna . . . 431 j at Halebid . . . . 444

252. Doorway of Great Temple at 262. Restored view of Temple at

Hanamkonda . . . 433

253. Kirtti-Stambha at Worangal . 434

254. Temple at Buchhanapalli . 436

Halebid . . . .445

263. Central Pavilion, Halebid, East
Front

447

LIST OF PLATES TO VOL. I.

PLATE

X. OLD DRAVIDIAN TEMPLE AT MAMALLAPURAM

(page 362) Frontispiece to Vohinie 1.

JAMES FERGUSSON To face page I

I. ALI MASJID — STUPA NO 5 ,, 92

II. THUPARAMA DAGABA (AS RESTORED) .... ,, 234

III. THE WATA-DA-GE AT POLONNARUWA .... ,, 246

IV. THE THUPARAMA TEMPLE, POLONNARUWA ... ,, 247
V. SAIVA TEMPLE AT POLONNARUWA ,, 248

VI. THE POTALA AT LHASA FROM THE W.S. W. ... ,, 292

VII. GOLDEN TEMPLE AT GYAN-TSE ..... ,, 294

VIII. MALEGITT1 SAIVA TEMPLE AT BADAMI .... ,, 356

IX. VIEW OF MAIN SHRINE, KAILASANATHA TEMPLE,

KANCHfPURAM ,, 359

XI. STONE CAR, AT THE TEMPLE OF VITTHALA, VIJAYANAGAR,

1881 (NOW DESTROYED) ,, 403

XII. GREAT TEMPLE AT ITTAGI, FROM S.W ,, 424

XIII. TEMPLE OF SOMESVAR AT GADAG, FROM N.E. . . ,, 427

XIV. DOORWAY OF THE SHRINE IN KASIVISVESVAR TEMPLE

AT LAKKUNDI „ 428

XV. CHAUDADAMPUR TEMPLE OF MUKTESVARA ... ., 429

XVI. GALAGANATH TEMPLE FROM N.W ,, 432

XVII. TEMPLE OF KEDARESVARA AT BALAGAMI ... ,, 441

Buddha preaching, (From a fresco painting at Ajanta.

JAMES FERGUSSON.

[To face page I, Volume I.

HISTORY

OF

INDIAN ARCHITECTURE.

VOL. I.

HISTORY

OF

INDIAN ARCHITECTURE.

INTRODUCTION.

IT is in vain, perhaps, to expect that the literature or the Arts
of any other people can be so interesting to even the best
educated Europeans as those of their own country. Until it
is forced on their attention, few are aware how much education
does to concentrate attention within a very narrow field of
observation. We become familiar in the nursery with the
names of the heroes of Greek and Roman history. In every
school their history and their arts are taught, memorials of
their greatness meet us at every turn through life, and their
thoughts and aspirations become, as it were, part of ourselves.
So, too, with the Middle Ages : their religion is our religion ;
their architecture our architecture ; and their history fades so
insensibly into our own, that we can draw no line of demarca-
tion that would separate us from them. How different is the
state of feeling, when from this familiar home we turn to such
a country as India ! Its geography is hardly taught in schools,
and seldom mastered perfectly ; its history is a puzzle ; l its
literature a mythic dream ; its arts a quaint perplexity.
But, above all, the names of its heroes and great men are so
unfamiliar that, except a few of those who go to India, scarcely
any ever become so acquainted with them, that they call up
any memories which are either pleasing or worth dwelling upon.
Were it not for this, there is probably no country — out of
Europe at least— that would so well repay attention as India:
none, where all the problems of natural science or of art are
presented to us in so distinct and so pleasing a form. Nowhere
does nature show herself in such grand and such luxurious

1 The last thirty years have added
greatly to the number and quality of
the text-books on Indian history, and

the general reader has no longer a valid
excuse for ignorance of it.

HISTORY OF INDIAN ARCHITECTURE.

features, and nowhere does humanity exist in more varied and
more pleasing conditions. Side by side with the intellectual
Brahman caste, and the chivalrous Rajput, are found the wild
Bhil and the naked Gond, not antagonistic and warring one
against the other, as elsewhere, but living now as they have
done for thousands of years, each content with his own lot,
and prepared to follow, without repining, in the footsteps of
his forefathers.

It cannot, of course, be for one moment contended that
India ever reached the intellectual supremacy of Greece, or
the moral greatness of Rome ; but, though on a lower step of
the ladder, her arts are more original and more varied, and her
forms of civilisation present an ever-changing variety, such as
are nowhere else to be found. What, however, really renders
India so interesting as an object of study is that "it is now a
living entity. Greece and Rome are dead and have passed
away, and we are living so completely in the midst of modern
Europe, that we cannot get outside to contemplate it as a
whole. But India is a complete cosmos in itself; bounded on
the north by the Himalayas, on the south by the sea, on the
east by jungles inhabited by rude tribes, and only on the west
having one door of communication, across the Indus, open to
the outer world. Across that stream, nation after nation have
poured their myriads into her coveted domain, but no reflex
waves ever mixed her people with those beyond her boundaries.

In consequence of all this, every problem of anthropology
or ethnography can be studied here more easily than anywhere
else; every art has its living -representative, and often of the
most pleasing form ; every science has its illustration, and
many on a scale not easily matched elsewhere. But, notwith-
standing all this, in nine cases out of ten, India and Indian
matters fail to interest, because they are to most people new
and unfamiliar. The rudiments have not been mastered when
young, and, when grown up, few men have the leisure or the
inclination to set to work to learn the forms of a new world,
demanding both care and study ; and till this is attained, it
can hardly be hoped that the arts and the architecture of India
will interest many European readers to the same extent as
those styles treated of in the volumes on ancient and mediaeval
architecture.1

Notwithstanding these drawbacks, it may still be possible
to present the subject of Indian architecture in such a form
as to be interesting, even if not attractive. To do this, however,

i 'History of Architecture in all vols. i. and ii., 3rd ed. (1893). Edited
Countries from the Earliest Times.' By .by R. Phen£ Spiers, F.S.A.
the late Jas. Fergusson, C.I.E., D.C.L.,

INTRODUCTION. 5

the narrative form must be followed as far as is compatible
with such a subject. All technical and unfamiliar names must
be avoided wherever it is possible to do so, and the whole
accompanied with a sufficient number of illustrations to enable
its forms to be mastered without difficulty. Even if this is
attended to, no one volume can tell the whole of so varied
and so complex a history. Without preliminary or subsequent
study it can hardly be expected that so new and so vast a
subject can be grasped ; but one volume may contain a complete
outline of the whole, and enable any one who wishes for more
information to know where to look for it, or how to appreciate
it when found.

Whether successful or not, it seems well worth while that
an attempt should be made to interest the public in Indian
architectural art ; first, because the artist and architect will
certainly acquire broader and more varied views of their art
by its study than they can acquire from any other source.
More than this, any one who masters the subject sufficiently
to be able to understand their art in its best and highest
forms, will rise from the study with a kindlier feeling towards
the nations of India, and a higher — certainly a corrector —
appreciation of their social status than could be obtained from
their literature, or from anything that now exists in their
anomalous social and political position.

Notwithstanding all this, many may be inclined to ask, Is
it worth while to master all the geographical and historical
details necessary to unravel so tangled a web as this, and then
try to become so familiar with their ever-varying forms as not
only to be able to discriminate between the different styles,
but also to follow them through all their ceaseless changes ?

My impression is that this question may fairly be answered
in the affirmative. No one has a right to say that he under-
stands the history of architecture who leaves out of his view
the works of an immense portion of the human race, which
has always shown itself so capable of artistic development.
But, more than this, architecture in India is still a living art,
practised on the principles which caused its wonderful develop-
ment in Europe in the I2th and I3th centuries; and there,
consequently, and there alone, the student of architecture has
a chance of seeing the real principles of the art in action. In
Europe, at the present day, architecture is practised in a
manner so anomalous and abnormal that few, if any, have
hitherto been able to shake off the influence of a false system,
and to see that the art of ornamental building can be based
on principles of common sense ; and that, when so practised,
the result not only is, but must be, satisfactory. Those who

6 HISTORY OF INDIAN ARCHITECTURE.

have an opportunity of seeing what perfect buildings the
uneducated natives of India produce, will easily understand
how success may be achieved, while those who observe what
failures the best educated and most talented architects in
Europe frequently perpetrate, may, by a study of Indian
models, easily see why this must inevitably be the result. It
is only in India that the two systems can be seen practised
side by side — the educated and intellectual European failing
because his principles are wrong, the feeble and uneducated
native as inevitably succeeding because his principles are right.
The Indian builders think only of what they are doing, and
how they can best produce the effect they desire. In the
European system it is considered more essential that a building,
especially in its details, should be a correct copy of something
else, than good in itself or appropriate to its purpose : hence the
difference in the result.

In one other respect India affords a singularly favourable
field to the student of architecture. In no other country of the
same extent are there so many distinct nationalities, each retain-
ing its old belief and its old feelings, and impressing these on
its art. There is consequently no country where the outlines of
ethnology as applied to art can be so easily perceived, or their
application to the elucidation of the various problems so pre-
eminently important. The mode in which the art has been
practised in Europe for the last three centuries has been very
confusing. In India it is clear and intelligible. No one can
look at the subject without seeing its importance, and no one
can study the art as practised there without recognising what
the principles of the science really are.

In addition, however, to these scientific advantages, it will
undoubtedly be conceded by those who are familiar with the
subject that for certain qualities the Indian buildings are un-
rivalled. They display an exuberance of fancy, a lavishness of
labour, and an elaboration of detail to be found nowhere else.
They may contain nothing so sublime as the hall at Karnak,
nothing so intellectual as the Parthenon, nor so constructively
grand as a mediaeval cathedral ; but for certain other qualities
— not perhaps of the highest kind, yet very important in
architectural art — the Indian buildings stand alone. They
consequently fill up a great gap in our knowledge of the
subject, which without them would remain a void.

HISTORV.

One of the greatest difficulties that exist — perhaps the
greatest — in exciting an interest in Indian antiquities arises

INTRODUCTION. 7

from the fact, that India has no history properly so called,
before the Muhammadan invasion in the I3th century. Had
India been a great united kingdom, like China, with a long
line of dynasties and well-recorded dates attached to them,
the task would have been comparatively easy ; but nothing
of the sort ever existed within her boundaries. On the
contrary, so far as our knowledge extends, India has always
been occupied by three or four different races of mankind,
who have never amalgamated so as to become one people,
and each of these races has been again subdivided into
numerous tribes or small nationalities nearly, sometimes wholly,
independent of each other — and, what is worse than all, not
one of them ever kept a chronicle or preserved a series of
dates commencing from any well-known era.1

The absence of any historical record is the more striking,
because India possesses a written literature equal to, if not sur-
passing in variety and extent, that possessed by any other nation,
before the adoption and use of printing. The Vedas themselves,
with their Upanishads and Brahmanas, and the commentaries
on them, form a literature in themselves of vast extent, and
some parts of which are as old, possibly older, than any written
works that are now known to exist ; and the Puranas, though
comparatively modern, make up a body of doctrine mixed with
mythology and tradition such as few nations can boast of.
Besides this, however, are the two great epics, surpassing in
extent, if not in merit, those of any ancient nation, and a drama
of great beauty, written at periods extending through a long
series of years. In addition to these we have treatises on law,
on grammar, on astronomy, on metaphysics and mathematics,
on almost every branch of mental science — a literature extend-
ing in fact to many thousand works, but in all this not one book
that can be called historical. No man in ancient India, so far
as is known, ever thought of recording the events of his own
life, or of repeating the previous experience of others, and it was
not till shortly before the Christian Era that they thought of
establishing eras from which to date deeds or events.

All this is the more curious because in Ceylon we have, in
the ' Dipawansa,' ' Mahawansa,' and other books of a like nature,

1 The following brief resume of the order to make it readable, all references

principal events in the ancient history and all proofs of disputed facts have

of India has no pretensions to being a I been here avoided. They will be found

complete or exhaustive view of the sub- 1 in the body of the work, where they are

ject. It is intended only as such a
popular sketch as shall enable the general
reader to grasp the main features of the
story to such an extent as may enable
him to understand what follows. In

more appropriate. But without some
such introductory notice of the political
history and ethnography, the artistic
history would be nearly, if not wholly
unintelligible.

8 HISTORY OF INDIAN ARCHITECTURE.

a consecutive history of that island, with dates which, with
certain corrections, may be depended upon within certain
limits of error, for periods extending from about B.C. 250
to the present time. At the other extremity of India, we
have also in the ' Rajatarangini ' of Kashmir, a work of the
1 2th century, which Professor Wilson characterised as "the
only Sanskrit composition yet discovered to which the title
of History can with any propriety be applied."1 It hardly
helps us, however, to any ancient historical data, its early
chronology being only traditional and confused ; but from the
beginning of the 9th Christian century, its materials are of
great value.2

In India Proper, however, we have no such guides as even
these, but for written history are almost wholly dependent on
the Puranas. They furnish us with a list of kings' names, with
the length of their reigns, so apparently truthful that they may,
within certain limits, be of use. They are only, however, of one
range of dynasties — probably also sometimes contemporary
— and extend only from the accession of Chandragupta — the
Sandrokottos of the Greeks — about B.C. 320, to the decline of
the Andhra dynasty, about the beginning of the 3rd century A.D.
It seems possible we may yet find sufficient confirmation of
these lists as far back as the 6th century B.C., so as to include
the period marked by the life and labours of Sakyamuni — the
last Buddha — in our chronology, with tolerable certainty. All
chronology before that period is as yet merely conjectural.
From the period of the Gupta dynasty in the 4th century
onwards, when the Puranas began to be put into their present
form, in consequence of the revival of the Brahmanical religion,
instead of recording contemporary events, they purposely
confused them so as to maintain their pretended prophetic
character, and prevent the detection of the falsehood of their
claim to an antiquity equal to that of the Vedas.

For Indian history after the 5th century we are consequently
left mainly to inscriptions on monuments or on copper-plates, to
coins, and to the works of foreigners for the necessary informa-
tion with which the natives of the country itself have neglected
to supply us. Inscriptions fortunately are more abundant in
India than, perhaps, in any other country, and nearly all of
them contain historical information ; and, thanks to the great
advances made in epigraphy during the last thirty-five years,
we are now able to piece together a tolerably accurate historical

1 'Asiatic Researches,' vol. xv. p. i. ! a valuable commentary and notes by
* Kalhana's ' Rajatarangini ' has been | Dr. A. M. Stein, 2 vols. (London,
very carefully translated and edited with 1900).

INTRODUCTION. 9

outline of the course of events from the 3rd century B.C.1 This
is more especially the case for the Dekhan and the north of India ;
in the Tamil country so much has not yet been done, but this is
more because there have been fewer labourers in the field than
from want of materials. There are literally thousands of
inscriptions in the south which have not been copied, and of
those that have been collected only a portion have yet been
translated ; but they are such as to give us assurance that, when
the requisite amount of labour is bestowed upon them, we shall
be able to fix the chronology of the kings of the south with a
degree of certainty sufficient for all ordinary purposes.2

It is a far more difficult task to ascertain whether we shall
ever recover the History of India before the time of the advent
of Buddha. Here we certainly will find no coins or inscriptions
to guide us, and no buildings to illustrate the arts, or to mark
the position of cities, while all ethnographic traces have become
so blurred, if not obliterated, that they serve us little as guides
through the labyrinth. Yet on the other hand there is so much
literature — such as it is — bearing on the subject, that we cannot
but hope that, when a sufficient amount of learning is brought
to bear upon it, the leading features of the history of even that
period may be recovered. In order, however, to render it
available, it will not require industry so much as a severe spirit
of criticism to winnow the few grains of useful truth out of the
mass of worthless chaff this literature contains. But it does not
seem too much to expect even this, from the severely critical
spirit of the age. Meanwhile, the main facts of the case seem to
be nearly as follows, in so far as it is necessary to state them, in
order to make what follows intelligible.

ARYANS.

At some very remote period in the world's history the Aryas
or Aryans3 — a people speaking an early form of Sanskrit —

1 The chronological results have been
systematically arranged in that useful
handbook.— Duff's 'Chronology of India'
(London, 1899).

2 Almost the only person who had
done anything in this direction till forty
years ago was the late Sir Walter Elliot.
Since 1872 the labours of Drs. Fleet,
Biihler, Kielhorn, R. G. Bhandarkar and
others have thrown a flood of light on the
history of southern as well as northern
India ; and within the last twenty years
Dr. Hultzsch's work among the Tamil
inscriptions of Madras has yielded very

important chronological and historical
information for the south of the penin-
sula. The Mysore Government has also
issued the great ' Epigraphia Carnatica,'
under the direction of Mr. Lewis Rice.

3 We have the word in the ' Aria ' and
' Ariana ' of the Greek writers, applied
to the country lying to the north-east of
Persia adjoining Baktriana. The early
Zoroastrians called their country ' Airy-
anavaej6' — the Afya. home, and in the
Behistun inscriptions it is styled ' Ariya.'
See Lassen, ' Indische Alterthumskunde,'
Bd. i. Ss. 5ff.

to HISTORY OF INDIAN ARCHITECTURE.

entered India across the Upper Indus, coming from Central
Asia. They were a fair complexioned people as compared with
the Aborigines, and for a long time they remained settled in the
Panjab, or on the banks of the Sarasvati, then a more important
stream than now, the main body, however, still remaining to the
westward of the Indus. If, however, we may trust our
chronology, we find them settled 150x3 to 2000 years before the
Christian Era, in Ayodhya and then in the plenitude of their
power. Naturally we look for some light on their early history
in the two great Indian epics — the Ramayana recording the
exploits of Rama, King of Ayodhya, of the Solar race, and in
much later times regarded as an incarnation of Vishnu ; and
the Mahabharata celebrating the contest between the Kurus
and Pandus, of the Lunar family. Both are steeped in Brahman
doctrines, almost certainly inserted in later ages among the
original legends. It thus becomes very difficult to separate
what belongs to the original spirit and aim of the works from
the interpolated materials. The Ramayana is so largely
allegorical and cast in the form it has reached us so long after
the period to which it refers that it is doubtful whether we can
draw any inference with safety from its contents, except that it
relates to the spread of Aryan civilisation — which had probably
then occupied most of the country north of the Vindhyan range
— into southern India, and as far as Ceylon.1 From a very
early period the Aryans had, doubtless, become mixed with
aboriginal races, and could not be regarded as pure at this
period. But whether they formed settlements in the Dekhan
or not, it was opened up to them, and by slow degrees imbibed
that amount of Brahmanism which eventually pervaded the
south. By B.C. 700, or thereabouts, they had begun to be
tolerably well acquainted with the whole of the peninsula.

The events that form the theme of the western epic — the
Mahabharata — may have occurred almost as early as, or even
several centuries later than the times of Rama. It opens up an
entirely new view of Indian social life. If the heroes of that
poem were Aryans at all, they were of a much less pure type
than those who composed the songs of the Vedas, or are
depicted in the verses of the Ramayana. Their polyandry,
their drinking bouts, their gambling tastes, and love of fighting,
mark them as a very different race from the peaceful shepherd
immigrants of the earlier age, and point much more distinctly
towards a Tartar, trans- Himalayan origin, than to the cradle of

1 For some account of the probable
spread of the Aryas southwards, see
Dr. R. G. Bhandarkar's 'Early History

of the Dekhan,' in Bombay Gazetteer
(1895), vol. i. pt. ii. pp. I32tf.

INTRODUCTION.

ii

the Aryan stock in Central Asia. We are tempted to ask
whether the Pandavas, who conquered in the great strife, were
not a confederacy of hostile tribes, headed by a band of warriors
of mixed or non-Aryan descent. Their helper and counsellor,
KrishnS, is a divinity unknown to the early Aryans, and with
him as an incarnation of Vishnu, the Siva, and Brahma, of modern
Hinduism, take the place of the older gods. As if to mark the
difference of which the warriors themselves felt the existence,
they distinguished themselves, by name, as belonging to a
Lunar race, distinct from, and generally antagonistic to, the
Solar race, which was the proud distinction of the purer and
earlier Aryan settlers in India.1

By about B.C. 700, we again find a totally different state of
affairs in India. The Aryans no longer exist as a separate
nationality, and neither the Solar nor the Lunar race are the
rulers of the earth. The Brahmans have become a priestly
caste, and share the power with the Kshatriyas, a race of far less
purity of descent. The Vai^yas, as merchants and husbandmen,
have become a power, and even the Sudras are acknowledged
as a part of the body politic ; and — though not mentioned in
the Scriptures — the Nagas, or Snake people, had become an
influential part of the population. They are first mentioned in
the Mahabharata, where they play a most important part in
causing the death of Parikshit, which led to the great sacrifice
for the destruction of the Nagas of Taksha^ila by Janamejaya,
which practically closes the history of the time. Destroyed,
however, they were not, for we find Naga dynasties ruling in
various parts of Central India and Rajputana from the 7th
century B.C., till at least the 4th century A.D.2

Although Buddhism was first taught probably by one belong-
ing to the Solar race, and of Aryan blood, and though its first
disciples were Brahmans, it had as little affinity with the religion
of the Vedas as Christianity had with the Pentateuch, and its fate
was the same. The one religion was taught by one of Jewish ex-
traction to the Jews ; but it was ultimately rejected by them, and
adopted by the Gentiles, who had no affinity of race or religion
with the inhabitants of Judaea. Though meant originally, no
doubt, for Aryans, the Buddhist religion was ultimately rejected
by the Brahmans, who were consequently eclipsed and superseded
by it for nearly a thousand years ; and we hear little of them and
their religion till they rise again at the court of the great Gupta

1 Orientalists have expressed very
varying opinions as to the historical
teachings of the epics. See Weber,
' On the Ramayana,' etc.

2 The Naga or Karkota dynasty of
Kashmir ruled as late as from about the
beginning of the 7th till the middle of
the gth century.

iz HISTORY OF INDIAN ARCHITECTURE.

kings in the 4th century A.D., when their religion began to
assume that strange shape which it now still retains in India.
In its new form it is as unlike the religion of the Vedas as it is
possible to conceive one religion being to another ; unlike that,
also, of the older portions of the Mahabharata ; but a confused
mess of local superstitions and imported myths, covering up and
hiding the Vedantic and Buddhist doctrines, which may some-
times be detected as underlying it. Whatever it be, however, it
was invented by and for as mixed a population as probably were
ever gathered together into one country — a people whose feelings
and superstitions it only too truly represents.

DRAVIDIANS.

Although, therefore, as was hinted above, there might be no
great difficulty in recovering the main incidents and leading
features of the history of the Aryans, from their first entry into
India till they were entirely absorbed into the mass of the popula-
tion some time before the Christian Era, there could be no greater
mistake than to suppose that their history would fully represent
the ancient history of the country. The Dravidians are a people
who, in historical times, seem to have been probably as numerous
as the pure Aryans, and at the present day form one-fifth of the
whole population of India. They belong, it is true, to a lower
intellectual status than the Aryans, but they have preserved
their nationality pure and unmixed, and, such as they were at
the dawn of history, so they seem to be now.

Their settlement in India extends to such remote pre-historic
times, that we cannot feel even sure that we should regard them
as immigrants, or, at least, as either conquerors or colonists on a
large scale, but rather as aboriginal in the sense in which that
term is usually understood. Generally it is assumed that they
entered India across the Lower Indus, leaving the cognate
Brahui in Baluchistan as a mark of the road by which they
came, and, as the affinities of their language seem to be with
the Ugrians and Northern Turanian tongues, this view seems
probable.1 But they have certainly left no trace of their migra-
tions anywhere between the Indus and the Narbada, and all the
facts of their history, so far as they are known, would seem to
lead to an opposite conclusion. The hypothesis that would
represent what we know of their history most correctly would
place their original seat in the extreme south, somewhere

1 Dr. Caldwell, the author of the I and most trustworthy advocate of this
' Dravidian Grammar,' is the greatest J view.

INTRODUCTION. 13

probably not far from Madura or Tanjor, and thence spread-
ing fan-like towards the north, till they met the Aryans in
the northern Dekhan. The question, again, is not of much
importance for our present purposes, as we do not know to
what degree of civilisation they had reached anterior to the
Christian Era, or when they were first able to practise the arts
of civilised life with such success as to bring them within the
scope of a work devoted to the history of art.1

It may be that at some future period, when we know more
of the ancient arts of these Dravidians than we now do, some
fresh light may be thrown on this very obscure part of history.
Geographically, however, one thing seems tolerably clear. If
the Dravidians came into India in historical times it was not
from Central Asia that they migrated, but from Persia, or some
southern region of the Asiatic continent.

DASYUS.

In addition to these two great distinct and opposite nationa-
lities, there exists in India a third, which, in pre-Buddhist times,
was as numerous, perhaps even more so, than either the Aryans
or the Dravidians, but of whose history we know even less than
we do of the two others. Ethnologists have not agreed on a
name by which to call them. I have suggested Dasyus,2 a slave
people, as that is the name by which the Aryans designated
them when they found them there on their first entrance into
India, and subjected them to their sway.3 Possibly they were
partly of Mongol-Tibetan origin, and partly they may have been
a mixed race allied to the Dravidians, and now represented by
Gonds, Santals, Bhils, etc.

The Dasyus, however, were not mere barbarians ; for they
had towns, and traces of at least a partial civilisation ; they
had leaders or chiefs possessed of strong fortified retreats, and
they possessed treasures of gold and rich jewels.4 Whoever
they were they seem to have been a people of less intellectual
capacity, less muscular, and less united than their invaders.
When the Aryans first entered India they seem to have found
them occupying the whole valley of the Ganges — the whole

1 In the ' Ramayana ' the monkey-
soldiers are directed to the countries
of the Andhras, Pandyas, Cholas, and
Keralas, in the south, and are told they
will there see the gate of the city of the
Pandyas adorned with gold and jewels.

2 'Tree and Serpent Worship,'
pp. 244-247.

1 ' Dasyu ' probably meant 'provincial,'

'aboriginal,' and was used much as
'Gentiles,' 'Pagans,' 'Barbarians,' in
early times. They are also termed
Yadvas, of which we may have a
survival in the 'Jats.' A

4 Vivian de Saint Martin's ' Etude sur
la Geographic et les populations primi-
tives du Nord-ouest de 1'Inde, d'apres
les Hymnes vediques.'

i4 HISTORY OF INDIAN ARCHITECTURE.

country in fact between the Vindhya and the Himalaya
Mountains. At present they are only found in anything like
purity in the mountain ranges that bound that great plain.
There they are known as Bhils, Gonds, Kandhs, Mundas,
Oraons, Hos, Kols, Santals, Nagas, and other mountain and
forest tribes. But they certainly form the lowest underlying
stratum of the population over the whole of the Gangetic plain.
So far as their affinities have been ascertained, they are with
the trans - Himalayan population, and it either is that they
entered India through the passes of that great mountain range,
or it might be more correct to say that the Tibetans are a
fragment of a great population that occupied both the northern
and southern slope of that great chain of hills at some very
remote pre-historic time.

Whoever they were, they were the people who, in remote
times, were apparently the worshippers of Trees and Serpents ; l
but what interests us more in them, and makes the enquiry
into their history more desirable, is that it was where the
people were largely of this aboriginal stock that Buddhism
seems to have been most readily adopted, and it is largely
among allied races that it is still adhered to. In Ceylon,
Tibet, Burma, Siam, and China — wherever a people allied to
the Mongol or Tibetan family exists, there Buddhism flourished
and still prevails. But in India a revival of Brahmanism
abolished it.

Architecturally, there is no difficulty in defining the limits of
the Dasyu province : wherever a square tower-like temple exists
with a perpendicular base, but a curvilinear outline above, such
as that shown in the woodcut (No. i), there we mav feel certain
of the existence, past or present, of a people of Dasyu extrac-
tion. No one can accuse the pure Aryans of introducing this
form into India, or of building temples at all, or of worshipping
images of Siva or Vishnu, with which these temples are filled,
and they consequently have little title to confer their name on
the style. The Aryans had, however, become so impure in
blood before these temples were erected, and were so mixed
up with the aboriginal tribes whose superstitions had so
influenced their religion and their arts that they accepted their
temples with their gods. Be this as it may, one thing seems
tolerably clear, that the regions occupied by the Aryans in
India were conterminous with those of the Dasyus, or, in
other words, that the Aryans conquered the whole of the
aboriginal or native tribes who occupied the plains of northern
India, and ruled over them to such an extent as materially to

1 See ' Indian Antiquary,' vol. iv. pp. 5f.

INTRODUCTION.

influence their religion and their arts, and also very materially
to modify even their language. So much so, indeed, that after
some four thousand years of domination we should not be
surprised if we have some difficulty in recovering traces of
the original population, and could probably not do so, if
some fragments of the people had not sought refuge in the
hills on the north and south of the great Gangetic plain, and
there have remained fossilised, or at least sufficiently permanent
for purposes of investigation.

i. Hindu Temple, at Bahulara, near Bankura.

SAISUNAGA DYNASTY, ABOUT B.C. 650 TO 318.

Leaving these, which must, for the present at least, be
considered as practically pre-historic times, we tread on surer
ground when we approach the period when Buddha was born,
and devoted his life to solve the problem of suffering and
transmigration. There seems little reason for doubting that
he was born about the year 560, taught during the reign of
Bimbisara, the fifth king of the SaLrunuga dynasty, and died
B.C. 480,1 at the age of eighty, in the eighth year of Ajatajatru,
the sixth king.2 New sources of information regarding these

1 Dr. J. F. Fleet, in 'Journal of the
Royal Asiatic Society,' 1906, pp. 9845.

the dates given are quite near enough.
The ' Matsya Purana ' alone inserts

• f *r A

. * / i -7 — ) t'r" ^^t*. * tit, ifirtta^d. i u; una, luune inseris

ces the death or Nirvana of Buddha in the reignsof Kan vSyana— nine years— and

B.C. 482, with other connected dates in j Bhumimitra — fourteen years — between

>rdance with it. For our purpose | Bimbisara and Ajata-ratru.

i6

HISTORY OF INDIAN ARCHITECTURE.

times are opening out, and we may before long be able to
recover a fairly authentic account of the political events of
that period, and as perfect a picture of the manners and the
customs of those days. It is too true, however, that those
who wrote the biography of Buddha in subsequent ages so
overlaid the narrative of his life with fables and absurdities,
that it is now difficult to separate the wheat from the chaff;
but we have sculptures extending back to within three centuries
of his death, at which time we may fairly assume that a purer
tradition may have prevailed. From what has already occurred,
we may hope to creep even further back than this, and eventu-
ally to find early illustrations which will enable us to exercise
so sound a criticism on the books as to enable us to restore
the life of Buddha to such an extent as to place it on a basis
of credible historicity.

Immense progress has been made during the last fifty or
sixty years in investigating the origin of Buddhism, and the
propagation of its doctrines in India, and in communicating
the knowledge so gained to the public in Europe. More, how-
ever, remains to be done before the story is complete, and
divested of all the absurdities which subsequent commentators
have heaped upon it. Still, the leading events in the life of
the founder of the religion are simple, and now sufficiently well
ascertained for all practical purposes.1

The founder of this religion was claimed by tradition as
one of the last of a long line of kings, known as the Solar
dynasties, who, from a period shortly subsequent to the advent
of the Aryans into India, had held paramount sway in Ayodhya
— the modern Oudh. About the I2th or I3th century B.C. they
were superseded by another race of much less purely Aryan
blood, known as the Lunar race, who transferred the seat of
power to capitals situated in the northern parts of the Doab.
But the tradition of the royal birth of Sakyamuni can hardly
be sustained historically. He seems to have been born at
Kapilavastu, at the foot of the Himalayas, as the son of

1 The most pleasing of the histories of
Buddha, written wholly from a European
point of view, is that of Barthelemy
St. Hilaire, Paris. Of those partially
native, partly European, are those of
Bishop Bigandet, from the Burmese
legends, and the ' Romantic History of
Buddha,' translated from the Chinese
by the Rev. S. Beal. The ' Lalita Vis-
tara,' translated by Foucaux, is more
modern than these, and consequently
more fabulous and absurd. In more

recent years a large literature has
appeared on the subject. Prof. H.
Oldenberg's ' Buddha : his Life, his
Doctrine, his Order,' translated from
the German by W. Hoey (1882) supplies
an able critical estimate of the teacher.
Dr. H. Kern's ' History of Buddhism in
India' has been translated into French
by G. Huet (Paris, 1901-1903); and
W. W. Rockhill's 'Life of Buddha,
and the Early History of his Order,'
1884, are also valuable works.

INTRODUCTION. 17

a petty chief. For twenty-nine years he is represented as
enjoying the pleasures, and following the occupations, usual to
the men of his rank and position ; but at that age, becoming
painfully impressed by the misery incident to human existence,
he determined to devote the rest of his life to an attempt
to alleviate it. For this purpose he forsook his parents and
wife, abandoned friends and the advantages of his position,
and, for the following fifty years, devoted himself steadily to
the task he had set before himself. Years were spent in
the meditation and mortification supposed to be necessary
to fit him for his mission ; the rest of his long life was
devoted to wandering from city to city, teaching and
preaching, and doing everything that gentle means could
effect to disseminate the doctrines which he believed were
to regenerate the world, and take the sting out of human
misery.

He died, or, in the phraseology of his followers, entered
Nirvana or Parinirvana — was absorbed into nothingness —
at Kujinara, in the eightieth year of his age, about 480
years B.C.

With the information that has accumulated around the
subject, there seems no great difficulty in surmising why the
mission of Sakyamuni was so successful as it proved to be.
He was born in an age when the purity of the Aryan races,
especially in eastern India, had become so deteriorated by
intermixture with aborigines, and with less pure tribes coming
from the north, that their power, and consequently their
distinctive influence was fading away. At that time, too, the
native and mixed races had acquired such a degree of civili-
sation as led them to claim something like equality with their
Aryan masters. In such a condition of things the preacher
was sure of a willing audience who ignored caste, and taught
that all men, of whatever nation or degree, had an equal chance
of reaching happiness, and ultimately Nirvana, by the practice of
virtue : in a word — to be delivered from the wearisome bondage
of ritual or caste observances and the depressing prospect of
interminable transmigration. Aboriginal or Turanian Dasyus,
perhaps even more readily than the mixed Aryans, would
hail him as a deliverer, and by the former the new religion
was specially adopted and propagated, whilst that of the
Brahman Aryans was, for a time at least, overshadowed and
obscured.

It is by no means clear how far Buddha was successful in

converting the multitude to his doctrines during his lifetime.

At his death, the first synod or council was held at Rajagriha,

and five hundred monks of a superior order, it is said, were

VOL. i. B

i8

HISTORY OF INDIAN ARCHITECTURE.

assembled there on that occasion,1 and if so they must have
represented a great multitude. But the accounts of this, and of
a second convocation, said, by the southern Buddhists, to have
been held one hundred years afterwards at Vaijali, are of doubtful
authenticity. Indeed, the whole annals of the .Saijunaga dynasty
from the death of Buddha till the accession of Chandragupta,
dr. B.C. 320, are about the least satisfactory of the time. Those
of Ceylon were falsified in order to make the landing of Vijaya,
the alleged first conqueror from Kalinga, coincident with the
date of Buddha's death, while a period of some length elapsed
between the two events.2 We have annals, and we may possibly
recover inscriptions3 and sculptures belonging to this period,
and though it is most improbable we shall recover any archi-
tectural remains, there are possibly materials existing which,
when utilised, may suffice for the purpose.

The kings of this dynasty seem to have been considered as
of a low caste, and were not, consequently, in favour either with
the Brahman or, at that time, with the Buddhist ; and no events
which seem to have been thought worthy of being remembered,
except the second convocation — the fact of which is doubtful —
are recorded as happening in their reigns, after the death of the
great Ascetic — or, at all events, of being recorded in such annals
as we possess.

MAURYA DYNASTY, B.C. 320 TO 180.

The case was widely different with the Maurya dynasty,
which was certainly one of the most brilliant, and is fortunately
one of the best known, of the ancient dynasties of India. The
first king was Chandragupta, the Sandrokottos of the Greeks, to

1 See Rockhill, ' Life of the Buddha,'
p. 156; Kern's 'Histoire du Bouddhisme
dans 1'Inde,' torn. ii. pp. 253ff. (French
trans. ).

2 There is an error of about sixty years
in the usual date — B.C. 543, derived from
the Singhalese chronicles, which is else-
where corrected. The revised date may
not be precisely correct, but it must be
approximately so.

* We have no very early Hindu coins ;
the earliest are square or oblong punch-
marked pieces, which seem to date from
about a century before Alexander, and
supply no historical data. The late Mr.
Ed. Thomas supposed a coin, bearing
the name of Amoghabhuti, a Kuninda,

belonged to one of the nine Nandas with
whom this dynasty closed ('Journal of
the Royal Asiatic Society,' N.S., vol. i.
pp. 447ff.). It is now known that such

coins do not belong to a date earlier
than about B.C. 100. The earliest coins
of historical value for India are those
of the Grseco-Baktrians and their con-
temporaries or successors on the north-
west frontier.

INTRODUCTION. 19

whom Megasthenes was sent as ambassador by Seleukos, the
successor of Alexander in the western parts of his Asiatic
empire. It is from his narrative — now unfortunately lost — that
the Greeks acquired almost all the knowledge they possessed of
India at that period.1 The country was then divided into 120
smaller principalities, but the Maurya residing in Palibothra
(Pataliputra) — the modern Patna — seems to have exercised a
paramount sway over the whole. It was not, however, this
king, but his grandson, the great A^oka (B.C. 262 to 225), who
raised this dynasty to its highest pitch of prosperity and power.
Though utterly unknown to the Greeks, we have from native
sources a more complete picture of the incidents of his reign
than of any ancient sovereign of India. The great event that
made him famous in Buddhist history was his conversion to
that faith, and the zeal he showed in propagating the doctrines
of his new religion. He did, in fact, for Buddhism, what
Constantine did for Christianity, and at about the same distance
of time from the death of the founder of the faith. From a
struggling sect he made it the religion of the State, and
established it on the basis on which it lasted supreme for nearly
1000 years. In order to render his subjects familiar with the
doctrines of his new faith, he caused a series of edicts embody-
ing them to be engraved on rocks near Peshawar, in Gujarat, in
the valley of the Dun under the Himalayas, in Hazara, in Katak
and Ganjam, in Mysore, and other places. He held a great
convocation or council of the faithful in his capital at Pataliputra,
and, on its dissolution, missionaries were sent to spread the
religion in the Yavana country, whose capital was Alexandria,
near the present city of Kabul. Others were despatched to
Kashmir and Gandhara ; one was sent to the Himawanta — the
valleys of the Himalaya, and possibly part of Tibet ; others
were despatched to the Maharatta country, and to Mysore, to
Vanavasi in Kanara, and to Aparantaka or the north Konkan.
Two missionaries were sent to Suvarnabhumi, now known
as Thatun on the Sitang river, in Pegu, and, tradition says, his
own son and daughter were deputed to Ceylon.2 All those
countries, in fact, which might be called foreign, were then
sought to be converted to the faith. He also formed alliances
with Antiokhos the Great, Antigonos, and with Ptolemy
Philadelphos, Alexander of Epeiros, and Magas of Cyrene,

1 For this period, see M'Crindle's | -' All these particulars, it need hardly

' Ancient India as described by Megas
thenes and Arrian ' (1877) ; the 'Inva-
sion of India by Alexander the Great '
(1896); and ' Ancient India as described
in Classical Literature ' (1901).

be said, are taken from the 1 2th and
1 5th chapters of the 'Mahawansa,' which
relates the traditions of a time six centuries
and more before its composition.

20

HISTORY OF INDIAN ARCHITECTURE.

for the establishment of hospitals and the protection of his
co-religionists in their countries. More than all this, he built
innumerable topes or stupas and monasteries all over the
country ; and, though none of those now existing can positively
be identified as those actually built by him, there seems no
reason for doubting that the sculptured rails at Bodh-Gaya
and Bharaut, the caves at Barabar in Bihar, some of those at
Udayagiri in Katak, and the oldest of those in the Western
Ghats were all erected or excavated during the existence of this
dynasty, if not under himself. These, with inscriptions and
such histories as exist, make up a mass of materials for a picture
of India during this dynasty such as no other can present ; and,
above all, they offer a complete representation of the religious
forms and beliefs of the kings and people, which render any
mistake regarding them impossible.1 It was Buddhism, but
without a deified Buddha, and with Tree and Serpent worship
cropping up in every unexpected corner.

There is certainly no dynasty in the whole range of ancient
Indian history that would better repay the labour of an
exhaustive investigation than that of these Maurya kings.
Not only were they the first in historical times who, so far
as we know, united nearly the whole of India into one great
kingdom, but they were practically the first who came in contact
with European civilisation and Western politics. More than
even this, it is probably owing to the action of the third king
of this dynasty that Buddhism, from being the religion of an
obscure sect, became, at one time, the creed of so large a
proportion of the human race, and influenced the belief and the
moral feelings of such multitudes of men in Asia.

It is to this dynasty, and to it only, that must be applied all
those passages in classical authors which describe the internal
state of India, and they are neither few nor insignificant. Though
the Hindus themselves cannot be said to have contributed much
history, they have given us, in the c Mudra-Rakshasa,' 2 a poetical
version of the causes of the revolution that placed the Mauryas
on the throne. But, putting these aside, their own inscriptions
supply us with a perfectly authentic contemporary account of
the religious faith and feelings of the period; while the numerous
bas-reliefs of the rails at Bodh - Gaya and Bharaut afford a
picture of the manners, customs, and costumes of the day, and a
gauge by which we can measure their artistic status and judge
how far their art was indigenous, how far influenced by foreign

1 For fuller information about Aroka
reference may be made to Edm. Hardy's
•KonigAsoka,'and V. A. Smith's 'Asoka,
the Buddhist Emperor of India.'

2 Wilson's ' Hindu Drama,' in
'Works,' vol. xii, pp. 151 et si
edition 1871.

his

INTRODUCTION. 21

elements. The dates of the kings of this dynasty are also
approximately known, and the whole framework of their history
depends so completely on contemporary native monuments, that
there need be no real uncertainty regarding any of the outlines
of the picture when once the subject is fairly grasped and
thoroughly handled.

It is the firmest standpoint we have from which to judge of
Indian civilisation and history, whether looking to the past or to
the future, and it is one that gives a very high idea of the
position at which the Hindus had arrived before they came
practically into contact with the civilisation of the West.

SUNGA DYNASTY, B.C. 180 TO 70.
KANWAYANA DYNASTY, B.C. 70 TO 36.

History affords us little beyond the lengths of the kings'
reigns for the next two dynasties, and we are obliged to trust
to the general correctness with which these are recorded in the
Puranas, and by degrees we are collecting inscriptions, and we
know of caves that belong to their time, so that we may hope to
breathe life into what has hitherto appeared only a dry list of
names. Possibly the Kanwas had usurped the power of the
Sungas, so that the two families may have been nominally con-
temporary during the period assigned to the latter, and that
both came to an end about 40 B.C. Anyhow we know that
the Andhras had risen to power on the decline of the Mauryas.
These dynasties were not, however, apparently known to the
Greeks, and possibly, being Buddhist, are passed over in com-
parative silence in the Puranas. It is thus only from their
monuments that we can hope to recover their history. Up to
the present time, the most important inscription discovered is
that of a prince Dhanabhuti who " in the time of the Sungas "
erected a gateway at the Bharaut stupa.

ANDHRA DYNASTY, ABOUT B.C. 170 TO A.D. 220.

The dynasty that ruled the Dekhan at least, contemporary
with these Rois faineants is — after the Mauryas — the most
important of all those about this period of Indian history. To
the classical authors they are known as the Andrae, in the
Puranas as Andhrabhrityas, and in the inscriptions as .Satakarnis
or .Satavahanas ; but under whatever name, notwithstanding
occasional periods of depression, they played a most important
part in the history of India, during more than four centuries.
They were a South-Indian dynasty, first mentioned in a

HISTORY OF INDIAN ARCHITECTURE.

Khandagiri inscription about B.C. I5O.1 Their capital was at
DhSnyakataka, on the lower Krishna, close to Amaravati ;
but, at a later date, they had a second capital at Paithan on
the upper Godavari. They ruled over Malwa and the Dekhan
from sea to sea, but about the end of the 1st or beginning
of the 2nd century the provinces north of the Narbada seem
to have been conquered by Saka. satraps, who were overthrown
about A.D. 125, by Gautamiputra Satakarni, who raised the
dynasty to the height of its power. The kings of this race
have, however, left many and most interesting inscriptions in
the western caves, and traces of their existence occur in many
parts of India.

Architecturally, their history begins with the gateways
of the Tope at Sanchi ; the southern of these was almost
certainly erected during the reign of the first Satakarni about
the middle of the 2nd century before our era — and the other
three in the course of that century. It ends with the com-
pletion of the rail at Amaravati, which was probably commenced
in the ist century, and completed before the end of the 2nd.2

Between these two monuments there is no great difficulty
in filling up the architectural picture from the caves, at Nasik,
Kanheri and Ajanta, and other places in western India, and
more materials may still eventually be discovered.

The history of this dynasty is more than usually interesting
for our purposes, as it embraces nearly the whole period during
which Buddhism reigned almost supreme in India. It became
a State religion, it is true, somewhat earlier under A^oka, but
there is no reason for believing that the Vedic religion or
Brahmanism then vanished. During four or five centuries,
however, after the A-roka Era we have not a trace of a Hindu
building or cave, and, so far as any material evidence goes, it
seems that Buddhism at the time was the predominant religion
of the land. It is not, of course, to be supposed that the Hindu
cult was wholly obliterated, but it was dormant, and in abeyance,
and, to use a Buddhist expression, the yellow robes shone over
the length and breadth of the land.

It was during the rule of these Andhras, though not by
them, that the fourth convocation was held by Kanishka, in
the north of India, and the new doctrine, the Mahayana,

1 ' Epigraphia Indica,' vol. ii. pp.
88-89.

2 For fuller details of the Sanchi and
Amaravati Stupas, the reader is referred
to 'Tree and Serpent Worship,' which
is practically devoted to a description
of them. For a further account of

Amaravati, and a rectification of the
dates in accordance with later dis-
coveries, reference may be made to ' The
Amaravati and Jaggayapeta Stupas '
(1887); and to Dr. R. G. Bhandarkar's
' Early Dekhan Dynasties ' in the Bombay
Gazetteer, vol. i. pt. ii.

INTRODUCTION. 23

introduced by Nagarjuna — a change similar to that made by
Gregory the Great when he established the Church, as opposed
to the primitive forms of Christianity, at about the same dis-
tance of time from the death of the founder of the religion. This
convocation was probably held about B.C. 40. Certain at least
it is, that it was about, or very soon after, that time that
Buddhism was first practically introduced into China, Tibet,
and Burma, and apparently by missionaries sent out from this
as they were from the A^oka convocation.

KSHATRAPAS OF KATHIAWAR, A.D. I2O TO 388.
GUPTAS, 320 TO ABOUT 535.
VALABHIS, ABOUT 600 TO 770.

The Andhras disappear from history early in the 3rd
century ; the Kshatrapas of Gujarat held sway in the west
for a century and a half longer, when they were superseded
by the Gupta dynasty who, at the end of the 4th century of
our era, seem to have attained to the position of lords
paramount in northern India. They date their inscriptions,
which are numerous and interesting, from an era established
by the Gupta king, Chandragupta I., dating 242 years after the
Saka era of A.D. 78, or in 320 ; but it was not apparently till
under the third king, Samudragupta, about 380, that they
really obtained the empire of northern India, which they
retained till the death of Skandagupta, about the year 465,
or it may be a little later.1

It was during the reign of the Guptas that Fah Hian visited
India (A.D. 400). As his objects in doing so were entirely of
a religious nature, he does not allude to worldly politics, nor
give us a king's name we can identify ; but the picture we
gather from his narrative is one of peace and prosperity in
so far as the country is concerned, and of supremacy generally
for his religion. Heretics are, it is true, mentioned occasionally,
but they are few and far between. Buddhism was then the
religion of the north, especially in the north-west of India ;
but even then there were symptoms of a change in the central
provinces and outlying parts of the country.

It is during their rule that we first perceive in high places
the germs of that change which was gradually creeping over
the religious system of India. That the Guptas were not
inimical to Buddhism may be inferred from the gifts that
Amarakardava, an officer of Chandragupta II., made to the

1 The Guptas and their inscriptions
have been dealt with by Dr. J. F. Fleet,

in his work, ' Inscriptions of the Early
Gupta Kings,' Corp. Inscr. Ind., vol. iii.

24 HISTORY OF INDIAN ARCHITECTURE.

stupa at Sanchi in the year A.D. 412, and recorded on the rail
of that Monument, but their other inscriptions, on the lats at
Allahabad, Junagadh, and Bhitari, show a decided tendency
towards Hinduism of the Vaishnava form, but which was still
far removed from the wild extravagances of the Puranas. There
seems little doubt that the boar at Eran, and the buildings
there, belong to this dynasty, and are consequently among the
earliest if not the very oldest temples in India, dedicated to
the new religion, which was then raising its head in defiance
to Buddhism.

From their coins and inscriptions, we may feel certain that
the Guptas possessed, when in the plenitude of their power,
the whole of northern India with the province of Gujarat, but
how far the boasts of Samudragupta (3/0-380) on the Allahabad
pillar were justified is by no means clear. If that inscription
is to be believed, the whole of the southern country as far as
Ceylon, together with, or up to the borders of Asam and Nepal,
were subject to their sway. However brilliant it may have been,
their power was of short duration. Gujarat, with Kathiawar,
from about A.D. 500, was held by the Maitrakas of Valabhi, at
first as feudatories of the Guptas, but, as the paramount power
declined, the Valabhi chiefs gradually assumed independence,
and founded a separate kingdom, which sometimes included
western Malwa, and lasted into the middle of the 8th century.

Although it was evident in the time of the Guptas that a
change was creeping over the religious belief of India, it was
not then that the blow was struck which eventually proved
fatal, but by a dynasty which succeeded them in Central India.

THE SIXTH CENTURY AND AFTER.

The Gupta power seems to have given way before the
inroads of Sakas or Huns, chiefly under Toramana and his son
Mihirakula, who succeeded him about 515, and was a bitter
persecutor of the Buddhists in the North- West. A coalition
was formed against him, and under Ya^odharman of Ujjain he
was totally defeated about 530. At this period the ' Raja-
tarangini' describes Vikramaditya-Harsha of Ujjain as sole
sovereign of India, the destroyer of the Sakas, and patron of
poets, who placed Matrigupta on the throne of Kashmir. It
is possible that this Yajodharman and Vikramaditya are only
birudas or titles of the same sovereign, who may have ruled
till 550 or thereabouts.1 Further, the period seems to suggest

1 Taranitha states that Vikramaditya- I Mlechchhas, massacring them ai Multan,
Harsha abolished the teaching of the | and was succeeded by ilia. The Man-

INTRODUCTION.

that this may have been that Vikramaditya, who, by his
liberality and magnificence, acquired a renown among the
Hindus, only second to that obtained by Solomon among the
Jews. By his patronage of literature and his encouragement
of art, his fame spread over the length and breadth of the
land, and to this day his name is quoted as the symbol of
all that is great and magnificent in India. What is more to
our present purpose, he was an undoubted patron of the
Brahmanical religion, and no tradition associates his name
directly or indirectly with anything connected with Buddhism.
Unfortunately we have no buildings which can be attributed
to him. But the main fact of a Brahmanical king reigning
and acquiring such influence in Central India at that time may
be significant of the declining position of the Buddhist religion
at that period.

His successor, Pratapa-rila-Siladitya's reign would fall about
the end of the 6th century, and he is spoken of by Hiuen
Tsiang as a patron of Buddhism.1 But it was usual with Hindu
kings to show favour to the various sects among their subjects
indiscriminately, and the Chinese pilgrim's statement that
during his long reign of about sixty years he honoured the
Buddhists and their doctrines, is no proof of his personal
religious creed.

In the beginning of the next century, after a short period of
anarchy, we find another Siladitya, Harsha-varddhana, seated
on the throne of Kanauj, and, during a prosperous reign of
about forty years (606-646), exercising supreme sway in that
country. It was during his reign that the Chinese pilgrim,
Hiuen Tsiang visited India, and gave a much more full and
graphic account of what he saw than his predecessor Fah Hian.
Nothing can be more characteristic of the state of religious
feeling, and the spirit of toleration then prevailing, than the fete
given by this king at Prayaga or Allahabad in the year 643,
at which the kings of Valabhi and Kamarupa (Asam) were
present. The first days of the festival were devoted to the
distribution, among the followers of Buddhism, of the treasures
accumulated during the previous five years, and then came the
turn of the Brahmans, who were treated with equal honour and
liberality ; then followed the fete of the other sects, among
whom the Jains appear conspicuous. All were feasted and

dasor inscriptions of A.D. 532-534,
describe Yarodharman as one who ruled
from the Brahmaputra to the western
ocean, and to whom even Mihirakula
paid ^homage. King Bhoja, the patron
of Kalidasa and others, may be the same

prince under a third name ; and Ballala-
mLrra says he reigned for fifty-five or
fifty-six years.

1 Conf. 'Journal des Savants,' Oct.
1905, pp. 534-548 ; and Beal, ' Buddhist
Records,' vol. ii. pp. 261, 267.

26 HISTORY OF INDIAN ARCHITECTURE.

feted, and sent away laden with gifts and mementoes of the
magnificence and liberality of the great king.

Pleasant as this picture is to look upon, it is evident that
such a state of affairs could hardly be stable, and it was in
vain to expect that peace could long be maintained between a
rising and ambitious sect, and one which was fast sinking into
decay ; apparently beneath the load of an overgrown priest-
hood. Accordingly we find that ten years after the death of
Harsha troubles supervened as prophesied,1 and the curtain
soon descends on the great drama of the history of northern
India, not to be raised again for nearly three centuries. It
is true, we can still follow the history of the Valabhis for some
time longer, and it would be satisfactory if we could fix the
date of their destruction with precision, as it was the event
which in the Hindu mind is considered the closing act of the
drama. If it was destroyed by a foreign enemy, it must have
been by the Moslim — perhaps by some expedition under
Amru ibn Jamal, the general of Hasham, ibn Amru al Taghlabi,
who was ruler of Sindh about 757 to 776.2 Valabhi was a
flourishing city in 640, when visited by Hiuen Tsiang, and
from that time, till the end of next century, the Moslims were
in such power on the Indus, and their historians tell us the
events of these years in such detail, that no other foreigner
could have crossed the river during that period. If it perished
by some internal revolution of convulsion, which is probable,
it only shared the fate that overtook all northern India about
this period. Strange to say, even the Moslims, then in the
plenitude of their power during the Khalifat of Baghdad,
retired from their Indian conquests, as if the seething cauldron
were too hot for even them to exist within its limits.

The more southern dynasty of the Western Chalukyas seem
to have retained their power down to about 757, and may, up
to that time, have exercised a partial sway to the north of the
Narbada, but after that we lose all sight of them for more than
two centuries till 973 when the dynasty was restored under
Taila II.; while, as a closing act in the great drama, the
' Rajatarangini ' boastfully represents the king of Kashmir —
Lalitaditya Muktapida (cir. 725-762) — as defeating Yajovarman
of Kanauj, conquering India from north to south, and subjecting
all the five kingdoms, into which it was nominally divided,
to his imperious sway.

We need not stop now to enquire whether this was exactly

1 'Vie et Voyages de Hiouen Thsang,' 'Elliot and Dowson's 'History of
trans, by Stanislas Julien, torn. i. p. 215 ; India,' vol. i. p. 444.
or Beal, ' Life of Hiuen Tsiang,' p. 156.

INTRODUCTION. 27

what happened or not. It is sufficient for present purposes to
know that about the middle of the 8th century a dark cloud
settled over the north of India, and that during the next two
centuries she was torn to pieces by internal troubles, which
have left nothing but negative evidence of their existence.
During that period the Rashtrakuta kings in the Dekhan,
having overthrown the Western Chalukyas, extended their
dominions from the Vindhyas to the Tungabhadra and
Krishna rivers, if not even farther south, waging war with
the Cholas. But after fully two centuries of successful domina-
tion, they were overthrown in 973 by the later Chalukyas
of Maharashtra. In the north were a number of smaller
kingdoms as the Chandellas in Bundelkhand, the Kalachuris
of Chedi, the Paramaras of Malwa, the Palas of Bengal, etc.

When light again appears in the middle of the loth century
the scene is wonderfully changed. Buddhism had practically
disappeared in the north and west at least, though it still
lingered on in Bengal, and Jainism had supplanted it in most
places ; but the mass of the people had become followers of
Vishnu or Siva. New dynasties had arisen which, though
they try to trace their lineage back to the troublous times
when Valabhi fell, were new to Indian history. Old India
had passed away, and the history of modern India was about
to open. The old dynasties had become extinct, and the
Rajput races were gaily stepping forward to assume their
places — too soon, alas ! to be engaged in a life or death
struggle with the most implacable foe to their race and religion
that India has ever known. It was a cruel Nemesis that
their victories over the Buddhists should soon have been
followed by the fatal siege of Somnath in 1025, and the
fight on the banks of the Ghaghar in 1193, which practically
laid India at the feet of the Moslim invader, and changed
the whole course of her subsequent career. But, as hinted
above, with the appearance of the Moslim on the scene, our
chronological difficulties cease, and the subject need not there-
fore be further pursued in this introduction.

IMMIGRATIONS.

From the above brief sketch of ancient Indian history it
may be gathered that it is doubtful whether we shall ever be
able to clothe with solid flesh the skeleton of history which is
all we possess anterior to the advent of Buddha. It is also
possible that pious frauds may have so confused the sequence
of events between his death and the rise of the Mauryas, that

28 HISTORY OF INDIAN ARCHITECTURE.

there will be great difficulty in restoring that period to any-
thing like completeness. But for the thousand years that
elapsed between "the revenge of Chanakya" and the fall of
Valabhi the materials are ample, and when sufficient industry
is applied to their elucidation there is little doubt that the
whole may be made clear and intelligible. It does not fall
within the scope of this work to attempt such a task ; but it
is necessary to endeavour to make its outlines clear, as, with-
out this being done, what follows will be utterly unintelligible ;
while, at the same time, one of the principal objects of this work
is to point out how the architecture, which is one important
branch of the evidence and the best aid we can have to the
teaching of history, may be brought to bear on the subject.

No direct evidence, however, derived only from events that
occurred in India itself, would suffice to make the phenomena
of her history clear, without taking into account the successive
migrations of tribes and peoples who, in all ages, so far as we
know, poured across the Indus from the westward to occupy
her fertile plains.

As mentioned above, the great master fact that explains
almost all we know of the ancient history of India is our know-
ledge that two thousand years or more before the birth of Christ
a Sanskrit-speaking nation migrated from the valleys of the
Oxus and Jaxartes. They crossed the Indus in such numbers
as to impress their civilisation and their language on the whole
of the north of India, and this to such an extent as practically
to obliterate, as far as history is concerned, the original
inhabitants of the valley of the Ganges, whoever they may
have been. At the time when this migration took place the
power and civilisation of Central Asia were concentrated on
the lower Euphrates, and the Babylonian empire never seems
to have extended across the Karmanian desert to the eastward.
The road, consequently, between Baktria and India was open,
and nations might pass and re-pass between the two countries
without fear of interruption from any other people.

If any of the ancient dynasties of Babylonia extended their
power towards the East, it was along the coast of Gedrosia,
and not in a north-easterly direction. It is, indeed, by no
means improbable, as hinted above, that the origin of the
Dravidians may be found among some of the Turanian peoples
who occupied southern and eastern Persia in ancient times,
and who may, either by sea or land, have passed to the
western shores of India. Till, however, further information is
available, this is mere speculation, though probably in the
direction in which truth may hereafter be found.

When the seat of power was moved northward to Nineveh,

INTRODUCTION. 29

the Assyrians seem to have occupied the country eastward of
the Caspian in sufficient force to prevent any further migration.
At least, after — say B.C. 1000 — we have no further trace of any
Aryan tribe crossing the Indus going eastward, and it seems
mainly to have been a consequence of this cutting off of the
supply of fresh blood that the purity of their race in India
was so far weakened as to admit of the Buddhist reform taking
root, and being adopted to the extent it afterwards attained.

During the period of the Akhaemenian sway (B.C. 558-334)
the Persians certainly occupied the countries about the Oxus
in sufficient strength to prevent any movement of the peoples.
So essentially indeed had Baktria and Sogdiana become parts
of the Persian empire, that Alexander was obliged to turn
aside from his direct route to conquer them, as well as the
rest of the kingdom of Darius, before advancing on India.

Whether it were founded for that purpose or not, the little
Greek kingdom of Baktria was sufficiently powerful, while it
lasted, to keep the barbarians in check ; but when, about or
after B.C. 160, the Yue-chi and other cognate tribes invaded
Sogdiana — driving out the Sakas, who next invaded Baktria,
and finally, about half a century later, the Yue-chi conquered
the whole of Baktria,1 they opened a new chapter in the history
of India, the effects of which are felt to the present day.

It is not yet quite clear how soon after the destruction of
the Baktrian kingdom these Turanian tribes conquered Kabul,
and occupied the country between that city and the Indus.
Certain it is, however, that they were firmly seated on the
banks of that river before the Christian Era, and under the
great king Kanishka of the Kushana tribe had become an
Indian power of very considerable importance. The date of
this king is, unfortunately, one of those puzzles that still remain
to be finally solved. It has been held that he was the founder
of the Saka. Era, A.D. 78, and that his reign must be placed
in the last quarter of the 1st century of our era.2 But this
era is only employed generally in the south and east ; and
it now seems almost certain that Kanishka's reign began in
B.C. 58 — the epoch of what was once known as the ' Malava
era,' and later as the 'Vikrama Samvat,' the reckoning in
common use in northern India.3

1 See Vivien de St Martin's ' Les Huns i Royal Asiatic Society,' 1907, p. 676.

blancs,' Paris, 1849; Franke, ' Beitrage I 2 Fergusson, in 'Journal of the Royal

aus chinesischen Quellen zur Kenntniss i Asiatic Society,' N.S. xii. pp. 259-285 ;

der Turkvolker und sky then Centralasien' Oldenberg in 'Indian Antiquary,' vol. x.

in'Abhandlungender konig. preussischen pp. 213-227.

Academic der Wissenschaften,' 1904 — 3 Fleet in ' Journal of the Royal Asiatic

summarised in 'Indian Antiquary,' vol. Society,' 1906, pp. 979-992.
xxxv. (1906), pp. 33-47 ; 'Journal of the

3o HISTORY OF INDIA ARCHITECTURE.

It also appears certain that the power of these Kushan
kings spread over the whole of the Panjab, and extended as
far at least as Mathura on the Jumna, before the commence-
ment of the Christian Era. Apparently the last of them was
Vasudeva, who ruled at least till A.D. 42. Soon after him we
meet with the name of a king Guduphara or Gondophernes,
which appears also in the legend of the Apostle Thomas :
an inscription of the 26th year of his reign is dated in the
iO3rd of the era, or A.D. 47 1 — when his rule must have
extended into the north of the Panjab. Next there followed
two (if not three) kings named Kadphises, who may have
ruled till the end of the century, after which northern India
was divided into separate kingdoms and tribal governments
till the rise of the Guptas in the 4th century.

Before the end of the first century another horde, known
to us only from coins and inscriptions in which they call
themselves Kshaharatas or Kshatrapas, occupied the whole of
the province of Gujarat ; one of the first of them — Nahapana,
for whom we have dates about A.D. 119 and in 124 — extended
his power over part of Malwa and the Nasik district. He was
overthrown by the Andhra king, Gautamiputra Satakarni, and
deprived of the districts south of the Narbada. Soon after,
we find another Kshatrapa, named Chashtana, ruling in Malwa,
and his successors founded a kingdom of their own. They date
their coins and inscriptions from the Saka Era, A.D. 78, and the
series extends from about 140 to 388 A.D. It thus happens
that this dynasty of Kshatrapas were only finally disposed of
by the rise of the Guptas.

The whole external history of northern India, from the time
of Kanishka to that of Ahmad Shah Durani (1761) is a narrative
of a continuous succession of tribes of Skythian origin, pouring
across the Upper Indus into India, each more Turanian than
the one that preceded it, till the whole culminated in the Mughal
conquest of India, in the I5th century, by a people as distinct
in blood from the Aryans as any that exist.

Of the older races, it seems probable that the Yavanas must
be distinguished from the Turanians. They were not Greeks,
though their name may be merely a mispronunciation of Ionian.
The term seems to have been applied by Indian authors to any
foreign race coming from the westward who did not belong to
one of the acknowledged kingdoms known to them. The
Kambojas seem to have been a people inhabiting the country
between Kandahar and Kabul, who, when the tide was setting

Grunwedel, ' Buddhist Art in India.' English ed. p, 84.

INTRODUCTION. 31

eastward, joined the crowd, and sought settlements in the more
fertile countries within the Indus.

The Sakas were well known to classical authors as the
Sacae, or Skythians. They were pressed on at first by the
Yue-chi, and became apparently most formidable during the
earlier centuries of the Christian Era.

Another important horde were the Ephthalites or White
Huns, who came into India apparently in the 5th century, and
one of whose kings, named Gollas, if we may trust Cosmas
Indicopleustes, was the head of a powerful state in northern
India, about the year 53O.1 They, too, seem to have been
conquered about the same time by the Hindus, and, as the
6akas, if not the Hiinas,2 were Buddhists, it may have been
their destruction that first weakened the cause of that religion,
and which led to its ultimate defeat a little more than a century
afterwards.

During the dark age, 750 to 950, we do not know of any
horde passing the Indus. The Muhammadans were probably
too strong on the frontier to admit of its being done, and after
that age they — and they only — conducted the various invasions
which completely changed the face and character of northern
India. For seven centuries they were continued, with only
occasional interruptions, and at last resulted in placing the
Muhammadan power supreme, practically, over the whole of
India, but only to fall to pieces like a house of cards, before the
touch of Western civilisation. All this, however, is written, and
written so distinctly, in so many books, that it need not be
recapitulated here.

SOUTHERN INDIA.

If the records of the ancient history of northern India are
unsatisfactory and untrustworthy, those of the southern part
of the peninsula are much more so. The Dravidians have no
ancient literature like that of the Vedas. They have no
traditions which point to any seat of their race out of India, or
of their having migrated from any country with whose inhabitants
they can claim any kindred. So far as they know, they are
indigenous and aboriginal. The utmost extent to which even
their traditions extend is to claim for their leading race of kings
— the Pandyas — a descent from Arjuna, one of the heroes of

1 'Christian Topography of Cosmas,'
translated by Dr. J. W. M'Crindle
(Hakluyt Soc.), pp. 370-371. This
Gollas seems probably the same as
Mihirakula or Mihiragula.

- We can hardly hope to discriminate
among these foreign invaders between
Hunas, Turushkas, .Sakas, Shahis,
Daivaputras, etc. , and may regard them
together as Indo-Skythians.

32 HISTORY OF INDIAN ARCHITECTURE.

the Mahabharata. He, it is said, when on his travels, married
a princess of the land, and she gave birth to the eponymous
hero of their race, and hence their name. But in later times all
the dynastic families got genealogies framed to trace their descent
from gods and early heroes. It is true, indeed, that they pro-
duce long lists of kings, which they pretend stretch back till
the times of the Pandavas. These were examined by the late
Professor Wilson in 1836, and he conjectured that they might
extend back to the 5th or 6th century before our era.1 But all that
has since come to light has tended to show that even this may
be an over-estimate of their antiquity. If, however, " the Choda,
Pada, and Keralaputra " of the second edict of Ajoka represent
the Cholas, Pandyas, Cheras, of more modern times, this triarchy
existed in the 3rd century B.C. In fact, all we really do know
is that, in classical times, there was a " Regio Pandionis " in the
country afterwards known as the Pandyan, kingdom of Madhura,
and it has been conjectured that the king who sent an embassy
to Augustus in B.C. 27 z was not a Porus, which would indicate
a northern race, but this very king of the south. Be this,
however, as it may, we do know, by the frequent mention of
this country by classical authors, that it was at least sufficiently
civilised in the early centuries of our era to carry on a consider-
able amount of commerce with the western nations, and there
is consequently no improbability that one or more powerful
dynasties may then have been established in the south. If one,
that dynasty was certainly the Pandyan. The Chola and the
Chera became important states only at a later date — preceded
by the Pallavas.

The discovery in 1892, by Mr. L. Rice, of a copy of the Ajoka
inscriptions so far south as Mysore, indicates that even in the
3rd century B.C., the Kanarese country was in communication
with, and subject to, the Maurya empire ; 3 and the civilisation
of the north must even then have penetrated into the south.

When we turn to their literature we find little to en-
courage any hope that we may penetrate further back into
their history than we have hitherto been able to do. Dr
Caldwell ascribes the oldest work in the Tamil, or any
southern language, to the 8th or pth century of our era,4
and it undoubtedly belongs to the Jains, who are originally a

1 'Journal of the Royal Asiatic
Society,' vol. iii. p. 202.

2 For an exhaustive description of this
subject see Priaulx, ' The Indian Travels
of Apollonius of Tyana, and the Indian
Embassies to Rome' (London, 1873),
pp. 65-87. We are now in a position to
prove a connection between the north of

India and Rome at that time. With
the south it seems to have been only
trade, but of this hereafter.

3 ' Epigraphia Carnatica,' vol. xi.
pp. if.

' Dravidian Grammar,' 2nd ed.
London, 1875, pp. 129 et seqq.

INTRODUCTION.

33

northern sect. According to the same authority, it was super-
seded by a Vaishnava literature about the I2th or I3th century,
and that again made way for one of Saiva tendency about the
latter date. There is no trace left of any Buddhist literature
in the south, and but little, consequently, that would enable us
to connect the history of the south with the chronology of
Ceylon or northern India, nor am I aware of the existence
of ancient Buddhist monuments south of the Krishna river
which would help us in this difficulty.1

Not having passed through Baktria, or having lived in
contact with any people making or using coins, the Dravidians
had none of their own, and consequently that source of informa-
tion is not available. Whatever hoards of ancient coins have
been found in the Madras Presidency have been of purely
Roman origin, brought there for the purpose of trade, and
buried to protect them from spoliation.

The inscriptions, which are literally innumerable all over the
Presidency, are the one source from which we can hope that
new light may be thrown on the history of the country, but,
with the exception of the edicts of A-roka found in Mysore,
none of these inscriptions hitherto brought to light go further
back than the 5th or 6th century, and it is not clear that
earlier ones may be found.2 It is, at all events, the most
hopeful field that lies open to future explorers in these dark
domains ; and, by the labours of epigraphists within the last
thirty years, most important light has been derived from them
for the mediaeval history of southern India. Those on the raths
of Mamallapuram and the caves at Badami, are in Sanskrit,
and consequently look more like an evidence of the northern
races pushing southward than of the southern races extending
their influence northward.

From a study of the architecture of the south we arrive at
the same conclusions as to the antiquity of Dravidian civilisation
that Dr Caldwell arrived at from a study of their literature.
The most important Buddhist monument yet discovered in the

1 The Buddhist tower at Negapattam,
destroyed in 1867, will be noticed in
Book I. chap. vi. p. 206.

a The Government of Mysore, with
laudable beneficence, employed Mr. L.
Rice with a staff of pandits for many
years, collecting and publishing the
inscriptions found in the state. The
results fill twelve volumes, forming the
'Epigraphia Carnatica' (1886-1905), and
when properly studied and analysed,
these must yield valuable results. For
Madras, Dr. E. Hultzsch was engaged
VOL. I.

in 1886 to collect the Tamil, Kanarese,
and Telugu inscriptions of the Pre-
sidency, and the results of his work
were published in six fasciculi of ' South
Indian Inscriptions' (3 vols. 1890-1903),
and, partly — with numerous Sanskrit
records from the other presidencies, — in
the 'Epigraphia Indica,' vol. iii. el stqq.
Previous to 1894, many Sanskrit and
Canarese inscriptions were published in
the 'Indian Antiquary' (vols. i.-xxiii.),
and in the ' Epigraphia Indica,' vols. i.
and ii.

34

HISTORY OF INDIAN ARCHITECTURE.

Presidency is that at Amaravati, on the Krishna ; and from that
vicinity northwards to Orissa there are remains showing that
there must have been flourishing communities there both of
Buddhists and Jains in early times. Whether the prevalence
of such structures in this region was due to a colony or settle-
ment formed by the northern Buddhists, at or near their port
of departure for Java and their eastern settlements, may be
doubted. The Andhras who ruled over the districts, were either
Buddhists or very liberal patrons of the sect. At Guntupalle in
the Godavari district have been found a group of rock-cut caves,
a structural chaitya, and a stupa, whilst at Chezarla, in the Nellor
district, another chaitya has been discovered almost entire,
though now used as a Hindu shrine.1 And remains of stupas
have been excavated in the Kistna district — at Jaggayyapeta,
Bhattiprolu, Gudivada, Guntupalle and Ghantarala ; 2 unfortun-
ately they have been utterly destroyed — some within living
memory.3

The rock-cut temples at Badami and Mamallapuram are the
works of Hindus in the 6th and /th centuries, and the structural
temples of Kailasanath and Vaikunthaperumal at Conjivaram are
of nearly the same age, and, with some others, they help materially
to illustrate the history of the style till the 8th century. From
that time forward their building activity was enormous. The
style culminated in the i6th and I7th centuries, to perish in the
1 8th.

When the history of the south does acquire something like
consistency it takes the form of a triarchy of small states. The
eldest and most important, that of Madura — so called after
Mathura (or Muttra) on the Jamna — was also the most civilised,
and continued longest as a united and independent kingdom.

The Cholas rose into power on the banks of the Kaveri, and to
the northward of it, about the year 1000, though no doubt they
existed as a small state about Conjivaram for some centuries
before that time. The third, the Chera, were located on the
west coast, extending from the Tulu country southwards, and
including Malabar and most of Travankor. Tradition assigns
to them a dynasty of kings called Perumals which ended in the
9th century. Chola and Chalukya inscriptions speak of their
being frequently defeated, but we have no inscriptions of any

1 See below Book I. chap. v. p. 166.
These very interesting structures were
surveyed several years ago, but the results
have not yet been fully published. The
caves are Buddhist of an early type.

2 'South Indian Buddhist Antiquities,'
by A. Rea, 1894 ; ' Epigraphia Indica,'
vol. ii. pp. 323-329.

3 The Gudivada and Bhattiprolu stupas,
were demolished by the Public Works
officers about thirty-six years ago, for
bricks to use in road-making, and the
marbles of the latter were built into the
walls and floor of the Vellatur sluice,
or burnt for lime. — ' Madras Government
Orders,' No. 1620, of 1st Nov. 1878.

INTRODUCTION. 35

Chera sovereign and, as yet, know very little for certain of their
history. With the other southern states, they were, however,
superseded, first by the Cholas, about A.D. 1000, and finally
eclipsed by the Hoysala Ballalas, a century or so afterwards.
These last became the paramount power in the south, till their
capital — Halebid — was taken, and their dynasty destroyed by
the Muhammadans, in the year 1310.

With the appearance of the Muhammadans on the scene the
difficulties of Indian chronology disappear in the south, as well
as in the north. From that time forward the history of India is
found in such works as those written by Firishta or Abul Fazl,
and has been abstracted and condensed in numerous books in
almost every European language. There are still, it must be
confessed, slight discrepancies and difficulties about the sequence
of some events in the history of the native principalities.1 These,
however, are not of such importance as at all to affect, much less
to invalidate, any reasoning that may be put forward regarding
the history or affinities of any buildings, and this is the class
of evidence which principally concerns what is written in the
following pages.

SCULPTURE.

In order to render the subject treated of in the following
pages quite complete, it ought, no doubt, to be preceded by an
introduction describing first the sculpture and then the mythology
of the Hindus in so far as they are at present known to us. There
are in fact few works connected with this subject more wanted at
the present day than a good treatise on these subjects. When
Major Moor published the ' Hindu Pantheon ' in 1810, the subject
was comparatively new, and the materials did not exist in this
country for a full and satisfactory illustration of it in all its
branches. When, in 1832, Coleman published his 'Mythology
of the Hindus,' he was enabled from the more recent researches
of Colebrooke and Wilson, to improve the text considerably, but
his illustrations are very inferior to those of his predecessor.
Moor chose his from such bronzes or marbles as existed in our
museums, and from an important private collection he formed
principally in western India.2 Coleman's were generally taken
from modern drawings, or the tawdry plaster images made for
the Durga puja of Bengali Babus.3 By the aid of photography

Much information on the history of : descendants in Suffolk,
these states will be found in Elliot and j 3 Similarly the small work — ' Hindu

Dowson's 'History of India,' and in
other recent works.

2 His collection, brought from Bombay
a century ago, is still preserved by his

Mythology, Vedicand Puranic,' by W. J.
Wilkins (Calcutta, 1893), is illustrated
solely from modern bazar pictures.

36 HISTORY OF INDIAN ARCHITECTURE.

any one now attempting the task would be able to select perfectly
authentic examples from Hindu temples of the best age. If this
were done judiciously, and the examples carefully reproduced,
it would not only afford a more satisfactory illustration of the
mythology of the Hindus than has yet been given to the public,
but it might also be made a history of the art of sculpture in
India, in all the ages in which it is known to us.

From its very nature, it is evident that sculpture can hardly
ever be so important as architecture as an illustration of the
progress of the arts, or the affinities of nations. Tied down to
the reproduction of the immutable human figure, sculpture
hardly admits of the same variety, or the same development, as
such an art as architecture, whose business it is to administer to
all the varied wants of mankind, and to express the multifarious
aspirations of the human mind. Yet sculpture has a history,
and one that can at times convey its meaning with considerable
distinctness. No one, for instance, can take up such a book as
that of Cicognara,1 and follow the gradual development of the
art as he describes it, from the first rude carvings of the
Byzantine school, till it returned in the present day to the
mechanical perfection of the old Greek art, though without its
ennobling spirit, and not feel that he has before him a fairly
distinct illustration of the progress of the human mind during
that period. Sculpture in India may fairly claim to rank, in
power of expression, with mediaeval sculpture in Europe, and to
tell its tale of rise and decay with equal distinctness ; but it is
also interesting as having that curious Indian peculiarity of being
written in decay. The story that Cicognara tells is one of
steady forward progress towards higher aims and better
execution. The Indian story is that of backward decline, from
the sculptures of the Bharaut and Amaravati topes, to the
illustrations of Coleman's or Wilkins's ' Hindu Mythology.'

When Hindu sculpture first dawns upon us in the rails at
Bodh-Gaya, and Bharaut B.C. 200 to 250, it is thoroughly original,
almost without a trace of foreign influence, but quite capable of
expressing its ideas, and of telling its story with a distinctness
that never was surpassed — at least in India. Some animals, such
as elephants, deer, and monkeys, are better represented there
than in any sculpture known in any part of the world ; so, too,
are some trees, and the architectural details are cut with an
elegance and precision which are very admirable. The human
figures, too, though very different from our standard of beauty
and grace, are truthful to nature, and, where grouped together,
combine to express the action intended with singular felicity.

1 ' Storia della Scultura, dal suo risorgimento in Italia sino al seculo di Napoleone.'
Venezia, 1813.

INTRODUCTION.

37

For an honest purpose-like pre-Raphaelite kind of art, there is
probably nothing much better to be found elsewhere.

The art had apparently begun to decline when the gateways
at Sanchi were executed somewhat later ; but whether this was
not mainly due to the more refractory character of the stone,
and a different school of workmen, it is hard to say. They may
then have gained a little in breadth of treatment, but it had lost
in delicacy and precision. Its downward progress was, how-
ever, arrested, apparently by the rise, in the extreme north-west
of India, of a school of sculpture strongly impregnated with the
traditions of classical art. The Graeco-Baktrians, driven out by
the Yue-chi, continued to hold some sort of domination in
Afghanistan till not very long before our era, and a vast inter-
change of ideas was, at that period, carried on between the east
and west by means of newly - opened highways. Thus Greek
models and art became familiar, and when once a demand arose
for such workmanship, a school of art would appear. For the
present it is sufficient to know that a quasi-classical school of
sculpture did exist in the Panjab, and to the west of the Indus
during the first four centuries after Christ, and it can hardly
have flourished there so long, without its presence being felt
in India.1

Its effects were certainly apparent at Amaravati in the 1st
and 2nd centuries, where a school of sculpture was developed,
partaking of the characteristics of both those of Central India
and of the west. Though it may, in some respects, be inferior
to either of the parent styles, the degree of perfection reached
by the art of sculpture at Amaravati may probably be considered
as the culminating point attained by that art in India.

When we meet it again in the early Hindu temples, and
later Buddhist caves, it has lost much of its higher aesthetic and
phonetic qualities, and frequently resorts to such expedients as
giving dignity to the principal personages by making them
double the size of less important characters, and of distinguish-
ing gods from men by giving them more heads and arms than
mortal man can use or understand.

All this is developed, it must be confessed, with considerable
vigour and richness of effect in the temples of Orissa and of
Mysore, down to the I3th or I4th century. After that, in the
north it was checked by the presence of the Moslims ; but, in
the south, some of the most remarkable groups and statues —
and they are very remarkable — were executed after this time,

1 For some account of Buddhist art in
Gandhara and of early Indian sculpture,
see Grunwedel's ' Buddhist Art in India,'
Eng. translation (Quaritch, 1901). A

work on 'Indian Sculpture and Painting,'
by Mr E. B. Havell, has recently been
published by Mr Murray.

38 HISTORY OF INDIAN ARCHITECTURE.

and continued to be executed, in considerable perfection, down
to the middle of the i8th century.

As we shall see in the sequel, the art of architecture
continues to be practised with considerable success in parts of
India remote from European influence ; so much so, that it
requires a practised eye to discriminate between what is new
and what is old. But the moment any figures are introduced,
especially if in action, the illusion vanishes. No mistake is then
possible, for the veriest novice can see how painfully low the
art of sculpture has fallen. Were it not for this, some of the
modern temples in Gujarat and Central India are worthy to
rank with those of past centuries ; but their paintings and their
sculptured decorations excite only feelings of dismay, and lead
one to despair of true art being ever again revived in the East.

To those who are familiar with the principles on which
these arts are practised, the cause of this difference is obvious
enough. Architecture being a technic art, its forms may be
handed down traditionally, and its principles practised almost
mechanically. The higher phonetic arts, however, of sculpture
and painting admit of no such mechanical treatment. They
require individual excellence, and a higher class of intellectual
power of expression, to ensure their successful development.
Architecture may, consequently, linger on amidst much
political decay ; but, like literature, the phonetic arts can only
be successfully cultivated where a higher moral and intellectual
standard prevails than, it is feared, is at present to be found in
India.

MYTHOLOGY.

Whenever any one will seriously undertake to write the
history of sculpture in India, he will find the materials abundant
and the sequence by no means difficult to follow ; but, with
regard to mythology, the case is different. It cannot, however,
be said that the materials are not abundant for this branch of
the enquiry also ; but they are of a much less tangible or
satisfactory nature, and have become so entangled, that it is
extremely difficult to obtain any clear ideas regarding them ;
and it is to be feared they must remain so, until those who
investigate the subject will condescend to study the architecture
and the sculpture of the country as well as its books. The
latter contain a good deal, but they do not contain all the
information available on the subject, and they require to be
steadied and confirmed by what is built or carved, which alone
can give precision and substance to what is written.

It is remarkable that, with all the present day activity
in every branch of Sanskrit research, so very little has been
done for the illustration of mythology, which is so intimately

INTRODUCTION.

39

connected with the whole literature. It would be a legitimate
part of the duty of the Archaeological Surveys to collect materials
on a systematic plan for this object ; and the production of illustra-
tions has now become so easy and inexpensive that photographs
from original materials of a satisfactory class might readily
be published to supply this most pressing desideratum. The
details of the emblems and symbols of the numerous divinities
of the pantheon could also be collected, along with the delinea-
tions, by those familiar with such symbols. All this could
easily be accomplished, and it is consequently hoped it may
before long be attempted.1

Much of the confusion of ideas that prevails on this subject
no doubt arises from the exaggerated importance it has been
the fashion to ascribe to the Vedas, as explaining everything
connected with the mythology of the Hindus. It would, indeed,
be impossible to over-estimate the value of these writings from
a philological or ethnological point of view. Their discovery
and elaboration have revolutionised our ideas as to the migra-
tions of races in the remote ages of antiquity, and established
the affiliation of the Aryan races on a basis that seems absolutely
unassailable ; but it cannot be too strongly insisted upon that
the Aryans are a race of strangers in India, distinct from the
Indian peoples themselves. They may, as hinted above, have
come into India some three thousand years before Christ, and
may have retained their purity of blood and faith for many
generations ; but with the beginning of the political Kaliyug
— or, to speak more correctly, at the time of the events detailed
in the Mahabharata, say 1200 years B.C. — they had lost much
of both ; while every successive wave of immigration that has
crossed the Indus during the last three thousand years has
impaired the purity of their race. From this cause, and from
their admixture with the aborigines, it may probably be with
confidence asserted that there is not now five per cent. — perhaps
not one — of pure Aryan blood in the present population of
India, nor, consequently, does the religion of the Vedas constitute
one-twentieth part of the present religion of the people.2

With the Vedas, however, we have very little to do in the
present work. The worship they foreshadow is of a class too
purely intellectual to require the assistance of the stonemason
and the carver to give it expression. The worship of the

1 Numerous excellent illustrations exist
among the materials already accumulated
by the Archseological Surveys in Southern
and Western India and in the Calcutta
Museum — but, at present, there seems
little prospect of their publication.

? For the mythology of the Vedas, see

Professor A. A. Macdonell's ' Vedic
Mythology ' in the ' Grundriss der indo-
arischen Philologie und Altertumskunde,'
in Bd. iii. A useful general handbook
is Dowson's ' Dictionary of Hindu
Mythology, '

4o

HISTORY OF INDIAN ARCHITECTURE.

Aryans was addressed to the sun and moon ; the firmament
and all its hosts ; the rain-bearing cloud ; the sun-ushering
dawn. All that was beautiful in the heavens above or beneficent
on earth, was sung by them in hymns of elevated praise, and
addressed in terms of awe or endearment as fear or hope pre-
vailed in the bosom of the worshipper.1 Had this gone on for
some time longer than it did, the objects worshipped by the
Aryans in India might have become imaged as gods, like those
of Greece and Rome, endowed with all the feelings and all the
failings of humanity. In India it was otherwise ; the deities
were dethroned, but were not degraded. There is no trace in
Vedic times, so far as at present known, of Indra or Varuna, of
Agni or Ushas, being represented in wood or stone, or of their
requiring houses or temples to shelter them. It is true indeed
that the terms of endearment in which they are addressed are
frequently such as mortals use in speaking of each other ; but
how otherwise can man express his feeling of love or fear, or
address his supplication to the being whose assistance he
implores ?

The great beauty of the Veda is, that it stops short before
the powers of nature are dwarfed into human forms, and when
every man stood independently by himself, and sought through
the intervention of all that was great or glorious on the earth,
or in the skies, to approach the great spirit that is beyond and
above all created things.

Had the Aryans been a numerical majority in India, and
able to preserve their blood and caste in tolerable purity, the
religion of India could hardly ever have sunk so low as it did,
though it might have fallen below the standard of the Veda.
What really destroyed it was, that each succeeding immigration
of less pure Aryan or of Turanian races rendered their numerical
majority relatively less and less, while their inevitable influence
so educated the subject races as to render their moral majority
even less important. These processes went on steadily and
uninterruptedly till, in the time of Buddha, the native religions
rose fairly to an equality with that of the Aryans, and after-
wards for a while eclipsed it. The Vedas were only ultimately
saved from absolute annihilation in India, by being connected
with the Vaishnava and »Saiva superstitions, where their
inanimate forms may still be recognised, but painfully degraded
from their primitive elevation.

When we turn from the Vedas, and try to investigate the
origin of those religions that finally absorbed the Vedas in

1 The ritual of the Veda is chiefly, if not
wholly, addressed to the elements, particu-
larly to fire. — H. H. Wilson, 'Asiatic Re-

searches,' vol. xvii. p. 194 ; ibid., p. 614.
Conf. Bergaigne, 'La Religion Vedique' ;
Oldenburg, ' Die Religion des Veda, '

INTRODUCTION.

their abominations, we find our means of information painfully
scanty and unsatisfactory. As will appear in the sequel, all that
was written in India that is worth reading was written by the
Aryans ; what was built was built by Turanians and Dravidians.
But the known buildings extend back only to the 3rd century
B.C., while the books may be ten centuries earlier, and, as might
be expected, it is only accidentally and in the most contemptuous
terms that the proud Aryans even allude to the abject Dasyus
or their religion. What, therefore, we practically know of them
is little more than inferences drawn from results, and from what
we now see passing in India.

Notwithstanding the admitted imperfection of materials, it
seems to be becoming more and more evident, that we have
in the north of India one great group of native religions,
which we know in their latest developments as the Buddhist,
Jaina, and Vaishnava religions. The first named we only
know as it was taught by vSakyamuni before his death about
480 B.C., but no one I presume supposes that he was the
first to invent that form of belief, or that it was not based
on some preceding forms. The Buddhists themselves, accord-
ing to the shortest calculation, admit of four preceding Buddhas
— according to the more formal accounts, of twenty-four. A
place is assigned to each of these, where he was born, and
where he died, the father and mother's name is recorded, and
the name, too, of the Bodhi - tree under whose shade he
attained Buddhahood. The dates assigned to each of these
are childishly fabulous, but they may have been real personages,
whose dates extended back to a very remote antiquity.1

The Jains, in like manner, claim the existence of twenty-
four Tirthankars, including Mahavira the last. Their places
of birth and death, ages and numbers of converts, are equally
recorded, all are in northern India, though little else is told
of them ; but, from their fabulous ages, stature, and the
immeasurable periods of the past when they are said to have
lived — they can only be looked on as purely fabulous. The
series ends with Mahavira, who was the contemporary of
vSakyamuni, and is said to have died before him at Pawa
in Bihar.

The Vaishnava series is shorter, consisting of only ten
Avatars ; but it, too, closes at the same time, Buddha himself
being the ninth, whilst the last is yet to come. Its fifth

1 A list of the twenty-four Buddhas,
with these particulars, is given in the
introduction to Tumour's ' Mahawansa,'
introd. p. 32. See also Spence Hardy's
' Manual of Budhism,' 2nd. ed. pp. 96ff.

Representations of six or seven of their
Bodhi-trees, with the names attached,
have been found at Bharaut and AjantS,
showing at least that more than four
were recognised,

HISTORY OF INDIAN ARCHITECTURE.

Avatar takes us back to Rama, who, if our chronology is
correct, may have lived B.C. 2000 ; the fourth — Narasimha, or the
man lion — may possibly point to the time the Aryans entered
India. The three first deal with creation and events anterior
to man's appearance on earth. In this respect the Vaishnava
list differs from the other two. They only record the exist-
ence of men who attained greatness by the practice of virtue,
and immortality by teaching the ways of emancipation from
rebirths. The Vaishnavas brought their god to earth, to mix
and interfere in mundane affairs in a manner that neither the
earlier Aryan nor the Buddhist dreamt of, and so degraded the
earlier religion of India into the monstrous system of idolatry
that now prevails in that country.

No attempt, so far as I know, has been made to explain
the origin of the .Saiva religion ; it was, however, most probably
an aboriginal superstition assimilated by the Brahmans. The
earliest authentic written allusion to it seems to be that of the
Indian ambassador to Bardisanes (cir. A.D. 220), who described a
cave in the north of India which contained an image of a god,
half-man, half-woman.1 This is beyond doubt the Ardhanama
form of Siva, so familiar afterwards at Elephanta and in every
part of India. The earliest engraved representations of this
god seem to be those on the coins of Kadphises II. (about 80 to
90 A.D.), where the figure with the trident and the Bull certainly
prefigure the principal personage in this religion.2 Besides all
this, it seems now tolerably well ascertained, that the practice
of endowing gods with a multiplicity of limbs took a much
greater development in Tibet and the trans-Himalayan countries
than in India, and that the wildest Tantric forms of Durga and
other divinities or demons are more common and more developed
in Nepal and Tibet than they are even in India Proper.3 If
this is so, it seems pretty clear, as the evidence now stands,
that Saivism is an aboriginal or northern superstition — possibly
introduced into India by some of the northern hordes who
migrated into India long before the Christian Era.

It is also only too true that no attempt has yet been made
to ascertain what the religion of the Dravidians was before
they adopted either the Jaina or the Vaishnava or Saiva forms
of religion. It is possible that among the Pdndu Kallus, and
other forms of ' Rude Stone Monuments ' that are found every -

1 Stobseus, ' Physica,' Gaisford's ed.
p. 54 ; see also Priaulx, ' India and
Rome,' p. 153; Burgess, 'Rock-Temples
of Elephanta,' 8vo ed. p. 67.

8 Wilson's ' Ariana Antiqua,' plates 10,
1 1 ; P. Gardner's ' Coins of the Greek
a.nd Scythic Kings of Bactria and

India,' pp. 124-128, plate 25, and
introd. p. 50.

3 Compare Griinwedel's ' Mythologie
des Buddhismus in Tibet and der
Mongolei. ' Buddhist figures have been
subjected to the same treatment as the
gods : to make them demoniac,

INTRODUCTION.

43

where in the south, we may find the fossil remains of the old
Dravidian religion before they adopted that of the Hindus.
These monuments, however, have not been examined with
anything like the care requisite for the solution of a complex
problem like this, and till it is done we must rest content with
our ignorance.1

In the north we have been somewhat more fortunate, and
enough is now known to make it clear that, so soon as
enquirers apply themselves earnestly to the task, we may
know enough to make the general outline at least tolerably
clear. When I first published my work on ' Tree and Serpent
Worship,' in 1869, no one suspected, at least no one had
hinted in type, that such a form of religion existed in Bengal.
Since that time, however, so much has been written on the
subject, and proofs have accumulated with such rapidity, that
few will now be bold enough to deny that Trees were
worshipped in India in the earliest times, and that a Naga
people did exist, especially in the north - west, who had a
strange veneration for snakes. In the Buddhist legends,
Buddha is constantly represented as converting Nagas, and
whilst a superhuman character is ascribed to them, they
doubtless represent people of Turanian descent.2 Further,
snake worship is prevalent still, especially among the lower
castes, and, though to a less extent, yet somehow connected
with it, is the veneration of trees.3 It is also quite certain that
underlying Buddhism we everywhere find evidence of a stratum
of Tree and Serpent Worship. Sometimes it may be repressed
and obscured, but at others it crops up again, and, to a certain
extent, the worship of the Tree and the Serpent, at some
times and in certain places, almost supersedes that of the
founder of the religion himself.

The five, or seven, or thousand-headed Naga is everywhere

1 A book was published in 1873 by
the late Mr. Breeks, of the Madras
Civil Service, on the Primitive -Tribes
of the Nilagiris, which gives a fuller
account of these "rude stone monu-
ments" than any other yet given to
the public. It can hardly, however, be
accepted as a solution of the problem,
which requires a wider survey than he
was able to make. See also Fergusson's
'Rude Stone Monuments' (1872), pp.

455-499-

'2 The Nagbansis of Chutia Nagpur, who
appear to have come from about Gorakh-
pur in northern Bihar, are evidently of
an early Naga or snake - worshipping

3 In Malabar, " a clump of wild jungle
trees luxuriantly festooned with graceful
creepers is usually found in the south-
west corner of the gardens of all respect-
able Malayali Hindus. The spot is left
free to nature to deal with as she likes ;
every tree and bush, every branch and
twig is sacred. This is the 'viskdttnm
kdvu ' (poison shrine) or ' ndga kotta '
(snake shrine). Usually there is a
granite stone (chittra kuta ka/lu) carved
after the fashion of a cobra's hood set up
and consecrated in this waste spot." —
Logan's ' Malabar,' vol. i. p. 183. For
some account of Trees worshipped in
Western India, see Bombay Gazetteer,
vol. ix. pt. i. pp. 382-388,

44

HISTORY OF INDIAN ARCHITECTURE.

present in the temples of the Jains, and pervades the whole
religion of the Vaishnavas.1 In the great act of creation the
Naga performs the principal part in the churning of the ocean,
and in almost every representation of Vishnu he appears either
as supporting and watching over him, or as performing some
subsidiary part in the scene. It is, in fact, the Naga that
binds together and gives unity to this great group of religions,
and it is the presence of the Tree and Serpent worship under-
lying Buddhism, Jainism, and Vaishnavism that seems to prove
almost incontestably that there existed a people in the north
of India, whether we call them Dasyus, Kolarians, or by any
other name, who were Tree and Serpent worshippers, before
they adopted any of the later Hindu forms of faith. Nothing
can be more antagonistic to the thoughts and feelings of a pure
Aryan race than such forms of worship, and nothing more
completely ante-Vedic than its rites.2 We seem, then, almost
forced to assume that it was an aboriginal superstition in the
north of India, and it was the conversion of the people to
whom it belonged that gave rise to that triarchy of religions
that have competed with each other in the north during the
last two thousand years.

This solution of the difficulty has the further advantage that
it steps in at once clearly to explain what philology is only
dimly guessing at, though its whole tendency, as well as
that of ethnology, now seems in the same direction. If this
view of the mythology be correct, it seems certain that there
existed in the north of India, before the arrival of the Aryans,
a people whose affinities were all with the Tibetans, Burmese,
Siamese, and other trans - Himalayan populations, and who
were not Dravidians, though they may have been intimately
connected with one division at least of the inhabitants of
Ceylon.

Both the pre- Aryan races of India belonged to the

1 Snake worship may have been intro-
duced into the south from the north ;
and it has been remarked that snake
images are very frequent about Jaina
temples in Mysore and Kanara. At
Negapatam is a temple dedicated to
Naganath, and at Subrahmanya in South
Kanara, at Nagarkoil, at Manarchal in
Travankor, and elsewhere, are also snake
temples much resorted to. No Brahman
ever officiates in a Naga temple. See
also Thurston, ' Ethnographic Notes in
Southern India,' 1906, pp. 283-293.

2 Though .Siva is always represented
with a black snake as one of his symbols

there does not seem to be any very close
connection between Snake worship and
.Saivism, though there are some coinci-
dences that may point that way ; in
Kanara, Naga images are set up facing
the east, under the shade of two pipal
trees — a male and female growing
together and married with proper rites.
Beside them grow a margosa and bilva
tree as witnesses ; now these latter trees
are more or less consecrated to Siva. On
the other hand no trace of Tree-worship
seems to be mingled with the various
forms of adoration paid to this divinity.
The tulasi or basil is sacred to Vishnu.

INTRODUCTION. 45

Turanian group ; but, as hinted above, the Dravidians belonged
to a different and more westerly branch of that great family
of mankind.

These, however, are speculations which hardly admit of
proof in the present state of our knowledge, and would con-
sequently be quite out of place here, were it not that some
such theory seems indispensable to explain the phenomena of
the architectural history of India. That of the north is so
essentially different from that of the south that they cannot
possibly belong to the same people. Neither of them are
Aryan ; and unless we admit that the two divisions of the
country were occupied by people essentially different in blood,
though still belonging to the building races of mankind, we
cannot possibly understand how they always practised, and
to the present employ, styles so essentially different. Until
these various ethnographical and mythological problems are
understood and appreciated, the styles of architecture in
India seem a chaos without purpose or meaning. Once,
however, they are grasped and applied, their history assumes
a dignity and importance far greater than is due to any
merely aesthetic merits they may possess. Even that, how-
ever, is in many respects remarkable, and, when combined
with the scientific value of the styles, seems to render them
as worthy of study as those of any other people with whose
arts we are acquainted.

STATISTICS.

It would add very much to the clearness of what follows
if it were possible to compile any statistical tables which would
represent with anything like precision the mode in which the
people of India are distributed, either as regards their religious
beliefs or their ethnographical relations. The late census of
1901 has afforded a mass of material for this purpose, but the
information is distributed through some thirty folio volumes,
in such a manner as to make it difficult to abstract what is
wanted so as to render it intelligible to the general reader.
Even, however, if this were done, the result would hardly, for
several reasons, be satisfactory. The uneducated masses have
hazy ideas even with regard to their religion, and can hardly
be expected to know to which of the larger sections of Hinduism
their particular sect belongs. Hence, in the tables we are given
the enumeration of the members of numerous Hindu sects, but
not classified under Saiva, Vaishnava, etc., though these larger
groups are the most interesting for us.

46 HISTORY OF INDIAN ARCHITECTURE.

The census results for all India may be briefly stated as
follows :

British Provinces in India . 221,409,000
Burma .... 10,490,000

Native States . . . 62,462,000

TOTAL . . 294,361,000

Notwithstanding difficulties or defects, it may be useful to
state here that the population of the whole of India, inclusive
of Burma — when arranged by religions — was found to stand
as follows :

Hindus of all sects . . 207,147,000, or 7-1 oths

Musalmans . . . 62,458,000, or fully i-5th

Buddhists — mostly in Burma 9,477,000, scarcely i-soth

Primitive or Animistic . 8,584,000, about i-34th

Christians .... 2,923,200, about i-ggth

Sikhs .... 2,195,300, about i-i34th

Jains 1,334,200, about r-22Oth

Parsis, Jews, and others . 242,300, about i-i2ooth

The tables of this census also afford us some information
with regard to the distribution of races among the people,
though it could not be expected that the ethnological survey
should yet be able to organise a satisfactory census of the
races, though the distribution of languages helps us somewhat.
Here, however, it is to be borne in mind that, especially in
northern India, many aboriginal or non-Aryan tribes have
changed their language for one of the Sanskritic family spoken
by their neighbours. Hence we must regard the Indo- Aryan
group of languages as including a vast number of people of
Turanian and mixed descent. The tables show that upwards
of 221,000,000 speak Indo- Aryan tongues — including Baloch,
Pashtu, Marathi, Bengali, etc. — that is about three-fourths of
the entire population, whilst close on 60,000,000 or a fifth are
Dravidians and Kolarians, Gonds, Brahuis, etc., which the
census has grouped together as a Dravida - Munda group.
Lastly, the Indo-Chinese and Malayan, including Tibetan,
Burmese, etc., number 11,720,000, or about four per cent, of the
whole.

The first linguistic group includes, of course, the Muham-
madans ; but we know that many of the Moslims of India
were recruited from slaves purchased and brought up in the
creed of their masters. In Bengal especially, where they are
most numerous, they are Bengalis pure and simple, many
perhaps most, of whom have adopted that creed quite recently
from motives it is not difficult to understand or explain.

INTRODUCTION. 47

Though there may consequently be 62,458,000 of Musalmans
in India at the present day, we may feel quite certain that not
one-half of this number are immigrants or the descendants
of emigrants who entered India during the last eight centuries.

The same is probably true of the Turanian races, who
entered India in the first ten centuries after our era. It is
scarcely probable that they were sufficiently numerous to be
the progenitors of thirty millions of people, and, if they were
so, the mothers, in nine cases out of ten, were most probably
natives of India.

Of the Aryans we know less ; but, if so great a number
as forty millions can trace anything like a direct descent from
them at the present day, the amount of pure Aryan blood in
their veins must be infinitesimally small. Yet, though their
blood may be diluted, the influence of their intellect remains
so powerfully impressed on every institution of the country
that, had they perished altogether,, their previous presence is
still an element of the utmost importance in the ethnic relations
of the land.

Another census may enable us to speak with still more
precision with regard to these various divisions of the mass
of the people of Hindustan, but meanwhile the element that
seems to be most important, though the least investigated
hitherto, is the extent of the aboriginal race. It has been
so overlooked, that putting it at a hundred millions may seem
an exaggeration. Its intellectual inferiority has kept it in the
background, but its presence everywhere seems to me the only
means of explaining most of the phenomena we meet con-
tinually, especially those connected with the history of the
architecture of the country. Except on some such hypothesis
as that just shadowed forth, I do not know how we are to
account for the presence of certain local forms of buildings we
find in the north, or to explain the persistence with which
they were adhered to.

When from these purely ethnographic speculations we turn
to ask how far religion and race coincide, we are left with
still less information of a reliable character. As a rule, the
Dravidians are 6aivas, and .Saiva in the exact proportion of
the purity of their blood. In other words, in the extreme
south of India they are immensely in the majority. In some
districts of the Madras Presidency they are as 6 or 7 to I of
the followers of Vishnu, and generally in the south, 2 to I ;
but as we proceed northward they become equal, and in some
of the northern districts of the Madras Presidency the pro-
portions are reversed.

In Bengal, and wherever Buddhism once prevailed, the

HISTORY OF INDIAN ARCHITECTURE.

Vaishnava sects are, as might be expected, the most numerous.
Indeed if it were not that so much of the present Hindu
religion is an importation into the south, and was taught to
the Dravidians by Brahmans from the north, it would be
difficult to understand how the Vaishnava religion ever took
root there, except in succession to Buddhism itself, which
existed to a considerable extent, but where it, too, was an
importation. If, however, it be correct to assume that Saivism
had its origin to the northward of the Himalayas, among the
Tartar tribes of these regions, there is no difficulty in under-
standing its presence in Bengal to the extent to which it is
found to prevail there. But, on the other hand, nothing can
be more natural than that an aboriginal Naga people, who
worshipped trees and serpents, should become Buddhists, as
Buddhism was originally understood, and, being Buddhists,
should slide downwards into the corruptions of the present
Vaishnava form of cult, which is that most fashionable and
prevalent in the north of India.

One of the most startling facts brought out by the census,
is that about one-third of the population of Eastern Bengal
are Muhammadan — 25,500,000, out of 74,750,000 — while in the
United Provinces the Muhammadans are scarcely more than
i-6th — 4,567,000 among 25,430,000; and in Madras little more
than i-i5th. It thus looks more like a matter of feeling than
of race ; it seems that as the inhabitants of Bengal were
Buddhists, and clung to that faith long after it had dis-
appeared in other parts of India, they came in contact with
the Moslim religion before they had adopted the modern form
of Vaishnavism, and naturally preferred a faith which acknow-
ledged no caste, and freed them from the exactions and tyranny
of a dominant priesthood. The Muhammadan religion is in
fact much more like Buddhism than are any of the modern
Hindu forms, and when this non-Aryan casteless population
came in contact with it and they were free to choose, after the
mysterious evaporation of their old beliefs, they adopted the
religion most resembling that in which they had been brought
up. It is only in this way that it seems possible to account
for the predominance of the Moslim religion in Lower Bengal l
and in the Panjab, where the followers of the Prophet out-
number the Hindus, in the proportion of 3 to 2, or as 14,000,000
in a population of 20,300,000.

1 In Bihar and West Bengal, the
Muhammadans number 4,050,000, or
less than 14 per cent, of the population,
whilst in Central, North, and East
Bengal, they number 20,870,000 or

quite 60 per cent., and in East Bengal
alone, there are 66 per cent, of Moslims
or 11,220,000 ; in several of the districts
they form quite three-fourths of the
population.

INTRODUCTION.

49

Where ^Saivism held its place or crept in was apparently
among those races who were Dravidians or Turanians, or had
affinities with the Tartar races, who immigrated from the north
long before the Muhammadan conquest.

To most people these may appear as rash generalisations,
and at the present stage of the enquiry would be so in reality,
if no further proof could be afforded. After reading the follow-
ing pages, I trust most of them at least will be found to rest
on the basis of a fair induction from the facts brought forward.
It might, consequently, have appeared more logical to defer
these statements to the end of the work, instead of placing
them at the beginning. Unless, however, they are read and
mastered first, a great deal that is stated in the following pages
will be unintelligible, and the scope and purpose of the work
can be neither understood nor appreciated.1

1 For a fuller statement of the author's
views on Ethnography as applied to
Architectural art, see his ' History of

Architecture in all Countries,' 3rd ed.
vol. i. introd. pp. 52-85.

2. Naga people worshipping the Triratna emblem of Buddha, on a fiery pillar.
(From a bas-relief at Amaravati. )

VOL. I.

I)

5°

Sri or Gaja Lakshml, seated on a Lotus, with two Elephants pouring
water over her. (From a modern sculpture from Indor. )

Map of

INDIA

Showing the Principal

Buddhist Locaiities

London.

BOOK I.
BUDDHIST ARCHITECTURE.

<s>

CHAPTER I.

INTRODUCTION AND CLASSIFICATION.

IT may create a feeling of disappointment in some minds
when they are told that there is no stone architecture in
India older than two and a half centuries before the Christian
Era ; but, on the other hand, it adds immensely to the clear-
ness of what follows to be able to assert that India owes the
introduction of the use of stone for architectural purposes, as
she does that of Buddhism as a state religion, to the great
A^oka, who reigned from about B.C. 265 to 228.

It is not, of course, meant to insinuate that the people of

India had no architecture before that date ; on the contrary,

it can be proved that they possessed palaces and halls of

assembly, perhaps even temples, of great magnificence and

splendour, long anterior to A^oka's accession ; but, like the

buildings of the Burmese at the present day, they were all

in wood. Stone, in those days, seems to have been employed

only for the foundations of buildings, or in engineering works,

such as city walls and gates, or bridges or embankments ; all

else, as will appear from the sequel, were framed in carpentry.

Much as we may now regret this, as all these buildings have

consequently perished, it is not so clear, as it may at first

appear, that the Indians were wrong in this, inasmuch as,

in all respects, except durability, wood is a better building

material than stone. It is far more easily cut and carved,

larger spaces can be covered with fewer and less cumbrous

points of support than is possible with stone, and colour and

gilding are much more easily applied to wood than to stone.

For the same outlay twice the space can be covered, and

more than twice the splendour obtained by the use of the

more perishable material, the one great defect being that it

52 BUDDHIST ARCHITECTURE. BOOK I.

is ephemeral. It fails also in producing that impression of
durability which is so essential to architectural effect ; while,
at the same time, the facility with which it can be carved
and adorned tends to produce a barbaric splendour far less
satisfactory than the more sober forms necessitated by the
employment of the less tractable material.

Be this as it may, it will, if I mistake not, become quite clear
when we examine the earliest " rock-cut temples " that, whether
from ignorance or from choice, the Indians employed wood and
that only, in the construction of their ornamental buildings,
before A-roka's time.1 From this the inference seems inevitable
that it was in consequence of India being brought into contact
with the western world, first by Alexander's raid, and then by
the establishment of the Baktrian kingdom in its immediate
proximity, that led to this change. We do not yet know pre-
cisely how far the Baktrian kingdom extended towards the Indus,
but we feel Greek influence on the coinage, on the sculpture,
and generally on the arts of India, from an early date, and it
seems as if we might be able to fix with precision not only the
dates, but the forms in which the arts of the Western world
exerted their influence on those of the East. Meanwhile it
may be sufficient to state here that we know absolutely nothing
of the temples or architecture of the various peoples or religions
who occupied India before the rise of Buddhism,2 and it is only
by inference that we know anything of that of the Buddhists
before the age of Asoka. From that time forward, however, all
is clear and intelligible; we have a sufficient number of examples
whose dates and forms are known to enable us to write a fairly
consecutive history of the architectural style during the 1000
years Buddhism was prevalent in India, and thence to trace its
various developments in the extra Indian countries to which it
was carried, and where it is still practised at the present day.3

1 These remarks must not be taken as
applying to sculpture also. It is quite
true that no stone sculptures have yet
been found in India of an earlier date
than the age of Aroka ; but, as will be
seen in the sequel, the perfection the
Indian artists had attained in stone
sculpture when they executed the bas-
reliefs at Bharaut (B.C. 200), shows a
familiarity with the material that could
only be attained by long practice.

2 No mention of temples, or, indeed,
of buildings is, I believe, found in the
Vedas, and though both are frequently
alluded to, and described in the Epic
Poems and the Puranas, this hardly
helps us ; first because, like all verbal
descriptions of buildings, they are too

vague to be intelligible, and secondly,
because there is no proof that the passages
containing these descriptions may not
have been interpolated after — possibly
long after — the Christian Era.

3 I believe I was the first to ascertain
these facts from a personal inspection of
the monuments themselves. They were
communicated to the Royal Asiatic
Society in a paper I read on the ' Rock-
cut Temples of India,' in 1842. Every
subsequent research, and every increase
of our knowledge, has tended to confirm
those views to such an extent that
they are not now disputed by any one
acquainted with the literature of the
subject.

CHAP. I. INTRODUCTION. 53

This being the case, it would be in vain to look for any
earlier architecture of any importance in India before A^oka's
time : such could be expected only in countries where stone
had been in use from the very earliest times. The Aryans, who
were dominant before the rise of Buddhism, wrote books and
expressed their ideas in words, like their congeners all the world
over, but they do not seem successfully to have cultivated the
aesthetic arts, or to have sought for immortality through the
splendour or durability of their buildings. That was the aspira-
tion of the Turanian and other races, and we owe it to this
circumstance that we are enabled to write with any certainty
the history of their rise and fall as evidenced in their archi-
tectural productions.

There is no d priori improbability that the Dravidian races
of the south of India, or the indigenous races of the north, may
not have erected temples or other buildings at a very early
date, but if so, all that can be said is that all trace of them is
lost. When we first meet the Buddhist style it is in its infancy
— a wooden style painfully struggling into lithic forms — and we
have no reason to suppose that other styles were then more
advanced. When, however, we first meet them, some six or
seven centuries afterwards, they are so complete in all their
details, and so truly lithic in their forms, that they have hitherto
baffled all attempts to trace them back to their original types,
either in the wood or brick work, from which they may have
been derived. So completely, indeed, have all the earlier
examples been obliterated, that it is now doubtful whether the
missing links can ever be replaced. Still, as one single example
of a Hindu temple dating before the Christian Era might solve
the difficulty, we ought not to despair of such being found,
while the central provinces of India remain so unexplored as
they are. Where, under ordinary circumstances, we ought to
look for them, would be among the ruins of the ancient cities
which once crowded the valley of the Ganges ; but there
the ruthless Moslim or the careless Hindu have thoroughly
obliterated all traces of any that may ever have existed. In
the remote valleys of the Himalaya, or of Central India, there
may, however, exist remains which will render the origin and
progress of Hindu architecture as clear and as certain as that
of the Buddhist ; but till these are discovered, it is with the
architecture of the Buddhist that our history naturally begins.
Besides this, however, from the happy accident of the Buddhists
very early adopting the mode of excavating their temples in
the living rock, their remains are imperishably preserved to us,
while it is only too probable that those of the Hindu, being in
less durable forms, have disappeared. The former, therefore,

54 BUDDHIST ARCHITECTURE. BOOK I.

are easily classified and dated, while the origin of the latter, for
the present, seems lost in the mist of the early ages of Indian
arts. Meanwhile, the knowledge that the architectural history
of India commences about B.C. 250, and that all the monuments
now known to us are Buddhist, or of cognate sects, for at least
five or six centuries after that time, are cardinal facts that
cannot be too strongly insisted upon by those who wish to clear
away a great deal of what has hitherto tended to render the
subject obscure and unintelligible.

CLASSIFICATION.

For convenience of description it will probably be found
expedient to classify the various objects of Buddhist art under
the five following groups, though of course it is at times
impossible to separate them entirely from one another, and
sometimes two or more of them must be taken together as parts
of one monument.

ist. Stambhas or Ldts. — These pillars are common to all the
styles of Indian architecture. With the Buddhists they were
employed to bear inscriptions on their shafts, with emblems or
animals on their capitals. With the Jains they were generally
Dipdans, or lamp-bearing pillars, but sometimes supporting
quadruple figures of a Jina ; with the Vaishnavas they as
generally bore statues of Garuda or Hanuman ; with the .Saivas
they bore the trLmla symbol or were Dipdans and flag-staffs ;
but, whatever their destination, they were always among the
most original, and frequently the most elegant, productions of
Indian art.

2nd. Stiipas or Topes. — These, again, were primarily relic-
shrines, but may be divided into two classes, according to their
destination : first, Stupas proper, or monuments containing relics
of Buddha or of some Buddhist saint ; x secondly, the stupas or
towers erected to commemorate some event or mark some
sacred spot dear to the followers of the Buddha. If it were
possible, these two ought to be kept separate, but no external
signs have yet been discovered by which they can be dis-
tinguished from one another, and till this is so, they must be
considered, architecturally at least, as one.2

1 The Jains in very early times had Mathura'; 'Epigraphia Indica,' vol. ii.

stupas and worshipped at them. Even plates at pp. 314-321; ' Actes du Vienna

still the Samosaranas in some^ of their Congres Int. Orient.' vol. iii. pt. ii.,

temples at 5atrunjaya, Girnar, Abu, etc., plate at p. 142; and z'«/ra pp. ill, 130.

are survivals of the earlier stupas. They \ 2 Dagaba is a Singalese word applied to

were also known as' Chaityas — as stupas a stupa, from the Sanskrit "dhatu," a

are still called in Nepal and Tibet. — | 'relic,' 'element,' and "garbha" (in

Biihler, 'Legend of the Jaina Stupa at Pali "gabbho") a 'womb,' 'receptacle,'

CHAP. I.

CLASSIFICATION.

55

3rd. Rails. — These must be recognised as one of the most
important features of Buddhist architecture. Generally they
are found surrounding Topes, but they are also represented
as enclosing sacred trees, temples, pillars, and other objects.
It may be objected that treating them separately is like de-
scribing the peristyle of a Greek temple apart from the cella.
The Buddhist rail, however, in early ages at least, is never
attached to the tope, and is used for so many other, and such
various purposes, that it will certainly tend to the clearness of
what follows if they are treated separately.

4th. Chaityas * or Assembly Halls. — Chaitya is a more
general term than stupa, and may be applied to any building
of the nature of a religious monument, but more correctly to
the second division of stupas, or those commemorative of acts,
miracles, etc., or not funerary.2 But it has further been restricted
so as to correspond with the churches of the Christian religion :
their plans, the position of the altar or relic casket, the aisles,
and other peculiarities are the same in both, and their uses are
identical in so far as the ritual forms of the one religion
correspond to those of the other.

5th. Vihdras or Monasteries. — With the Buddhists and
Jains a Vihara was a hall where the monks met and walked
about ; afterwards these halls came to be used as temples,
and sometimes became the centres of monastic establishments.
Like the Chaityas, they resemble very closely the corresponding
institutions among Christians. In the earlier ages they
accompanied, but were detached from, the Chaityas or churches.
In later times they were furnished with chapels in which the
service could be performed independently of the Chaitya halls
which may or may not be found in their proximity.

or 'shrine.' Dhatugarbha is thus the
relic-receptacle or inner shrine, and is
strictly applicable only to the dome of
the stupa, sometimes called the "anda"
or egg. ' Dhatus ' were not merely
relics in the literal sense, but memorials
in an extended acceptation, and were
classified as — corporeal remains ; objects
belonging to the teacher, as his staff,
bowl, robe, holy spots, etc. ; and any
memorial, text of a sacred book, cenotaph
of a teacher, etc. Stupas are known as
Chaityas in Nepal, and as Dagabas in
Ceylon.
1 The word Chaitya, like Stupa, means

primarily a heap or tumulus, but it also
means a place of sacrifice or religious
worship, an altar — from chitd, a heap, an
assemblage, etc. Properly speaking,
therefore, the chaitya caves ought perhaps
to be called " halls containing a chaitya,"
or "chaitya halls," and this latter term
will consequently be used wherever any
ambiguity is likely to arise from the use
of the simple term Chaitya.

2 All structures of the nature of
sanctuaries are Chaityas, so that sacred
trees, statues, religious inscriptions and
sacred places come also under this general

BUDDHIST ARCHITECTURE.

BOOK I.

CHAPTER II.

STAMBHAS OR LATS.

IT is not clear whether we ought to claim a wooden origin for
these, as we can for all the other objects of Buddhist archi-
tecture. Certain it is, however, that the lats of Ajoka, with
shafts averaging twelve diameters in height, are much more like
wooden posts than any forms derived from stone architecture,
and in an age when wooden pillars were certainly employed to
support the roofs of halls, it is much more likely that the same
material should be employed for the purposes to which these
stambhas were applied, than the more intractable material of
stone.

The oldest authentic examples of these lats that we are
acquainted with, are those which King Ajoka set up in the
twenty-seventh year after his consecration — the thirty-first of
his reign — to bear inscriptions conveying to his subjects the
leading doctrines of the new religion he had adopted. The
rock-cut edicts of the same king are dated in his twelfth year,
and convey in a less condensed form the same information —
Buddhism without Buddha — but inculcating respect to parents
and priests, kindness and charity to all men, and, above all,
tenderness towards animal life.1

The best known of these lats is that removed from Topra in
Ambala district, and set up in 1356, by Firoz Shah Tughlak, in
his Kotila at Delhi, without, however, his being in the least aware
of the original purpose for which it was erected, or the contents
of the inscription. A fragment of a second was found lying on
the ridge, north of Delhi, where it had been set up by Firoz

1 These inscriptions have been pub-
lished in various forms and at various
times by Sanskrit scholars, such as
Burnouf, Kern, Senart, Biihler, etc.
Among these reference may be made to
]£. Senart, ' Les Inscriptions de Piyadasi '
(2 vols.) Paris, 1881-1886; Buhler, in
' Epigraphia Indica,' vols. i. and ii. ; and
'Archaeological Survey of Southern India,'
vol. i. Among other things, they explain

to us negatively why we have so little
history in India in these days. Ajoka is
only busied about doctrines. He does
not even mention his father's name ; and
makes no allusion to any historical event,
not even those connected with the life of
the founder of the religion. Among a
people so careless of genealogy, history
is impossible.

CHAP. II.

STAMBHAS OR LATS.

57

5. Assyrian honeysuckle ornament from capital

of Lat, at Allahabad.

tree — a deodar pine, for instance — than anything
designed in stone. Like all the others of this class,
this lat has lost its crowning ornament, which

Shah, in his hunting lodge ; and was re-erected in
1867.* Two others exist in Champaran district at
Radhia, and Mathia,2 and a fragment of another was
recognised — utilised as a roller for the station roads
by an utilitarian member of the Civil Service. The
most complete shaft, however, is that which, in
1837, was found lying on the ground in the fort at
Allahabad, and then re-erected with a pedestal,
from a design by Captain Smith.3 This pillar is
more than usually interesting, as in addition to
the A-yoka inscriptions it contains one by Samu-
dragupta (A.D. 380 to 400), detailing the glories
of his reign, and the great deeds of his ancestors.4
It seems again to have been thrown down, and
was re-erected, as a Persian inscription tells us,
by Jahangir (A.D. 1605), to commemorate his
accession. It is represented without the pedestal
(Woodcut No. 4). The shaft, it will be observed,
is more than 3 ft. wide at the base, diminishing to f
2 ft. 2 in. at the summit, which in a length of 33 £
ft.5 looks more like the tapering of the stem of a j

it

v.y.y.

r&A

w>"
|
W«Y:j

. i'i/i'

SL

i~l

•&

i

1 ' Journal of the Asiatic Society of Bengal,' vol. vi. p. 794.
had been brought from Mirath in 1356.

2 The first of these is known as the Lauriya-Araraj or Bakhira
pillar, being at the village of Lauriya about a mile from the temple
of Mahadeva Araraj, the shaft of which rises nearly 40 ft. above
the water level ; and the second is the Lauriya Navandgarh lat,
3 miles north of Mathia. Cunningham, 'Archaeological Survey
Reports,' vol. i. pp. 67 and 73; xvi. plate 17; ' Epigraphia I
Indica,' vol. ii. pp. 2450*".

3 ' Journal of the Asiatic Society of Bengal,' vol. vi. plate 40, at
P- 798.

Fleet's 'Gupta Inscriptions,' pp. 1-17.

5 These dimensions are taken from Capt. Burt's drawings pub-
lished in the 'Journal of the Asiatic Society of Bengal,' vol. iii.
pp. 105-123 and plate 3. 4. L4t at Allahabad.

BUDDHIST ARCHITECTURE.

BOOK I.

probably was a Buddhist emblem — a wheel or the triratna orna-
ment1— but the necking still remains (Woodcut No. 5), and
is almost a literal copy of the honeysuckle ornament we are
so familiar with as used by the Greeks with the Ionic order.
In this instance, however, it is hardly probable that it was in-
troduced direct by the Greeks, but is more likely to have been
borrowed, through Persia, from Assyria, whence the Greeks also
originally obtained it. The honeysuckle ornament, again,
occurs as the crowning member of a pillar at Sankisa, in the
Doab, half-way between Mathura and Kanauj (Woodcut No. 6),
and this time surmounting a capital of so essentially Persepolitan
a type, that there can be little doubt that the design of the
whole capital came from Persia. This pillar, of which the
greater part of the shaft is lost, is surmounted by an elephant,

but so mutilated that even in the 7th
century the Chinese traveller Hiuen
Tsiang mistook it for a lion, if this is
indeed the effigy he was looking at,
as General Cunningham supposes,2
which, however, is by no means so
clear as might at first sight appear.

6. Capital at Sankisa. (From a Drawing
by Gen. Cunningham.)

Capital of Lat in Tirhut. (From a
Drawing by Capt. Kittoe.)

Another capital of a similar nature to that last described
crowns the Lauriya Navandgarh lat in Champaran — this time
surmounted by a lion of bold and good design (Woodcut
No. 7). In this instance, however, the honeysuckle ornament

1 ' Tree and Serpent Worship,' plates
9, 10, loa, et passim.

2 ' Archaeological Reports,' vol. i. p.
274, plate 46.

CHAP. II.

STAMBHAS OR LATS.

59

is replaced by the more purely Buddhist ornament of a flock
of the sacred hansas or geese. In both instances there are
cable ornaments used as neckings, and the bead and reel so
familiar to the student of classical art. The last named form
is also, however, found at Persepolis. These features it may
be remarked are only found on the lats of Ajoka, and are
never seen afterwards in India, though common in Gandhara
and on the Indus for long afterwards, which seems a tolerably
clear indication that it was from Persia that he obtained those
hints which in India led to the conversion of wooden archi-
tecture into stone. After his death, these classical features
disappear, and wooden forms resume their sway, though the
Persian form of capital long retained its position in Indian
art. Whatever the Hindus copied, however, was changed, in
the course of time, by decorative additions and modifications,
in accordance with their own tastes.

To the preceding five we have to add four more lats
found in recent years. These are: (i) one at Rampurwa near
Pipariya, also in Champaran district, on the edge of the Tarai,
discovered in 1881, but not quite excavated; (2) at Nigliva
in the Nepal Tarai, about 18 miles north from Chilliya in Basti
district, a broken pillar, bearing an inscription by Asoka,
stating that it had marked the birthplace of Kanakamuni
Buddha ; (3) at Rummindei about 13 miles south-east from the
preceding and north of Paderiya, the lower 22 ft. of a lat,
discovered by Dr. A. Fiihrer in 1896, and bearing an inscription
stating that it marked the spot where Sakyamuni was born,1
and (4) a large fragment and capital found, about three years
ago, at Sarnath, bearing a portion of an Asoka. inscription.

It is more than probable that each of these A^oka lats
stood in front of, or in connection with some stupa, or building
of some sort ; but all these have disappeared, and the lats
themselves have — some of them at least — been moved more
than once, so that this cannot now be proved. So far, however,
as can now be ascertained, one or two stambhas stood in front
of, or beside each gateway of every great tope, and one or
two in front of each chaitya hall. At least we know that
six or seven can be traced at Sanchi, and nearly an equal
number at Amaravati,2 and in the representation of topes at
the latter place, these lats are frequently represented both
outside and inside the rails.

At Karle, one still stands in front of the great cave sur-

1 Cunningham, ' Archaeological Re-
ports,' vol. xvi. pp. nof. and plate 28 ;
Fiihrer, ' Buddha Sakyamuni's Birth-
place.'pp. 2;f. ,33f.,andplates4-6. There
is also a fragment of a stambha at the

great Sanchi stupa, with remains of an
A^oka inscription on it.

2 ' Tree and Serpent Worship,' plates
I, 5, 89 and 90.

6o

BUDDHIST ARCHITECTURE.

BOOK I.

mounted by four lions, which, judging from analogy, once bore
a chakra or wheel, probably in metal.1 (Woodcut No. 8.) A
corresponding pillar probably once stood on the opposite side of

the entrance bearing some
similar emblem. Two such
are represented in these
positions in front of the
great cave at Kan her i,
which is a debased copy of
the great Karle cave.2

The lat at Eran and the
iron pillar at Meharauli
nearDelhi,thoughsimilarin
many respects to those just
described, seem certainly
to belong to the era of the
Guptas during the 5th
century of our era — the
latter about 415, and the
former in A.D. 484 — and
to be dedicated to the
Vaishnava creed, and in
consequence belong to a
subsequent chapter. That
at Pathari in Bhopal bears
a much obliterated in-
scription of a Rashtrakuta
king, Parabala, dated in
A.D. 86i.3

This is a meagre ac-
count, it must be confessed,
of Buddhist lats, which probably at one time could be counted
by hundreds in the important Buddhist localities in Bengal ;
but it is feared we shall hardly be able to add many more to
our list. They are so easily overthrown and so readily utilised
in populous localities, that all trace of most of them has pro-
bably been irrecoverably lost, though one or two more examples
may possibly be found in remote, out-of-the-way places.

There is no instance, so far as I am aware, of a built
monumental pillar of ancient date now standing in India.
This is sufficiently accounted for by the ease with which they

Capital of the Lion pillar at Karl£.
(From a sketch by Mr. H. Cousens.)

1 'Tree and Serpent Worship,' plate 42.

1 In the description accompanying
Daniell's view of this cave he says : " On
the pillar to the right, above the capital,
is a group of lions, from the centre of
which a few years since arose the chakra
. . . though not the least appearance of

it at present remains."

3 ' Journal of the Asiatic Society of
Bengal, 'vol. xvii. pp. 3O5f. ; Cunningham,
' Archaeological Reports, ' vol. x. p. 70 ;
Kielhorn, in ' Nachrichten Gottingen
Gesellschaft : Phil. Hist. Kl.,' 1901, pp.

CHAP. II.

STAMBHAS OR LATS.

61

could be thrown down and their materials removed, when
they had lost the sanctity which alone protected them. There
are, however, two such pillars among the topes of Kabul, and
evidently coeval with them, now called the Surkh Min^r and
the Minar Chakri (Woodcut No. 9). These are ascribed by
the traditions of the place to Alexander the Great, though
they are evidently Buddhist monuments, meant to mark some
sacred spot, or to commemorate some event, the memory of which
has passed away. There can be little doubt that their upper
members are meant to be copies of the tall capitals of the
Persepolitan pillars, which were probably common throughout
this part of Asia, but their shape and outline exhibit great
degeneracy from the purer forms with which that architecture
commenced in India, and which were there retained in their
purity to a much later period than in this remote province.
No reliable data seem to exist for ascertaining what the age
of these monuments may be. It probably was the 3rd or 4th
century of our era, or it may be even earlier.

9. Min&r Chakri, Kabul.

(From a Drawing by Mr. Masson, in Wilson's ' Ariana Antiqua. ')

62

BUDDHIST ARCHITECTURE.

BOOK I.

CHAPTER III.

STtJPAS.
CONTENTS.

Relic Worship— Bhilsa Topes— Topes at Sarnath and in Bihar— Amaravati
Stupa— Gandhara Topes— Jalalabad Topes— Manikyala Stupa.

THERE are few subjects of like nature that would better reward
the labour of some competent student than an investigation into
the origin of Relic Worship and its subsequent diffusion over
the greater part of the old world. So far as is at present known,
it did not exist in Egypt, nor in Greece or Rome in classical
times, nor in Babylon or Assyria. In some of these countries
the greatest possible respect was shown to the remains of
departed greatness, and the bones and ashes of persons who
were respected in life were preserved with care and affection ; l
but there was no individual so respected that a hair of his head,
a tooth, or a toe-nail, even a garment or a utensil he had used,
was considered as a most precious treasure after his death. In
none of these countries does it appear to have occurred to any
one that a bone or the begging-pot of a deceased saint was a
thing worth fighting for ; or that honour done to such things
was a meritorious act, and that prayers addressed to them were
likely to be granted. Yet so ingrained do these sentiments
appear to be among the followers of Buddha, that it is difficult
to believe that the first occasion on which this sentiment arose,
was at the distribution of his remains on his attaining Nirvana
at Kusinagara, about B.C. 480. On that occasion, eight cities
or principalities are said to have contended for the honour of
possessing his mortal remains, and the difficulty was met by
assigning a portion to each of the contending parties, who are
said to have erected stupas to contain them in each of their
respective localities.2 None of these can now be identified with

1 Examples of this may be cited in the
reverence of the Athenians for the
remains of Theseus and Oedipus, and
the honours paid to those of Demetrius ;
but this bears no analogy to the relic-
worship of India, and Central Asia.

2 These were Rajagriha, Vauali,
Kapilavastu, Allakappa, Rimagrama,
Vethadipa, Pava, and Kunnara.—' Bud-
dhist Art in India,' p. 15 ; S. Hardy's
'Eastern Monachism,' p. 221.

The fame of this distribution seems to

CHAP. III.

STUPAS.

certainty — everything in future ages being ascribed to Aj-oka,
who, according to popular tradition, is said to have erected
the fabulous number of 84,000 relic shrines, or towers to mark
sacred spots.1 Some of these may be those we now see, or are
encased within their domes ; but if so, they, like everything else
architectural in India, are the earliest things we find there. It
is true, the great pagoda — the Shwe Dagon at Rangun — is said
to contain relics of all the four Buddhas of the present Kalpa,
the staff of Kakusandho ; the water-dipper of Kon^gamano ;
the bathing garment of Kajyapa, and eight hairs from the head
of Gautama Buddha ; 2 but supposing this to be true, we only
now see the last and most modern, which covers over the older
erections. This is at least the case with the great dagaba at
Bintenne, near Kandy, in Ceylon, in which the thorax-bone of
the great ascetic is said to lie enshrined. The ' Mahawansa,' or
Buddhist history of Ceylon, describes the mode in which this
last building was raised, by successive additions, in a manner
so illustrative of the principle on which these relic shrines
arrived at completion, that it is well worth quoting: — "The
chief of the Devas, Sumana, supplicated of the deity worthy of
offerings for something worthy of worship. The Vanquisher,
passing his hand over his head, bestowed on him a handful of his
pure blue locks from the growing hair of the head. Receiving
it in a superb golden casket, on the spot where the divine
teacher had stood, he raised an emerald stupa over it and
bowed down in worship.

" The thero Sarabhu, at the demise of the supreme Buddha,
receiving at his funeral pile the Thorax-bone, brought and
deposited it in that identical dagaba. This inspired personage
caused a dagaba to be erected twelve cubits high to enshrine
it, and thereon departed. The younger brother of King
Devanampiyatissa (B.C. 244), having discovered this marvellous
dagaba, constructed another encasing it, thirty cubits in height.
King Dutthagamini (cir. B.C. 96), while residing there, during his
subjugation of the Malabars, constructed a dagaba, encasing
that one, eighty cubits in height." Thus was the " Mahiyangana

have reached Europe at least as early as
the 1st century of the Christian Era,
inasmuch as Plutarch (' Moralia,' p. 1002,
Diibner, ed., Paris, 1841) describes a
similar partition of the remains of
Menander, among eight cities who are
said to have desired to possess his
remains ; but as he does not hint that
it was for purposes of worship, the
significance of the fact does not seem to

have been appreciated. Conf. ' Questions
of King Milinda ' in ' Sacred Books of
the East,' vol. xxxv. introd. p. 20.

1 ' Mahawansa,' p. 26, ' Hiouen
Thsang,' torn. ii. p. 417 ; Beal, 'Buddhist
Records,' vol. ii. pp. 87-88.

2 Account of the great bell at Rangun.
— Hough, ' Asiatic Researches,' vol. xiv,
p. 270.

BUDDHIST ARCHITECTURE.

BOOK I.

dagaba completed."1 It is possible that at each successive
addition some new deposit was made ; at least most of the
topes examined in Afghanistan and the Panjab, which show
signs of these successive increments, seem also to have had
successive deposits, one above the other.

Of the four canine teeth of Sakyamuni, one is said to have
been honoured among the Devas or gods, another among the
Nagas or water-spirits, the third was carried to Gandhara,2 and
the fourth to Kalinga. Little or nothing is related of the first
three ; the most celebrated is the left canine tooth. At the
original distribution it is said to have fallen to the lot of Orissa,
and to have been enshrined in a town called from that circum-
stance " Dantapura." This, most probably, was near the modern
town of Kalingapatam ; or possibly, as has been supposed,
the celebrated temple of Jagannath, which now flourishes at
Puri, may be on the site of the temple to which the tooth
belonged. Be this as it may, it seems to have remained there
in peace for more than eight centuries, when Guha^iva, the king
of the country, being attracted by some miracles performed by
it, and by the demeanour of the priests, became converted from
the Brahmanical cult, to which he had belonged, to the religion
of Buddha. The dispossessed Brahmans thereon complained to
his suzerain lord, resident at Pataliputra, in the narrative called
only by his title Pandu. He ordered the tooth to be brought
to the capital, when, from the wonders it exhibited, he was con-
verted also; but this, and the excitement it caused, led to its
being ultimately conveyed surreptitiously to Ceylon, where it is
said to have arrived about the year 310; and in spite of various
vicissitudes, its representative still remains in British custody,
the palladium of the kingdom, as it has been regarded during
the last sixteen centuries.3

Almost as celebrated was the begging -pot of Sakyamuni,
which was long kept in a dagaba or vihara erected by Kanishka
at Peshawar, and worshipped with the greatest reverence.4 After

1 Abstracted from the ' MaMwansa,'
chap. I.

2 It was preserved at Nagara or
Nagarahara near Jalalabad, where Fah-
Ilian, A.D. 400, in his I3th chapter
describes it as perfect. Hiouen Thsang,
' Memoires,' tome ii. p. 97, describes the
stupa as ruined, and the tooth having
disappeared. — Beal, ' Buddhist Records,'
vol. i. p. 92.

3 The principal particulars of this
story are contained in a Singalese work
called the ' Daladavamsa,' translated by
Sir Mutu Comara Swamy. See also
'Journal of the Royal Asiatic Society' i

(N.S.) vol. iii. pp. 132 et seqq. ; 'Tree
and Serpent Worship,' pp. 174 et seqq. ;
S. Hardy's ' Eastern Monachism,' pp.
224(7"; and Dr J. Gerson da Cunha's
' Memoir on the History of the Tooth
Relic,' in 'Journal of the Bombay Br.
Royal Asiatic Society,' vol. xi. pp. 115-
146.

4 ' Foe Koue Ki,' ch. xii. pp. 77, 82,
83 ; Beal's 'Travels of Fah-Hian,' pp. 36-
37, or ' Buddhist Records,' vol. i. introd.
pp. xxxii., xxxiii. Conf. Cunningham,
' Archreological Reports,' vol. xvi. pp.
8-1 1, and plate 3.

CHAP. III.

STUPAS.

paying a visit to Benares,1 it was conveyed to Kandahar, and is
said to be still preserved there by the Musalmans, and looked
upon even by them as a most precious relic.2

All this will become plainer as we proceed, for we shall
find every Buddhist locality sanctified by the presence of relics,
and that these were worshipped apparently from the hour of
the death of the founder of the religion to the present day.
Were this the place to do it, it would be interesting to try
and trace the path by which, and the time when, this belief
in the efficacy of relics spread towards the west, and how and
when it was first adopted by the Catholic Church, and became
with them as important an element of worship as with
the Buddhists.3 That would require a volume to itself; mean-
while, what is more important for our present purpose is the
knowledge that this relic-worship gave rise to the building of
these great stupas or dagabas, which are the most important
feature of Buddhist architectural art.

No one can, I fancy, hesitate in believing that the Buddhist
stupa is the direct descendant of the sepulchral tumulus of
the Turanian races, whether found in Etruria, Lydia, or among
the Skyths of the northern steppes. The Indians, however,
never seem to have buried, but always to have burnt, their
dead, and consequently never, so far as we know, had any
tumuli among them. It may be in consequence of this that
the stupas, in the earliest times, took a rounded or domical
form, while all the tumuli, from being of earth, necessarily
assumed the form of cones. Not only out of doors, but in
the earliest caves, the forms of dagabas are always rounded ;
and no example of a straight - lined cone covering a stupa
has yet been discovered. This peculiarity, being so universal,
would seem to indicate that they had been long in use before
the earliest known example, and that some other material
than earth had been employed in their construction ; but we
have as yet no hint when the rounded form was first employed,
nor when it was refined into a relic shrine. We know, indeed,
from the caves, and from the earliest bas-reliefs, that all the

1 ' Hiouen Thsang,' tome i. p. 83, or
Beal, ' Life of Hiuen Tsiang,' p. 63.

2 ' Foe Koue Ki,' pp. 351-352 ; Beal's
'Travels of Fah-Hian,' p. 161. A de-
tailed account of its transference from
the true Gandhara — Peshawar — to the
new Gandhara in Kandahar will be found
in a paper by Sir Henry Rawlinson,
'Journal of the Royal Asiatic Society,'
vol. xi. p. 127. Conf. 'Indian Anti-

VOL. I.

quary,' vol. iv. p. 141.

3 The craze for relics that sprang up in
the 5th century was largely stimulated
by the writings of such authorities
as Augustine, Ambrose, Basil, and
Chrysostom. It was strictly akin to the
belief of the Buddhists. The reverence
for the " qadam-i-rasul," and relics of
Muhammad by his followers is also of
a similar character.

E

66

BUDDHIST ARCHITECTURE.

BOOK I.

roofs of the Indians were curvilinear — probably derived from
the roofs of thatched huts ; and if one can fancy a circular
chamber with a domical roof — not in stone, of course — as the
original receptacle of the relic, we may imagine that the form
was derived from this.1

The worship of stupas probably arose from the popular
idea that the sanctity of the relics was shared by their shrines ;
and gradually stupas, simply in memory of the Buddha or
of any of his notable followers, came to be multiplied and
reverenced everywhere. Many were solid blocks without any
receptacle for a relic ; but in those inside chaitya halls, the
casket was placed in the capital or Tee, whence it could be
readily transferred, or taken out on the occasion of a festival.

The earlier ones were very plain, consisting of a base or
drum and dome, with a square capital in the form of a box ;
the dome was regarded as the essential feature of a stupa,
and with the "chhattra" or umbrella over it, as a symbol of
dignity, and a surrounding path for " pradakshina " or circum-
ambulation fenced off by a wall or railing, it was complete.
In course of time they came to be honoured almost as the
Buddha himself — had his image affixed to their drums, and
were decked with parasols, garlands of flowers, and flags or
long ribbons, whilst presentations of money were made for
their service.

BniLSA TOPES.

The most extensive, and taking it altogether, perhaps the
most interesting, group of topes in India is that known as the
Bhilsi Stupas or Topes, from a town of that name on the
north border of Bhopal, near which they are situated. There,
within a district not exceeding 10 miles east and west and
6 north and south, are five or six groups of topes, containing
altogether between twenty-five and thirty individual examples.
The principal of these, known as the great tope at Sanchi-
Kanakheda, has been frequently described, the smaller ones
are known from General Cunningham's descriptions only ; 2 but

1 Among the bas-reliefs of the Bharaut
tope is one representing just such a
domical roof as this (Woodcut No. 81).
It is not, however, quite easy to make
out its plan, nor to feel sure whether
the object on the altar is a relic, or
whether it may not be some other kind
of offering.

2 ' Bhilsa Topes, or Buddhist Monu-
ments in Central India,' 1854. One

half of the work on ' Tree and Serpent
Worship," and forty-five of its plates,
besides woodcuts, are devoted to the
illustration of the great Tope ; and
numerous papers have appeared on the
same subject in the 'Journal of the
Asiatic Society ' and elsewhere. A cast
of the eastern gateway is in the South
Kensington Museum ; also in Edinburgh,
Dublin, Berlin, and Paris Museums.

CHAP. III.

BHILSA TOPES.

67

altogether they have excited so much attention that they are
perhaps better known than any group in India. We are not
however, perhaps, justified in assuming, from the greater extent
of this group as now existing, that it possessed the same pre-
eminence in Buddhist times. If we could now see the topes
that once adorned any of the great Buddhist sites in the Doab
or in Bihar, the Bhilsa group might sink into insignificance. It
may only be that, situated in a remote and thinly peopled part
of India, they have not been exposed to the destructive energy
of opposing sects of the Hindu religion, and the bigoted Moslim
has not wanted their materials for the erection of his mosques.
They consequently remain to us, while it may be that nobler
and more extensive groups of monuments have been swept from
the face of the earth.

Notwithstanding all that has been written about them, we
know very little that is certain regarding their object and their
history.1 Our usual guides, the Chinese Pilgrims, fail us here.
Fah Hian never was within some hundreds of miles of the place ;
and if Hiuen Tsiang ever was there, it was after leaving Valabhi,
when his journal becomes so confused and curt that it is always
difficult, sometimes impossible, to follow him. He has, at all
events, left no description by which we can now identify the
place, and nothing to tell us for what purpose the great tope or
any of the smaller ones were erected. The ' Mahawansa,' it is
true, helps us a little in our difficulties. It is there narrated that
A^oka, when on his way to Ujjeni (Ujjain), of which place he had
been nominated governor, tarried some time at Chetyagiri, or, as
it is elsewhere called, Wessanagara, the modern Besnagar, close
to Sanchi. He there married the daughter of a chief, and by
her had twin sons, Ujjenia and Mahinda, and afterwards a
daughter, Sanghamitta. The two last named are said to have
entered the priesthood, and played a most important part in the
introduction of Buddhism into Ceylon. Before setting out on
this mission, Mahinda visited his royal mother at Chetyagiri,
and was lodged in " a superb vihara," which had been erected
by herself.2 In all this there is no mention of the great tope,
which may have existed before that time ; but till some building
is found in India which can be proved to have existed before that
age, it may be assumed that this is one of the 84,000 topes said
to have been erected by A^oka. Had Sanchi been one of the

1 Colonel F. C. Maisey was sent by
Government in 1849 to make drawings
of the gateways and sculptures at Sanchi-
Kanakheda. These drawings — which
had been first used in ' Tree and Serpent
Worship,' — here-published in 1892, with

letterpress based on a fanciful theory as
to their age and origin.

2 ' Mahawansa,' chap. 13. See also
'Tree and Serpent Worship,' pp. 99 et
seqq., where all this is more fully set out
than is necessary here.

68

BUDDHIST ARCHITECTURE.

BOOK I.

eight cities which obtained relics of Buddha at the funeral pyre,
the case might have been different ; but it has been dug into, and
found to be a stupa without relics.1 It consequently may have
been erected to mark some sacred spot or to commemorate
some event, and we have no reason to believe that this was done
anywhere before A^oka's time.

On the other hand, two smaller topes at the same place
contained relics of a historical character. That called No. 2 Tope
contained those of some of the Buddhist teachers who took part
in the third great convocation held under A^oka, and some of
whom were sent on missions to the Himalayas, to disseminate
the doctrines then settled, and with these were associated the
names of others, probably contemporaries, but of whom we
know nothing otherwise.2 No 3 Tope contained two relic
caskets, represented in the accompanying woodcuts (Nos. 10

10. Relic Casket of Moggalana.

ii. Relic Casket of Sariputra.

and 11). One of these contained relics of Maha Moggalana,
the other of Sariputra, friends and companions of Buddha him-
self, and usually called his right and left hand disciples.3 It
does not of course follow from this that this dagaba is as
old as the time of Buddha ; on the contrary, the probability
seems to be that these relics were deposited there in A^oka's
time, in close proximity to the sacred spot, which the great
tope was erected to commemorate. The tope containing relics
of his contemporaries may, of course, be more modern, possibly
contemporary with the gateways.4

The general appearance of the Sanchi - Kanakheda Stupa
will be understood from the view of it on Woodcut No. 12, and
its shape and arrangement from the plan and section, Nos. 13

1 At least the excavations failed in the
discovery of a deposit.

2 The ' Dipawansa ' names the four
missionaries who accompanied Kasapa-
gota Kotiputa to convert the tribe of
Yakkas in Himavanta, as — Majjhima,
Dudubhisara, Sahadevaand Mulakadeva.
Kasapagota, Majjhima and Dudubhisara,
are named on relic-boxes from Sanchi and

Sonari. — 'Journal of the Royal Asiatic
Society,' 1905, pp. 6836°.

3 Cunningham, 'Bhilsa Topes,' pp.
297, 299 et seqq.

* The Chandragupta inscription on the
rail near the eastern gateway is a subse-
quent addition, and belongs to Chandra-
gupta II., of the year A.D. 412. Fleet,
' Early Gupta Inscriptions,' pp. 29-34.

CHAP. III. BHILSA TOPES. 69

and 14. From these it will be observed that the principal

12. View of the great Tope at Sanchi, north-east' side.

W

Scale HO ft. to 1 in.

13. Plan of great Tope at Sanchi.

Scale 50 ft. to 1 in.

14. Section of great Tope at Sanchi.

BUDDHIST ARCHITECTURE.

BOOK I.

building consists of a dome somewhat less than a hemisphere,
1 06 ft. in diameter, and 42 ft. in height.1

On the top of the stupa is a flat space about 34 ft. in
diameter, formerly surrounded by a stone railing, some parts
of which were found still lying there ; and in the centre of this
once stood a feature known to Indian archaeologists as a ' Tee.'2

The woodcut (No. 15), from
a rock-cut example at Ajanta,
represents the usual form at
this age. The lower part is
adorned with the usual Bud-
hist rail (to be described here-
after), the upper by the con-
ventional window pattern,
two features which are uni-
versal. It is crowned by a

15. -Tee 'cut in the rock on a Digaba at Ajanta. Hd of three slabs> an<j nQ

doubt either was or simulated

a relic casket. No tope, and no representation of a tope —
and we have hundreds — are without this feature, and generally
it is or was surmounted by one or more discs representing
the umbrellas of state ; in modern times by as many as nine
of these. The only ancient wooden one now known to exist
is that in the cave at Karle (Woodcut No. 70), but the repre-
sentations of them in stone and painting are literally thousands
in number.

The dome rests on a sloping base, 14 ft. in height by 121
ft. in diameter, having an offset on its summit about 6 ft.
wide.3 This, to judge from the representations of topes on
the sculptures, must have been surrounded by a balustrade,
and was ascended by a broad double ramp on one side. It
was probably used for processions round the monument, which
seem to have been among the most common Buddhist cere-
monials. The centre of this great mound is quite solid, being
composed of bricks laid in mud ; but the exterior is faced with
dressed stones. Over these was laid a coating of cement nearly
4 inches in thickness, which was, no doubt, originally adorned
either with painting or ornaments in low relief.

1 These views, plans, etc., are taken
from a Memoir by Capt. J. D. Cunning-
ham, ' Journal of the Asiatic Society of
Bengal,' vol. xvi. (August, 1847), pp.

739-763-

2 This is an Anglicised form of the
Burmese " Hti," by which this member
is known ; in Sanskrit it is called harmikd.
The lower part is called gala in Nepal —
the " neck," and the whole cMddmani —

"thecrestorcrestjewel.'' The umbrellas
grouped over it were termed chhattrd-
vali, and the shaft on which they are
supported the yashti. — Foucher, 'L'Art
Greco-Bouddhique,' tome i. p. 97.

3 ' Journal of the Royal Asiatic
Society,' 1902, pp. 29-44 5 an^ Griin-
wedel's ' Buddhist Art in India,'
(English ed.), pp. 24-26.

CHAP. III. TOPES AT SARNATH AND IN BIHAR. 71

Besides the group at Sanchi, which compromises six or seven
stupas — mostly ruined, there is at Sonari, 6 miles distant, another
group of eight topes. Two of these are important structures,
enclosed in square courtyards, and one of them yielded numerous
relics to the explorers.

At Satdhara, 3 miles further on, is a great tope 101 ft.
in diameter, but which, like that at Sanchi, yielded no relics.
No. 2, however, though only 24 ft. in diameter, was found to
contain relics of vSariputra and Moggalana, like No. 3 at Sanchi.
Besides these there are several others, all small, and very much
ruined.

The most numerous group, however, is situated at Bhojpur,
7 miles south-east from Sanchi, where thirty-seven distinct
topes were grouped together on various platforms. The largest
is 66 ft. in diameter, but No. 2 is described as one of the
most perfect in the neighbourhood, and, like several others in
this group, contained important relics.

At Andher, about 5 miles west of Bhojpur, was a fine
group of three small, but very interesting topes. With those
above enumerated, this makes up about sixty distinct and
separate topes, in this small district, which certainly was not
one of the most important in India in a religious point of
view, and consequently was probably surpassed by many, not
only in the number but in the splendour of its religious edifices.1

Without more data than we at present possess, it is of
course impossible to speak with certainty with regard to the
age of this group of topes, but, so far as can be at present
ascertained, there seems no reason for assuming that any of
them are earlier than the age of A^oka, B.C. 250, nor is it
probable that any of them can be of later date than, say,
the first century before our era. The topes themselves seem
all to be included within these two centuries, or possibly
even less.

TOPES AT SARNATH AND IN BlHAR.

Not only is there no other group of topes in India Proper
that can be compared, either in extent or in preservation, to
those of Bhilsa, but our knowledge of the subject is now so
complete that it is probably safe to assert that only two, or
at most three, topes exist between the Satlaj and the sea,
sufficiently perfect to enable their form and architectural

1 As the particulars regarding all these ; ' Bhilsa Topes,' published in 1854, it has
topes, except those at Sanchi, are taken ' not been thought necessary to repeat the
from Gen. Cunningham's work entitled ! reference at every statement.

72 BUDDHIST ARCHITECTURE. BOOK I.

features to be distinguished. There are, of course, numerous
mounds near all the Buddhist cities which mark the site, and
many of which probably hide the remains, of some of the
hundreds of stupas or dagabas mentioned by the Chinese

Tope at S3.rna.th, near Benares. (From a Photograph.)

Pilgrims, besides many that they failed to distinguish. All,
however, with the fewest possible exceptions, have perished ;
nor is it difficult to see why this should be so. All, or nearly
all, were composed of brick or small stones, laid either without
mortar, or with cement that was little better than mud. They

CHAP. III. TOPES AT SARNATH AND IN BIHAR.

73

consequently, when desecrated and deserted, formed such
convenient quarries for the villagers, that nearly all have been
utilised for building huts and houses of the Hindus, or the
mosques of the Musalmans. Their rails, being composed
of larger stones and not so easily removed, have in some
instances remained, and some will no doubt be recovered
when looked for ; and as these, in the earlier ages at least,
were the iconostasis of the shrine, their recovery will largely
compensate for the loss of the topes which they surrounded.

The best known, as well as the best preserved of the Bengal
topes, is that called Dhamek, at Sarnath, near Benares (Wood-
cut No. 16). It was explored by General Cunningham in
1835-36, and found to be a stupa or chaitya — not containing
relics, but erected to mark some spot sanctified by the
presence of Buddha, or by some act of his during his long
residence there. In 1904-05 further excavations were made
under Mr. Oertel, but the results have not yet been made
public. It is situated in the Deer Park, where Buddha took
up his residence with his five disciples when he first removed
from Gaya on claiming to have attained Buddhahood, and
commenced his mission as a teacher. That it commemorates
this event, is exceedingly probable, since that stupa, of all
others, would be religiously preserved and restored. There
are several mounds in the neighbourhood, but the descriptions
of the Chinese Pilgrims are not sufficiently precise to enable
us always to discriminate between them.

The building consists of a stone basement, 93 ft. in diameter,
and solidly built, the stones being clamped together with iron
to the height of 43 ft. Above that it is in brickwork, rising to
a height of no ft. above the surrounding ruins, and 128 ft.
above the plain.1 Externally the lower part is relieved by eight
projecting faces, each 21 ft. 6 in. wide, and 15 ft. apart. In each
is a small niche, intended apparently to contain a seated figure
of Buddha, and below them, encircling the monument, is a
band of sculptured ornament of the most exquisite beauty. The
central part consists — as will be seen by the cut (No. 17) on
the next page — of geometric patterns of great intricacy, but
combined with singular skill ; and, above and below, foliage
equally well designed, and so much resembling that carved by
Hindu artists on the earliest Muhammadan mosques at Ajmir
and Delhi, as at first sight might suggest that they may not
be very distant in date.

1 These dimensions and details are
taken from Gen. Cunningham's 'Archaeo-
logical Reports,' vol. i. pp. 107 et seqq. ;
for his account of the exploration, see

'Journal Bengal Asiatic Society,' vol.
xxxii. ; Sherring's ' Sacred City of the
Hindus,' pp. 236-243.

74

BUDDHIST ARCHITECTURE.

BOOK I.

The carvings round the niches and on the projections have
been left so unfinished — in some instances only outlined— that
it is impossible to guess what ultimate form it may have been
intended to give them. The upper part of the tower was

, ;;:;,; ^,; y ,.,.,., „,*.• ^J^^^ '^"^-^^

la^s&Kgassg^^r^^r^:^---: — •&**• .rS^T^

—^l^r^rJ::' "J "7""' T :;i"'- /" v> -

17-

Panel on the Tope at Sarnath. (From a Photograph. )

possibly an addition of even later date than the lower part,
and may never have been finished at all ; but from our
knowledge of the Afghanistan topes we may surmise that it
was intended to encircle it with a range of pilasters, and then
some bold mouldings, before covering it with a hemispherical
dome.

CHAP. III. TOPES AT SARNATH AND IN BIHAR.

75

In his excavations, General Cunningham found, buried in
the solid masonry, at the depth of io£ feet from the summit, a
large stone on which was engraved the usual Buddhist formula,
" Ye dharmma hetu," etc., in characters belonging, he thought,
to the /th century, from which he inferred that the monument
belongs to the 6th century. But I cannot accept the conclusion ;
it seems to me much more probable that this stone may have
belonged to some building which had fallen into decay, or to
have been the pedestal of some statue which had been dis-
used, and was consequently utilised in the erection or repair
of this structure. One feels consequently more inclined to adopt
the tradition preserved by Captain Wilford,1 to the effect that
the Sarnath monument was erected by the sons of Mahipala, and
destroyed (interrupted ?) before its completion.2 We know that
the Deer Park, where the Buddha preached his first sermon, was
the site of one of the early and most sacred stupas, and the
excavations recently made have brought to light remains of all
ages from that of A^oka down to the nth century at least.
An inscription, found long since, and dated in A.D. 1026,
records the repair of a Buddhist stupa and Dharmachakra,
and the erection of a Gandhakuti temple in the time of
Mahipala — probably by his sons.3 Whether it refers to this
stupa or not, it indicates that large restorations did take
place as late as the nth century, when this also was probably
encased, as we now find it, with a modernised exterior. The
form of the monument with the eight projecting faces that
decorate its drum, the character of its sculptured ornaments,
the unfinished condition in which it is left, and indeed the
whole circumstances of the case, render this date so probable
that it may be accepted for the present at least, though it is
quite possible that further research may require us to modify
this opinion.

The only stone building yet found in India that has any
pretension to be dated before Ajoka's time is one at Rajgir,
having the popular name of Jarasandha-ka-baithak. As will
be seen from the annexed woodcut (No. 18), it is in the form
of a platform 85 ft. square at the base and sloping upwards for
20 or 28 ft. to a platform measuring 74 by 78 ft.4 It is built
wholly of stones, neatly fitted together without mortar ; and
its most remarkable peculiarity is that it contains fifteen cells,
one of which is shown in the woodcut. They are from

1 ' Asiatic Researches,' vol. ix. p. 203,
and vol. x., p. 130.

2 See also paper by Vesey Westmacott,
'Calcutta Review,' vol. lix., 1874, p. 68.

1 ' Asiatic Researches,' vol. v. p. 133 ;

' Indian Antiquary,' vol. xiv. p. 139.
The contents of this inscription was prob-
ably the basis of Wilford's statement.
4 ' Indian Antiquary,' vol. i. p. 72.

BUDDHIST ARCHITECTURE.

BOOK I.

6 to 7 ft. in length, by about half that in breadth, and they
are occupied at times up to the present day by Jogis whose
bodies are constantly smeared with ashes.1

"U

up

CELL

18. View and Plan of Jarasandha-ka-baithak. (From Cunningham, 'Archaeological
Survey Report,' vol. iii. plate 42.)

The other Bengal tope existing nearly entire, also known as
Jarasandha-ka-baithak, is opposite to the village of Giriyek,
about 6 miles east - north - east of the preceding. General
Cunningham states its dimensions to be 28 ft. in diameter by
21 ft in height, resting on a basement 14 ft. high, so that its
total height, when complete, may have been about 55 ft. As
it was not mentioned by Fah Hian, A.D. 400, and is, apparently,
by Hiuen Tsiang, A.D. 640, its age is probably, as General
Cunningham states, intermediate between these dates, or about
A.D. 5<x>.2 It is a bold, fine tower, evidently earlier than that
at Sarnath, and showing nothing of the tendency towards
Hindu forms there displayed. It has, too, the remains of a
procession-path, or extended basement which is wholly want-
ing at Sarnath, but which is always found in the earlier monu-
ments. It was erected, as Hiuen Tsiang tells us, in honour
of a Hama — goose — who devoted itself to relieve the wants of
a starving community of Bhikshus.3

The third stupa, if it may be so called, is the celebrated
temple or, properly, chaitya at Bodh - Gaya, which stands

1 'Cave-Temples,' pp. 331". ; Cunning-
ham, ' Archaeological Survey Report,'
vol. i. p. 20 ; vol. iii. p. 142.

2 'Archaeological Survey Report,' vol. i.

pp. 16-19, and plate 15.

3 ' Hiouen Thsang,' torn. iii. p. 60; or
Beal, 'Buddhist Records,' vol. ii. p. 181.

CHAP. III.

BODH-GAYA.

77

immediately in front of the celebrated Bodhi - tree (Ficus
religiosaj- under whose shade Buddha is said to have attained
complete enlightenment in the thirty-fifth year of his age, dr.
B.C. 525. Its history is told in such detail by Hiuen Tsiang2
that there seems little doubt as to the main facts of the case.
According to this authority, Ajoka built a small yjMra here,
but long afterwards this was replaced by a temple 160 ft.
high and 60 ft. (20 paces) wide, which are the exact dimen-
sions of the present building, according to Cunningham,3 and
we are further told that it was erected by a Brahman, who
was warned by Mahe^wara (Siva), in a vision, to execute
this work.4 In this temple there was a cella corresponding
with the dimensions of that found there, in which the
Brahman placed a statue of Buddha, seated cross-legged, with
one hand pointing to the earth.6 The date of the erection of
this temple is still obscure : General Cunningham laboured,
on rather doubtful data, to prove that it was erected in the
reign of the Kushan king, Huvishka, in the century B.C.,
and, from a coin found in a later image, he concluded that
additions and restorations were made in the 4th century.6
There are sculptures and inscriptions that must belong
between the 2nd and 7th centuries, and they are numerous
between the 9th and end of the I2th; but none of them
help us in definitely fixing the date of the temple. From
the style and what remains of the older sculptures, we can
only assume that it may belong to somewhere about the 6th'
century, though considerably altered in later times by successive
restorations. From an Arakanese inscription on the spot, first
translated by Colonel H. Burney, we further learn that the

1 Buchanan Hamilton was told by the
priests on the spot, in 1811, that it was
planted there 2225 years ago, or B.C. 414,
and that the temple was built 126 years
afterwards, or in 289 B.C. : — not a bad
guess for A^oka's age in a locality where
Buddhism has been so long forgotten. —
Montgomery Martin's ' Eastern India,'
vol. i. p. 76.

2 ' Hiouen Thsang,' torn. ii. pp. 464-
468 ; or Beal, 'Buddhist Records,' vol. ii.
pp. 118-121.

3 'Archaeological Reports,' vol. i. p. 5;
and ' Mahabodhi,' p. 18.

4 Wassiliev, in his work on the
Doctrine, History, etc., of Buddhism,
pp. 41-42, gives a somewhat different
account, stating that it was erected by
Punya, one of three brothers converted
by Uttara, one of whom built a temple
in the Deer Park at Benares, and the
other a temple in the Venuvana garden
at Rajagriha. Gen. Cunningham refers

to a story related by Taranath of two
brothers, one of whom built the Nalanda
temple, and, from the resemblance in
style, he infers that the MahSbodi temple
belongs to the same date, if not to one of
the brothers.

5 That is in the attitude known as the
' Bhumisparja mudrji. ', in which .Sakya-
mum sat wKen Ee attained supreme
knowledge.

8 'Mahabodhi,' pp. 17-25. This was
a coin of Pa^upati, Raja of Nepal, whose
name — possibly the Pampreksha of the
chronologies — is well-known to numis-
matists from his coins ; but nothing is
certain about his date, except that most
probably he lived subsequently to the
7th century. — 'Indian Antiquary,' vol.
xiii. p. 412; Wright's, 'History of
Nepal,' p. 113; Sylvain Levi, ' Le
Nepal,' tome ii. pp. 108-111; 'Journal
Royal Asiatic Society,' 1908, p. 68 1.

BUDDHIST ARCHITECTURE.

BOOK I.

place, having fallen into decay, was restored by the Burmese
in the year 1105, and again in i2gS.1

From the data these accounts afford us we gather that the
building we now see before us (Woodcut No. 19) is substantially

19. Temple at Bodh-Gaya with Bo-tree. (From a Photograph by Mr. Peppe, C.E.)

that erected in the 6th century, but the niches Hiuen Tsiang
saw, containing golden statues of Buddha, cannot be those
now existing — most of the images round the basement are

f l ' Asiatic Researches,' vol.xx. pp. 161-
189 ; Rajendralal Mitra, ' Buddha Gaya,'
p. 209 ; Phayre's ' History of Burma,'
p. 46 ; and 'Journal of the Asiatic Society
of Bengal,' vol. xxxvii. p. 97, note. Gen.
Cunningham read the dates as equivalent
to 1079 and 1086 A. D. ; ' Mahabodhi,' pp.

27, 28, and 77. But though a scholarly
translation of the inscriptions has yet to
be made, the readings of the Burmese or,
Arakanese dates as 467 and 660 of their lj
era, can hardly be questioned — i.e. 1105
and 1298 A.D,

CHAP. III.

BODH-GAYA.

79

distinctly of Burmese type, though some few of them appear
to be of about the 6th century — and the sculptures he mentions
find no place in the present design ; the amalakas of gilt copper
that crowned the whole, as he saw it, have also disappeared.1

The changes in detail, as well as the introduction of radiating
arches in the interior, must belong to the Burmese restorations
in the beginning of the I2th and end of the I3th centuries.
Though these, consequently, may have altered its appearance in
detail, it is probable that, until the "restoration" in 1 880-81,
we still had before us a straight-lined pyramidal nine-storeyed
temple of about the 6th century, retaining all its essential forms
— anomalous and unlike anything else we find in India, either
before or afterwards, but probably the parent of many nine-
storeyed towers found beyond the Himalayas, both in China
and elsewhere.

Unfortunately, instead of carefully preserving this very
interesting monument, the Government of Bengal was advised
to " restore " it, and this was carried out under the superintend-
ence of General Cunningham and his assistant, by which — as
might have been anticipated — it was materially modified, and
from an archaeological point of view seriously injured.2

Eventually we may discover other examples which may
render this noble tower less exceptional than it now appears
to be. At Kesariya in Champaran, about 20 miles south-
east from Lauriya-Araraj, where one of the pillars of A^oka
mentioned above is found, are the ruins of what appears to
have been a very large tope. It is, however, entirely ruined
externally, and has never been explored, so that we cannot
tell what was its original shape or purpose.3 At Piprahwa also
in the north of Basti district, on the Nepal frontier, a mound
containing the remains of a stupa was excavated in 1897-98,
but, apparently, more attention was given to the discovery of
the relic casket, than to the construction and dimensions of the
stupa. It seems, however, to have been about 90 feet, in
diameter at the base, and about 13^ ft. from the ground level ;
the dome began with a diameter of about 62 ft.4 The inscription

1 Beat's 'Buddhist Records,' vol. ii.
pp. 118 and 136, note 2.

2 Cunningham's 'Mahabodhi,' preface,
p. ix. The restoration cost the Govern-
ment somewhere about 200,000 rupees ;
and then the Mahant of the neighbouring
monastery appropriated the renovated
temple for his Vaishnava followers and
consecrated the image by applying to it the
"tilak" or frontal mark of Vishnu, so that
it might be worshipped as that divinity,
whilst he set both the Buddhists and

Government at defiance to re-occupy the
shrine.

3 Cunningham, 'Archaeological Re-
ports,' vol. i. pp. 64ff. and plate 24.

4 'Journal of the Royal Asiatic Society,'

1898, p. 577 ; and ' Report on a Tour of
Exploration of the Antiquities in the
Tarai, Nepal,' by Babu P. C. Mukherji,

1899, pp. 43-47, in which the "approxi-
mate total diameter of the whole stftpa "
is given as 9p£ ft. ; but both illustrations
and descriptions are very unsatisfactory.

8o

BUDDHIST ARCHITECTURE.

BOOK I.

on one of the relic caskets has attracted much attention on the
part of scholars as recording the deposit of relics of the Sakya
clansmen of Buddha.1

All along this line of country numerous Buddhist remains
are found, all more or less ruined, and they have not yet been
examined with the scientific care necessary to ascertain their
forms. This is the more to be regretted as this was the native
country of the founder of the religion, and the place where his
doctrines appear to have been originally promulgated. If any-
thing older than the age of A^oka is preserved in India, it is
probably in this district that it must be looked for.

A MAR A VAT!

Although not a vestige remains in situ of the central stupa
at Amaravatt, there is no great difficulty, by piecing together
the fragments of it now in the British Museum — as is done in
Plates 48 and 49 of 'Tree and Serpent Worship' — in ascertain-
ing what its dimensions and general appearance were. When
Colonel Mackenzie first saw it, in 1797, the central portion of
the mound was still untouched, and rose in a turreted shape to
a height of 20 ft. with a diameter of about 90 ft. at the top,
and had been cased round with bricks, and so may have been
40 or 50 ft. in height. This indicates a dome of considerable
size ; the base or drum was probably 162^ ft. in diameter, and
wainscotted with sculptured marble ; how broad it was above
we have no means of knowing, or whether there may not have
been even a second terrace ; but if, as is most probable, there
was only one, the dome may have been 120 to 140 ft. in
diameter. The perpendicular part was covered with sculptures
in low relief, representing stupas and scenes from the life of
Buddha. The domical part was covered with stucco, and with
wreaths and medallions either executed in relief or painted.
No fragment of them remains by which it can be ascertained
which mode of decoration was the one adopted.2

Altogether, there seems no doubt that the representation
of a stupa (Woodcut No. 20), copied from the Amaravati
marbles, fairly represents the central building there. There
were probably forty-eight such representations of dagabas on
the basement of the stupa. In each the subject of the sculpture
is varied, but the general design is the same throughout ; and,
on the whole, the woodcut may be taken as representing the

1 'Journal of the Royal Asiatic Society,'
1899, pp. 149-180; 1905, pp. 679f; and
1907, pp. I05f.

2 For a detailed account of the Amara-

vati Stupa, see ' Archaeological Survey of
Southern India : The Buddhist Stupas of
Amaravati and Jaggayyapeta, ' 1887.

CHAP. III.

AMARAVATI.

81

mode in which a Buddhist stupa was ornamented in the
ist or 2nd century, at which time the style seems to have
reached its highest point of elaboration, in India at least.1

20. Re

sentation of a Stupa from the Rail at Amaravati. (From a bas-relief in the
British Museum.)

In the Andhra country — or, at least, in the districts adjoining
the deltas of the Krishna and Godavan rivers — Buddhism must

1 The recent discovery of the base of
a stupa, about 1 1 ft. in diameter, outside
where the south gate of the great Amara-
vati stupa was, has revealed the style in
which the base or drum of these eastern
stupas was decorated — by marble slabs

VOL. I.

richly carved with representations of
stupas, placed at intervals, and with
other sculptures between. There was
no "inner rail" around the large stupa,
as was at one time assumed.

82

BUDDHIST ARCHITECTURE.

BOOK I.

have had a very strong hold in early times. This is abundantly
evidenced by the numbers of their remains in the area. A list
of the mounds or " dibbas " — as they are called — indicates about
three hundred in the Kistna district alone ; probably most of
these cover ancient remains, and, as excavation has shown, many
of them were stupas.1 How many may exist in neighbouring
districts we do not yet know. Unfortunately many, including
the largest of these, containing the most important remains,
have been used as quarries for brick and marble — not by natives
only — but by Government Public Works engineers, the record
of whose vandalism in utilising the materials is most deplorable.2

The stupa at Bhattiprolu, about 6 miles north of Repalle, was
for long one of the best preserved in the district — presenting a
circular mound or dome of 40 ft. or more in height, though ruined
at the top, a marble pillar 1 5 ft. high standing erect beside it, and
with clear indications of the procession-path round it.3 This was
destroyed, about 1 870 ; 4 what was left of it, on being surveyed in
1892, showed that the dome had been about 132 ft. in diameter,
while the basement was of about 148 ft, the procession-path
had been 8 ft. 4 in. wide and fenced on the outside by a
marble rail — of which the bases of six piers were found in situ.
Towards each of the cardinal points the base projected about
2 ft. 4 in., with a straight front, probably for the support of
the five monoliths — represented on all the sculptured stupas
from Amaravati, and as was the case at Jaggayyapeta.5 The
sculptures of the latter stupa, indeed, bear a close resemblance
in their archaic character, to the only two fragments recovered
here, and, so far as they go, indicate that this stupa may
have been of considerably earlier date than even the great
rail at Amaravati. We might suppose also that the sculptures
would be confined chiefly to the projecting facades, whilst the
rest of the basement was faced with plain slabs and pilasters.

At least one relic casket had been found by the first
excavators at a considerable height above the ground level ;
but the enclosing slabs, whether inscribed or not, were broken
and cast aside, and the casket was smashed on the voyage to
England and thrown away. During the survey, Mr. Rea dis-
covered three more relic boxes at a lower level, and bearing

1 ' Madras Government Orders,' No.
462 of 2gth May 1889.

2 See e.g. ' Madras Government Orders,'
No. 467, 30th April 1888, p. 15.

3 'Indian Antiquary,' vol. i. pp. 153,
348» 374 > ' Madras Journal of Literature
and Science,' vol. xix. p. 225 ; Sewell's
' Topographical Lists of Antiquarian
Remains,' pp. 77-78.

4 Ante, p. 34, note 3.

5 ' The Buddhist Stupas of Amaravati
and Jaggayyapeta,' p. lio. These pro-
jecting pedestals with their five Aryaka
(worshipful) columns, may be analogous
to the chapels for the Dhyani-Buddhas at
the bases of the dagabas of Ceylon and of
the chaityas of Nepal.

CHAP. III.

JAGGAYYAPETA, ETC.

inscriptions of considerable palaeographic interest, the alphabet
of which can hardly be placed later than 200 B.C.1

At Gudivada, 20 miles north-west from Masulipatam, there
was a "dibba" containing the remains of a stupa, which was
also demolished by the local engineers, it is said about 1860,
and so little was left of it that its dimensions cannot now be
ascertained. Four relic caskets are said to have been found,
though we can learn nothing about their age ; but considerable
numbers of Andhra coins, mostly of lead, are turned up about the
site belonging, probably, to the first three centuries of our era.2

The stupa at Jaggayyapeta or Betavolu, 30 miles north-
west from Amaravati, had been plundered of its rail, and of
much of the marble casing of its basement, the dome had been
destroyed, and relic casket dug out before it was surveyed
in 1882. The basement was 31^- ft. in diameter, and portions
of the facing remained, chiefly on the south side where the
slabs on the projection for the support of the five stelae bore
archaeic sculptures. The procession-path had been about 5 ft.
wide, surrounded by a rail or wall, of which every fragment
had disappeared.8

At Ghantayala, 13 miles west from Masulipatam, a mound
was surveyed by Mr. Rea in 1892, and was found to contain
the remains of a stupa with a diameter at the ground level
of about in ft. contained by a circular brick wall 18 ft. thick,
forming the drum of the stupa. Inside this was a curious
reticulation of walls, between which the spaces were packed
with black earth : a circular wall of 56 ft. outside diameter
was connected by sixteen radiating partitions with the outer
wall ; and inside this was a square of 26 ft, in the middle
of which was a column of brick 10 ft. square, joined to the
preceding by four partitions from the middle of its sides,
which ran right through the whole interior, while the sides of
the outer square were continued to the inner circle.* In the
centre of the column was a well, varying in width from 9 in.
to 2 ft, 6 in. square, in which was a relic casket, but without
anything to indicate its age. On each of the four faces of
the base were projections, as at Jaggayyapeta, about 17 ft. in
length by 5 1 ft. broad and 4^ ft. high. Sculptured slabs were

1 ' Epigraphia Indica,' vol. ii. , preface,
pp.ix.-xiii. , and pp. 323-329; Rea's' South
Indian Buddhist Antiquities,' pp. 7-16 ;
and conf. Fleet 'Journal of the Royal
Asiatic Society,' 1903, pp. 996°.

2 Sewell, ' Topographical Lists of
Antiquarian Remains, 'p. 52 ; Rea, op.
<r;V.pp. 18-23.

3 ' The Buddhist Stupas of Amaravati

and Jaggayyapeta,' pp. 107-113, and
plates 53-55.

4 Rea, ' South Indian Buddhist
Antiquities,' plate 14 ; a less complex
arrangement of interior partitions was
found by Dr Fiihrer in the Kankali-Tila
at Mathura. — Foucher, ' L'Art Greco-
Bouddhique,' torn. i. p. 95 ; and V. Smith,
'Jain Stupa,' plates I and 3.

BUDDHIST ARCHITECTURE.

BOOK I.

found in the neighbourhood bearing a strong resemblance to
those of Amaravati ; and some had been utilised to carve on
them Hindu divinities.1

It is not quite clear how this stupa was completed ; possibly
there may have been more than one platform round the dome ;
but we may suppose, as is most probable, that the massive brick
wall formed the true basement or drum, faced with sculptures,
and crowned by a procession - path round the dome, as at
Sanchi ; this latter may have had a diameter of about 84 ft,
rising as a hemisphere to about 54 ft. from the ground.2

At many other places, both in the Godavari and Kistna
districts, remains of other stupas, as also of rock-temples and
other Buddhist antiquities, including some structural chaitya
chapels, have been found — testifying to the predominance of
Buddhism in this province, and their prosperity for some
centuries before and after the Christian Era. Jaina images
are also met with — evidencing their spread southwards from
Orissa.3

GANDHARA TOPES.

The extreme paucity of examples retaining their archi-
tectural form, in the valley of the Ganges, is, to some extent,
compensated for by the existence of a very extensive range
of examples in Afghanistan and the western Panjab. In his
memoir of these topes, published by Professor Wilson, in his
' Ariana Antiqua,' Mr. Masson enumerates and describes, in
more or less detail, some sixty examples,4 or almost exactly
the same number which General Cunningham described as
existing in Bhopal. In this instance, however, they extend
over a range of 200 miles, from Kabul to the Indus, instead
of only 1 6 or 17 miles from Sonari to Andher. To these must
be added some fifteen or twenty examples, found at Manikyala

1 One slab had represented a fine
stupa, with Buddha, having a halo about
his head, as the central figure, and with
the usual five stelae above ; another,
represented a Bodhi tree, under which
is a throne on which lie two round
objects like cushions (perhaps relics),
whilst, curiously enough, the figures at
each side are not the usual worshippers,
but Mara and his hosts of Marakayakas.
Photographs of such sculptures would be
preferable to pen and ink drawings. —
Rea, loc. cit. plates 27 and 28.

2 Mr Rea, loc. cit., pp. 33-41, has pro-
posed a somewhat fanciful theory of the
construction of this stupa.

3 Since the publication of the volume
on the Amaravati and Jaggayyapeta stupas

in 1886, the sole addition has been Mr.
Rea's report on the excavation of three
sites — Bhattiprolu, Ghantajala, and
Gudivada. Detailed accounts of the
rock-temples at Guntupalle or Jilligera-
gudem, and of the structural chaityas at
Chezarla, Vidyadharapuram, .Sankaram,
etc., illustrated with photographs and
plans, would be of great archaeological
importance. See below, page 167.

* Masson, however, distinguished be-
tween topes, of which portions of the
masonry were visible, and "mounds"
that, in most instances, cover the remains
of stupas, such as the Ahin-posh Stupa,
and in some cases at least, they cover
whole groups of stupas.

CHAP. III.

GANDHARA TOPES.

or in its neighbourhood, together with those discovered in Swat
and on the north-west frontier within the last twenty-five years,
and it is certain that numbers still exist undescribed, making
altogether quite a hundred stupas in this province.

Notwithstanding this wealth of examples, we miss one,
which was probably the finest of all. When Fah Hian passed
through the province in A.D. 400, he describes the dagaba
which King Kanishka had erected at Peshawar as " more than
470 ft. in height, and decorated with every sort of precious
substance, so that all who passed by, and saw the exquisite
beauty and graceful proportions of the tower and the temple
attached to it, exclaimed in delight that it was incomparable
for beauty " ; and he adds, " Tradition says this was the highest
tower in Jambudwipa." l When Hiuen Tsiang passed that
way more than 200 years afterwards, he reports the tower
as having been 400 ft. high, but it was then ruined — " the
part that remained, a li and a half in circumference (1000
feet) and 1 50 ft. high " ; and he adds, in twenty-five stages of
the tower there were a "ho" — 10 pecks — of relics of Buddha.2
No trace of this monument now exists.

These north-western stupas are so important for our history,
and all have so much that is common among them, and are
distinguished by so many characteristics from those of India
Proper, that it would be extremely convenient if we could
find some term which would describe them without involving
either a theory or a geographical error. The term Afghanistan
topes, by which they have been designated, is too modern,
and has the defect of not including Peshawar and the western
Panjab. " Ariana," as defined by Professor Wilson, describes
very nearly the correct limits of the province ; for, though it
includes Baktria and the valley of the Upper Oxus, where
no topes have yet been found, we know from the Chinese
Pilgrims that in the 5th and 7th centuries these countries, as
far as Khotan, were intensely Buddhist, and monuments exist
there, and have recently been found in Khotan both by Dr. Sven
Hedin and by Dr. M. A. Stein.3 The name, however, has of
late almost disappeared in favour of Gandhara — the early Indian
name of the eastern portion of the district under notice.

When the Sanskrit-speaking races first broke up from their
original settlements in the valley of the Oxus, they passed
through the valley of the Kabul river on their way to India,

1 Beal's ' Fah-Hian,' p. 35 ; ' Buddhist
Records,' vol. i., introd. p. xxxii.

2 ' Vie et Voyages de Hiouen Thsang,'
torn. i. p. 8-) ; Beal, ' Life of Hiuen
Tsiang,' p. 63.

3 Stein's ' Archjeological Exploration
in Chinese Turkistan ' (1901); 'Sand-
buried Ruins of Khotan' (1903); and
'Ancient Khotan,' 2 vols. (4to), 1907.

86 BUDDHIST ARCHITECTURE. BOOK I.

and lingered, in all probability, both there and in the Panjab
before reaching their first permanent position on the Saraswati
— in the true " Arya Varta " — between the Satlaj and the Jamna.
It is also nearly certain that they remained the dominant race
in these countries down to the time of Alexander's invasion, and
during the supremacy of the Baktrian kingdom. About 150
years, however, before the Christian Era, if we may trust the
Chinese accounts,1 the Yue-chi, and other tribes of Tartar origin,
were on the move in this direction. Somewhat later they
struck down the Baktrian monarchy, and appear from thence-
forward to have permanently occupied their country. It is
not clear whether they immediately, or at what interval they
penetrated into the Kabul valley ; but between that time and
the Christian Era successive hordes of Yue-chi, Sakas, Turushkas,
and Hunas, had poured into the valley and the western Panjab
to such an extent as to obliterate, or, at least for the time,
supersede the Aryan population, and supplant it by one of
Turanian origin, and with this change of race came the change
of religion. Gandhara is, however, a local name, which certainly,
in early times, included the best part of this province, and in
Kanishka's time seems to have included all he reigned over,
and, if so, is the most appropriate term we could find.

It has, moreover, this advantage, that it is essentially
Buddhist. In the time of A^oka, it was Kashmir and Gandhara
to which the Buddhist Council sent its missionaries, and from
that time forward Gandhara is the term by which, in all
Buddhist books, that kingdom is described, of which Taxila was
at one time the capital, and which is, as nearly as can now be
ascertained, conterminous with our architectural province.

It is not clear whether Kanishka was or was not the first
Buddhist king of this country ; but, so far as is at present
known, he seems to have done for Buddhism in Gandhara what
A^-oka did for that religion in Central India. He elevated it
from its position as a struggling sect to that of being the
religion of the State. We know, however, that A^oka's Council
sent missionaries to this country ;2 and, more than this, that he
engraved a complete set of his edicts on a rock at Kapurdigiri,
30 miles north-east from Peshawar, but we do not know what
success they or he attained. Certain it is, as Professor Wilson
remarks, that " no coin of a Greek prince of Baktria has ever
been met with in any tope." 3 The local coins that are found
in them all belong to dynasties subsequent to the destruction
of the Baktrian kingdom, and, according to the same authority

1 De Guigne's ' Histoire des Huns,' vol. ii. pp. 40 et seqq.

2 Tumour's ' Mahawansa,' p. 71. 3 'Ariana Antiqua,' p, 43,

CHAP. III.

GANDHARA TOPES.

(p. 322), "were selected from the prevailing currency, which was
not of any remotely previous issue ; " " while the Graeco-Baktrian
coins had long ceased to be current, though they had not,
perhaps, become so scarce as to be enshrined as rarities " (p. 44).
Under these circumstances, Professor Wilson arrives at the con-
clusion that the topes " are undoubtedly all subsequent to the
Christian Era" (p. 322). It is true that some of the kings
whose coins are found in the topes, such as Hermaeus, Azes,
and others, probably lived prior to that epoch, but none of
their coins show a trace of Buddhism. With Kanishka, how-
ever, all this is altered. He is represented as a Buddhist,
beyond all doubt ; he held the convocation, called the third by
the northern Buddhists — the fourth according to the southern
— at which Nagarjuna was apparently the presiding genius.
From about that time the Tibetans, Burmese, and Chinese
date the first introduction of Buddhism into their countries ;
not, however, the old simple Buddhism, known as the Hinayana,
which prevailed before, but the corrupt Mahayana, which, as
a new revelation, Ndgarjuna spread from Peshawar over the
whole of central and eastern Asia. It was precisely analogous
to the revolution that took place in the Christian Church, about
the same time after the death of its founder. Six hundred
years after Christ, Gregory the Great established the hierarchical
Roman Catholic system, in supersession of the simpler primitive
forms. In the fifth century after the Nirvana, Nagarjuna intro-
duced the complicated ritualistic and idolatrous Mahayana,1
though, as we learn from the Chinese Pilgrims, a minority still
adhered in after times to the lesser vehicule or Hinayana
system.

Although, therefore, we are probably safe in asserting
that none of the Gandhara stupas date much before the
Christian Era, it is not because there is any inherent, d priori
improbability that they should date before Kanishka, as there
is that those of India Proper cannot extend beyond A^oka.
There is no trace of wooden construction here : all is stone
and all complete, and copied probably from originals that
may have existed two centuries earlier. Their dates depend
principally on the coins, which are almost invariably found
deposited with the relics, in these topes. Coins have rarely
been found in any Indian tope.2 They are found in hundreds
in these north-western ones, and always fix a date beyond
which the tope cannot be carried back, and generally enable
us to approximate to the true date of the monument in

1 Vassilief, ' Le Bouddhisme, ses
Dogmes,' etc., Paris, 1865, p. 31, et
passim.

2 A silver coin of one of the Andhra
kings, belonging to the 2nd century A. D.,
was found in the Sopara stupa.

88

BUDDHIST ARCHITECTURE.

BOOK I.

question. If those of Kanishka are the earliest, which appears
to be the case, the great one which he commenced, at
Manikyala, is probably also the last to be finished in its
present form, inasmuch as below 12 ft. of solid masonry, a
coin of Yajovarman of Kanauj was found, and his date cannot
be carried back beyond A.D. 720. Between these dates, there-
fore, must be ranged the whole of this great group of Buddhist
monuments.

There were perhaps no great Buddhist establishments in
Gandhara before Kanishka, and as few, if any, after the 8th
century, yet we learn that, during the earlier part of the period
between these dates this province was as essentially Buddhist as
any part of India. Fah Hian tells us, emphatically, that the law of
Buddha was universally honoured, and mentions 500 monasteries,1
and Hiuen Tsiang makes no complaint of heretics, while both
dilate in ecstasies on the wealth of relics everywhere displayed.
Part of the skull, teeth, garments, staffs, pots of Buddha —
impressions of his feet, even his shadow — was to be seen in this
favoured district, which was besides sanctified by many actions
which had been commemorated by towers erected on the spot
where these meritorious acts were performed. Many of these
spots have been identified, and more will no doubt reward the
industry of future investigators, but meanwhile enough is known
to render this province one of the most interesting of all India
for the study of the traditions or art of Mediaeval Buddhism.

The antiquities of the western part of the province were
first investigated by Dr. Honigberger, in the years i833-i834,2 and
the result of his numismatic discoveries published in Paris and
elsewhere ; but the first account we have of the buildings them-
selves is that given by Mr. Masson, who, with singular per-
severance and sagacity, completed what Dr. Honigberger had
left undone.3 Those of the eastern district and about Manikyala
were first investigated by General Ventura and M. Court, officers
in the service of Ranjit Singh, and the result of their researches
published by Prinsep in the third volume of his ' Journal ' in
1 830 ; but considerably further light was thrown on them by
the explorations of General Cunningham, published in the
second volume of his 'Archaeological Reports' for 1863-1864
(pp. 82 et seqq.}. Since then still further additions have been

1 Beal's translation, p. 26 ; ' Buddhist
Records,' introd. pp. xxx. , xxxi.

2 Honigberger, ' Reise' ; see also J. G.
Gerard's 'Memoir' in 'Jour. Asiat. Soc.
Beng.' vol. iii. p. 321 ; and Jacquet's
'Notice' in 'Jour. Asiatique' I He.
serie, tome ii. p. 234 ; iv. p. 401 ; and
vii. p. 385.

3 Mr. Masson's account was com-
municated to Professor Wilson, and by
him published in his ' Ariana Antiqua,'
with lithographs from Mr. Masson's
sketches which, though not so detailed
as we could wish, are still sufficient to
render their form and appearance
intelligible.

CHAP. Ill, JALALABAD TOPES. 89

made to our knowledge of the Gandhara remains. In 1879 the
late Mr. Wm. Simpson, whilst accompanying the British forces,
excavated the Ahin-posh stupa, to the south of Jalalabad, and
his account of it formed a valuable contribution to our knowledge
of this and other remains in the valley.1 Under Major Cole
and others, extensive excavations were also made in Yusufzai
and Swat in search of sculptures ; but plans of the structures
have but seldom been secured.2 Lastly, in 1895-1897, a mission
was sent out from France, under the direction of Mons. A.
Foucher, to make a scientific survey of the Buddhist remains in
the Yusufzai and Swat districts. The first volume of the results
of this expedition was published in 1905, containing an account
of the whole subject, abundantly illustrated and thoroughly
scientific.3

JALALABAD TOPES.

The topes examined and described by Mr. Masson as exist-
ing round Jalalabad are thirty-seven in number, viz., eighteen
distinguished as the Daranta group, six at Chahar Bagh, and
thirteen at Hidda. Of these about one-half yielded coins and
relics of more or less importance, which proved the dates of
their erection to extend from somewhat before the Christian Era
to the 7th or 8th century.

One of the most remarkable of these is No. 10 of Hidda or
Hada, which contained, besides a whole museum of gems and
rings, five gold solidi of the emperors Theodosius (A.D. 408),
Marcian and Leo (474) ; two gold Kanauj coins ; and 202
Sassanian coins extending to, if not beyond, the Hijra.4 This
tope, therefore, must belong to the 7th century, and would be a

1 ' Transactions of the Royal Institute
of British Architects,' 1879-80, pp. 37-64.
The plates in Gen. Cunningham's paper,

to a native jamadar, is severely criticised
(' L'Art Greco-Bouddhiquedu Gandhara,'
tomei. p. 2 1 ), as is also the destruction of

— ' Proceedings of the Asiatic Society of interesting ancient structures by the Public

Bengal,' 1879, pp. 205-210, supplied to Works Department. — Foucher, in ' Le

him by Lieut. Mayne, R.E. — do not add j Tourdu Monde, 1899,' p. 486, and 'L'Art

to Mr. Simpson's drawings. Mr. J. D. ; Greco-Bouddhique,' pp. I4ff, 47, etc.

Beglar, Gen. Cunningham's assistant, was ; 3 Foucher, ' L'Art Greco Bouddhique

sent to excavate the stupas at 'Ali Masjid, ; du Gandhara,' contains a masterly account

at the same time, but no account of his I of the origin of the classical influence on

work seems to have been given — not even I Buddhist art in India and the East, and

a plan of the stupas excavated ; a few | is richly illustrated from his surveys and

photographs in the Calcutta Museum (of | the sculptures stored in various museums,

which six were published in the ' Ancient j * The length of time over which these

Monuments of India,' plates 103-108) are j coins range — more than 200 years — is

the only evidence of this excavation. ! sufficient to warn us what caution is

2 The way in which the excavations of | requisite in fixing the date of buildings

so many of these sites were conducted, \ from their deposits. A tope cannot be

must ever be regretted. Major Cole's i earlier than the coins deposited in it, but,

" lointaine direction" of the excavations as in this case, it may be one or two

in Yusufzai, leaving the entire supervision hundred years more modern.

BUDDHIST ARCHITECTURE.

BOOK I.

most convenient landmark in architectural history, were it not
that the whole of its exterior is completely peeled off, so that
no architectural mouldings remain, and, apparently from the
difficulty of ascertaining them, no dimensions are quoted in
the text.1 About one - half of the others contained relics, but
none were found to be so rich as this.

In general appearance they differ considerably from the great
Indian topes just described, being taller in proportion to their
breadth, and having a far more tower-like appearance, than any
found in India, except the Sarnath example. They are mostly
smaller, one of the largest at Daranta being only 160 ft in
circumference or about 51 ft. in diameter. This is about the
usual size of the topes in Afghanistan, the second class being a
little more than 100 ft, while many are much smaller. There
are, however, some of larger size, for Mr. William Simpson
found the circumference of the Ahin-posh stupa to be about
250 ft., and of the Umar Khel tope at Daranta, a rough
measurement gave 300 ft. circumference to the circular drum.2

In every instance they seem to have rested on a square
base, though in many this has been removed, and in others it is
buried in rubbish ; in many cases also, if not always, there was a
deep plinth or low terrace below this base. Above this rises a
circular drum, crowned by a belt sometimes composed merely of
two architectural string-courses, with different coloured stones
disposed as a diaper pattern between them. Sometimes a range
of plain pilasters occupies this space. More generally the
pilasters are joined by arches sometimes circular, sometimes
of an ogee form. In one instance — the Red (Surkh) Tope — they
are alternately circular and three-sided arches. That this belt
represents the enclosing rail at Sanchi and the pilastered base
at Manikyala need not be doubted. It shows, however, a very
considerable change in style to find it elevated so far up the
monument as it is here.

Generally speaking, the dome or roof rises immediately
above this, but no example in this group retains its termination
in a perfect state. Some appear to have had hemispherical
roofs, some more nearly conical, of greater or less steepness of
pitch ; and some (like that represented in Woodcut No. 21) had,
perhaps, only a slight elevation in the centre. It seems possible
there may have been some connection between the shape of the
roof and the purpose for which the tope was raised, or the age

1 ' Ariana Antiqua,' p. 109.

- ' Transactions of the Royal Institute
of British Architects,' 1879-80, p. 53. The
Nagara-gundi tope, at old Nagarahara,
appears to have been also of similar size.

One between Cherat and Giiniyar passes
in Swat, is 140 ft. in circumference at
the bottom of the cylindric part. —
Foucher, ' L'Art Greco-Bouddhique du
Gandhara,' tome i. p, 65.

CHAP. III.

JALALABAD TOPES.

21. Tope at Bimaran. (From a Drawing by Mr. Masson,
in Wilson's ' Ariana Antiqua.')

of its erection. But we have no evidence to lead us to any
decision of this point.

One interesting peculiarity was brought to light by Mr.
Masson in his
excavation of the
tope at Sultanpur,
9 miles west of
Jalalabad, and is
shown in the an-
nexed section
(Woodcut No. 22).
It appears that the
monument origin-
ally consisted of a
small tope on a
large square base,
with the relic
placed on its sum-
mit. This was
afterwards in-
creased in size by a
second tope being built over it.1

Among the later discoveries in the North- West Frontier
districts, the following may be mentioned as adding to our
knowledge of these stupas.
At Chakpat near Chakdarra
fort in the Swat valley, a
mound was excavated by
Mr. A. Caddy in 1896, in
search of sculptures for the
Calcutta Museum, and was
found to contain a small
stupa in the form of a simple
hemispherical dome, about 20
ft. in diameter. This unique
form, probably one of the
earliest types for such a
structure, points to the derivation of these structures from
the simple tumulus (Woodcut No. 23). The dome was fairly
complete and was encircled by the base of a wall about 30 ft in
diameter. This was evidently the remains of an outer casing
that had been built at a later date over the original structure —
enlarging it and probably altering the contour. The debris of
this — when the stones of the outer surface were carried off —

22. Tope at Sultanpur. (From a Drawing by
Mr. Masson, in Wilson's ' Ariana Antiqua.')

1 At Hada, near Jalalabad, Mr. Simpson found a somewhat similar instance, -
'Transactions of the Royal Institute of British Architects,' 1879-80, p. 56.

BUDDHIST ARCHITECTURE.

BOOK I.

had concealed and helped to preserve the original. A circular
stone disc, of quite 1 1 ft. 6 in. in diameter, which had formed
one of the chhattras or umbrellas crowning the larger stupa,

23-

Stupa at Chakpat. (From a Photograph.)

and had slid down from the higher dome when first ruined,
is seen in the illustration.1

Half a mile below 'Alt Masjid in the Khaibar Pass, on a
small hill, are the remains of a religious establishment sur-
rounded by a group of ruined stupas of a very interesting
character. They were excavated in 1879 by General Cunning-
ham's assistant, and are said to have yielded important materials,
never published. Little more than the bases of these stupas
remained, but they were very rich in stucco figure decoration ;
the accompanying reproduction of a photograph of No. 5 of the
series (Plate i) will convey some idea of their form.2

Beyond 'Ali Masjid and near Lalabeg is the Ishpola tope
— M. Court's Pishbulak 3 — placed on the summit of a rock
projecting into, and dominating the valley. It appears about
nearly as large as the Manikyala stupa, and, like it, the
hemispherical dome that crowns it, is only slightly raised
on a short cylindrical neck resting on a square base which
is further supported by walls forming a second and lower
quadrangular terrace. The base had fourteen pilasters on

1 Foucher, ' L'Art Greco-Bouddhique,'
tome i. pp. 56-59, 67, and 74.

2 Burgess, ' Monuments of Ancient
India,' plate 106. Three coins of Vasu-
deva, the third Kushan King, were found

here, but the stupas can hardly be
ascribed to an earlier date than the 3rd
century A.D.

3 'Journal of the Asiatic Society of
Bengal,' vol. v. p. 393.

PLATE I.

;^ ?

-r .^M.. " *"' -•" •' •

ALI MASJID STUl'A NO. 5.

[To face page cp, Vol. I.

CHAP. III.

MANIKYALA.

93

each side with quasi-Corinthian capitals ; and three of the
interspaces have recesses as if for images of the Buddha, whilst
on the lower member of the base was a continuous row of
Buddhist figures about 18 in. high and placed close together.
The stair leading to the top of the base was on the east
side, and both the base and south side of the dome were
still fairly entire in iS?g.1

Of the Ahin-posh stupa mentioned above, only the base
remained and was 100 ft. square with extensions on the sides
for the stairs, of which the principal appear to have been
on the north and south sides — those on the other two sides
having been added afterwards. Parts of the lowest course
of the drum or dome showed that its diameter had been
80 ft., and the base seems to have been about 23 ft. high,
with fourteen pilasters on the east and west faces. Under
this base is a plinth 5^ ft. high and projecting 6 ft. In the
relic chamber were found three Roman coins of Domitian
(A.D. 81-96), Trajan, and the empress of Hadrian (117-138),
and seventeen of Kanishka, Huvishka and Kadphises : hence
this stupa can hardly be ascribed to an earlier date than the
end of the 2nd century, but probably considerably later.2

The most imposing ruin of a stupa noticed in Swat is one
between the passes of Cherat and Guniyar, to the south of
Chakdarra, of which the circumference of the drum is about
140 ft, or 45 ft. in diameter. Like the Ishpola stupa, its
dome rises on a double drum, by which the vertical lines in
the outline seem to gain in importance at the cost of the
downward curve. The case is similar with the Barikot tope,
about 9 miles up the valley from Chakdarra, which, as
M. Foucher points out, is very probably the Uttarasena stupa
mentioned by Hiuen Tsiang.3 It has lost its square base
or platform, of which the stones have been used to build
modern walls in the vicinity, but it has preserved the belt
of arches which divides the cylindrical portion into two sections,
one above the other ; and the dome is pretty entire.

About a mile south-east from Haibatgram, in the same
locality, is another stupa, hid away in a valley that takes
from it the name of Top-darra. It has a circular base about
31 ft. in diameter and 7 ft. in height, introduced above the
square base. This latter is some 12 ft. high and 58 ft. square,
has nine pilasters on each face, and remains almost intact

1 Simpson, ' Transactions of the Royal
Institute of British Architects,' 1879-80,
pp. 40, 41 ; Foucher, ' L'Art Greco-
Bouddhique,' tome i. p. 74 and fig. 14.

2 Simpson, ut sup. p. 49.

3 Beal's 'Buddhist Records,' vol. i.
pp. 126-127, and 132-133; Julien,
' Memoires,' tome ii. pp. 139, 146-149.
A view is given in Foucher, ' L'Art
Greco-Bouddhique,' tome i. p. 67.

94

BUDDHIST ARCHITECTURE.

BOOK I.

except the stair on one side ascending to the platform, which
a trifling excavation would probably disinter.1 The height
of this structure to the top of the dome must have been nearly
40 ft.

Besides those already mentioned there are about twenty or
thirty topes in the neighbourhood of Kabul, but all much ruined,
and few of any striking appearance. So at least we are led to
infer from Mr. Masson's very brief notice of them. No doubt
many others still remain in spots hitherto unvisited by
Europeans.

In the immediate vicinity of these topes are found caves and
tumuli, the former being the residences of priests, the latter
partly burying - places, and partly ruined viharas, perhaps in
some instances smaller relic-shrines. Their exact destination
cannot be ascertained without a careful investigation by persons
thoroughly conversant with the subject. There are still, also,
many other points of great interest which require to be cleared
up by actual examination. When this has been done we may
hope to be able to judge with some certainty of their affinity
with the Indian buildings on the one hand, and those of Persia
on the other.

MANIKYALA.

The most important group, however, of the Gandhara topes
is that at Manikyala in the Panjab, situated between the Indus
and the Jehlam or Hydaspes, about 20 miles south-east of
Rawalpindi. Fifteen or twenty examples are found at this
place, most of which were opened by General Ventura and
M. Court about the year 1830, when several of them yielded
relics of great value, though no record has been preserved of the
greater part of their excavations. In one opened by M. Court,
a square chamber was found at a height of 10 ft. above the
ground level. In this was a gold cylinder enclosed in one
of silver, and that again in one of copper. The inner one
contained four gold coins, ten precious stones, and four pearls.
These were, no doubt, the relics which the tope was intended
to preserve. The inscription has been read, and is dated in the
eighteenth year of Kanishka,2 so that we may feel assured it
was erected during, or not long after, his reign. Seven Roman
coins were found much worn, as if by long use,3 before they
reached this remote locality ; and, as they extend down to

1 Foucher, loc. cit. pp. 70, 71 and 74,
and figs. 17, 1 8.

2 'Journal Asiatique,' IX* Ser., tome

vii. , 1896, pp. 1-25.

3 'Journal of the Asiatic Society of
Bengal,' vol. iii. p. 559.

CHAP. III.

MAN IKY ALA.

95

B.C. 43,1 it is certain the monument was erected after that
date. The gold coins were all those of Kanishka. This tope,
therefore, could hardly have been erected earlier than thirty
years before Christ. To the antiquary the enquiry is of con-
siderable interest, but less so to the architect, as the tope is
so completely ruined that neither its form nor its dimensions
can now be distinguished.

Another was opened in 1863 by General Cunningham, in
the relic chamber of which he found a copper coin, belonging
to the Satrap Zeionises, who is
supposed to have governed this
part of the country about the be-
ginning of the Christian Era, and
we may therefore assume that the
tope was erected by him or in his
time. This and other relics were
enclosed in a glass-stoppered vessel,
placed in a miniature representa-
tion of the tope itself, 4^ in. wide
at base, and 8|- in. high (Woodcut
No. 24), which may.be considered
as a fair representation of what
a tope was or was intended to be,
in that day. It is, perhaps, taller,
however, than a structural example
would have been ; and the tee,
with its four umbrellas, is, possibly,
exaggerated.2

The principal tope of the group
is, perhaps, the most remarkable of
its class in India, though inferior
in size to several in Ceylon. It
was first noticed by Mountstuart
Elphinstone, and a very correct
view of it published by him, with the narrative of his mission
to Kabul in iSoQ.3 It was afterwards thoroughly explored by
General Ventura, in 1830, and a complete account of his investi-
gations published by Prinsep in the third volume of his 'Journal.'
Since then its basement has been cleared of the rubbish that
hid it to a depth of 12 ft. to 15 ft. all round, by the officers

24.

Relic Casket from Tope at
Manikyala. (Found and drawn
by Gen. Cunningham.)

1 Thomas in ' Prinsep,' vol. i. p. 150;
and 'Journal R. Asiatic Soc. ,' vol. ix.
(N.S.), pp. 217-218.

2 ' Archaeological Reports,' vol. ii. p.
167, plate 65. A similar reliquary, with
five umbrellas or chattras, was found by
Gerard in the Burj-i-yak-dereh tope to

the east of Kabul; Jacquet, in 'Jour.
Asiatique,' III6 serie, tome vii. pp. 394-
395 ; Foucher, 'L'Art Greco-Bouddhique
du Gandhara,' tome i. pp. iqf. and 75.

3 Elphinstone's 'Account of Caubul,'
pp. 78, and 376, ist ed.

BUDDHIST ARCHITECTURE.

BOOK I.

of the Public Works Department. They also made careful
plans and sections of the whole.1

From these it appears that the dome is an exact hemisphere,
127 ft. in diameter, and consequently as nearly as may be, 400
ft. in circumference. The outer circle measures in like manner
159 ft. 2 in., or 500 ft. in circumference, and is ascended by
four very grand flights of steps, one in each face, leading to
a procession-path 16 ft. in width, ornamented both above
and below by a range of dwarf pilasters, representing the
detached rail of the older Indian monuments. It is, indeed,
one of the most marked characteristics of these Gandhara topes,
that none of them possess, or ever seem to have possessed,
any trace of an independent rail ; but most have an ornamental

belt of pilasters, joined generally by arches simulating the
original rail. This can hardly be an early architectural form,
and leads to the suspicion that, in spite of their deposits, their
outward casing may be more modern than the coins they
contain ; yet, on the other hand, we must admit that the simple
hemispherical dome, without drum, placed on a low platform,

1 Cunningham, 'Archaeological Reports,' vol. v. pp. 75-79, and plates 21-24.

CHAP. III.

MANIKYALA.

97

(as at Chakpat) must have been an early form, and that the rail
was a feature confined to purely Indian stupas, whilst among
those of the Gandhara region it never appeared.

The outward appearance of the Manikyala tope, in its half-
ruined state, may be judged of from the view (Woodcut No. 25).
All that it really requires to complete its outline is the tee,
which was an invariable adjunct to these buildings : no other

26. Restored Elevation of the Tope at Manikyala. Scale 50 ft. to i in.

feature has wholly disappeared. The restored elevation, half-
section, half- elevation (Woodcut No. 26), to the usual scale,
50 ft. to i in., will afford the means of comparison with other
monuments ; l and the section and elevation of the base.
Woodcut No. 27, on the next page, will explain its architectural
details in so far as they can be made out.

On digging into this monument, General Ventura found
three separate deposits of relics, arranged at apparently equal
distances of 25 ft. from the surface of the finished monument
and from each other, and each apparently increasing in value
or importance as it descended. The first was at the base of a
solid cubical mass of squared masonry, and contained, inter a/z'a,
some Sassanian coins and one of Yajovarman (about A.D. 720),
and one of Abdullah bin Hazim, struck at Merv A.H. 66, or
A.D. 685.2 The second, at a depth of 50 ft., contained no coins.
The principal deposit, at a depth of 75 ft, was on the exact
level of the procession-path outside. It consisted of a copper
vessel, in which was a relic casket in brass, represented in the

1 The restored elevation here, omits
the stairs (sapana) in front and at the
sides, as also the umbrellas that crowned
the whole. — See Cunningham, 'Archceo-

VOL. I.

logical Survey Reports,' vol. v. plates
21 and 22.
2 Thomas's ' Prinsep,' vol. i. p. 94.

G

BUDDHIST ARCHITECTURE.

BOOK I.

woodcut No. 28 on next page, containing a smaller vessel of
gold, filled with a brown liquid, and with an inscription on the

PART OF FRONT ELEVATION.

27. Elevation and Section of portion of Basement of Stupa at Manikyala.

lid which has not yet been fully deciphered, but around it were
one gold and six copper coins of the Kanishka type.

If this were all, it would be easy to assert that the original
smaller tope, as shown in the section (Woodcut No. 26), was
erected under Kanishka, or in his time, and that the square
block on its summit was the original tee, and that in the 8th
century an envelope 25 ft. in thickness, but following the
original form, was added to it,1 and with the extended procession-
path it assumed its present form, which is very much lower than
we would otherwise expect from its age.

Against this theory, however, there is an ugly little fact. It
is said that a fragment2 or, as it is printed, three Sassanian coins

1 It is not to be assumed that, when
a stupa was enveloped by an addition, the
enlarged form was symmetrical with the
original ; rather it would usually add pro-
portionately more to the height than to
the diameter.

2 In the text it is certainly printed
"three" with a reference to 19 in the
plate 21 of vol. iii. The latter is un-
doubtedly a misprint, and I cannot help
believing the former is so also, as only

one fragment is figured ; and Prinsep
complains more than once of the state
of the French MS. from which he was
compiling his account. I observe that
Gen. Cunningham adopts the same
views. At p. 78, vol. v., he says: "I
have a strong suspicion that Gen.
Ventura's record of three Sassanian coins
having been found below deposit B may
be erroneous."

CHAP. III.

MANIKYALA.

99

were found at a depth of 64 ft. (69 ft. from the finished surface) ;
and if this were so, as the whole masonry was found perfectly
solid and undisturbed from the surface to the
base, the whole monument must be of the age
of this coin. As engraved, however, it is such
a fragment l that it seems hardly sufficient to
base much upon it. Unless the General had
discovered it himself, and noted it at the
time, it might so easily have been mislabelled
or mixed up with other Sassanian fragments
belonging to the upper deposits that its posi-
tion may be wrongly described. If, however,
there were three, this explanation will not
suffice. It may, however, be that the princi-
pal deposit was accessible, as we know was
sometimes the case,2 in this instance at the
bottom of an open well-hole or side gallery,
before the time of the rebuilding in the 8th
century, and was then, and then only, built
up solid. If we may disregard this deposit,
its story seems self-evident as above explained.
But whatever its internal arrangements may
have been, it seems perfectly certain that its
present external appearance is due to a rebuilding, possibly as
late as the early part of the 8th century.

General Cunningham attempted to identify M. Court's tope
with that erected to commemorate the Buddha, in a previous
stage of existence, offering his body to appease the hunger
of a tigress, or — according to another version — of its seven
famishing cubs ; 3 but this was based on a mistaken reading of
some words in the inscription. The stupa of the " body-
offering" must have been far to the north-east of Taxila —
probably in the Hazara country.4 Unfortunately nothing of
the exterior coating now remains on any of the sixteen topes
at this place, and, what is worse, of all the fifty or fifty-five
which can still be identified at Taxila. As General Cunningham
remarks, of all these sixty or seventy stupas there is not one,
excepting the great Manikyala tope, that retains in its original
position a single wrought stone of its outer facing;5 none

28. Relic Casket,
Manikyala.

1 ' Journal of the Asiatic Society of
Bengal,' vol. iii. plate 21, fig. 18.

2 ' Foe Koue Ki,' ch. xiii. pp. 85-
86; Beal, 'Buddhist Records,' vol. i.
p. xxxiv.

3 S. Julien's ' Memoires sur les Con-
trees Occidentals, ' tome i. p. 164, and
'Vie de H. Th.' p. 89; or Beal's

' Buddhist Records,' vol. i. pp. 145^ ; and
' Life of H. T.' p. 67 ; Sp. Hardy's
' Manual of Budhism,' 2nd ed. pp. 93f.

4 E. Chavannes, ' Song Yun ' in
'Bulletin de 1'Ecole Fran9aise de 1'Ex-
treme-Orient,' tome iii. (19x33), p. 411.

5 ' Archaeological Reports,' vol. ii. p.
172.

BUDDHIST ARCHITECTURE.

BOOK I.

consequently, are entitled to a longer notice in a work wholly
devoted to architecture.

Over all India there must have been large numbers of stupas
in Buddhist times, though now so very few remain above
ground. There is, however, near Daulatpur in the Haidarabad
district of Sind, a large tope, known as Thai Rukhan, fully
50 ft. in diameter, and about the same in height. The inner
core is of sun-dried brick cased outside with good burnt bricks —
moulded for the cornices and capitals of pilasters. It has been
surrounded by a platform about 6 yards wide, now ruined and
covered with debris.1 The lower portion of the stupa is much
peeled and injured ; but above are two belts of pilasters — about
twenty in each — with moulded bases and quasi-Corinthian
capitals. Over the lower belt a cornice ran, above which the
diameter of the tower is contracted by perhaps 5 ft. In the
upper section the pilasters are better preserved, and though
the top is much ruined, the dome probably began at about
8 ft. above this.2 The bricks are very large, measuring 16 in.
in length by n in breadth and 3 in. thick. The stupa
probably belongs to an early age, and, as M. Foucher remarks,
in style, it appears to have descended directly from those of
the Swat valleys and Kabul.

In 1877 a stone box containing a crystal reliquary was
found embedded in brick when excavating a small mound at
Kolhapur ; the casket was broken, but the lid of the box bore
a short inscription in early letters, of the maker's name and of
the person for whom it was made.3

In 1882 Mr J. M. Campbell of the Civil Service excavated
the remains of a stupa at Sopara, 5 miles north of Basein.
The dimensions of it, given by Pandit Bhagwanlal Indraji are
unfortunately not consistent.4 If the diameter of the stupa was,
as stated, 67 ft, and the circular platform on which it stood
was 268 ft. in circumference and 18 ft. high, the ramp round
the dome would be just 9 ft. wide (not 18 ft). But possibly the

1 Among the debris were found some
27 bricks, about 7 inches square, each
bearing a small figure of Buddha, seated
with the legs down : these may have
formed part of a string course. Others
bore representations of the birth of
Buddha and of the bedroom scene before
his leaving home.

2 Burgess, ' Ancient Monts. of India,'
plate 62. Other stupas, more ruined,
have been noted in Sindh. At Kahu near
Mirpur Khas, one was excavated for
bricks when making the Haidarabad
Umarkot railway, and figures of Buddhas
moulded in brick and other ornamental

forms were found and appropriated by
officials. At Depar, 4 miles from
Brahmanabad, and atTando near Tando-
Muhammad - Khan, are brick mounds
which are remains of stupas. — ' Jour.
Bombay B. R. Asiatic Soc.' vol. xix.

P- 45-

3 'Jour. Bombay B. R. Asiat. Soc.'
vol. xiv. pp. 147-151.

* 'Jour. Bombay B. R. Asiat. Soc.' vol.
xv. pp. 292-311. It is twice stated (pp.
293 and 295) that the terrace was ' 18 ft.
wide.' The plan and section on plate
3 are evidently not drawn to any scale,
and afford no help.

CHAP. III.

STUPAS.

101

diameter of the stiipa was only 49 ft., and the height of it,
before excavation, must have been about 50 ft, The special
interest of the excavators, however, was in the relic caskets,
which, with some interesting figures of Buddhas, contained a
silver coin of Sri Yadnya Gautamiputra, who reigned about the
end of the 3rd century A.D.1

1 In 1889 Mr Campbell had the Boria
slupa, at the foot of Mount Girnar
excavated ; but the account given in the
' Journal of the Asiatic Society of Bengal,'

vol. Ix. pp. 17-23, is defective in details.
In April 1898 Mr. Cousens found the
basement of another stiipa at Sopara with
the empty relic casket.

29. Parinirvana of Buddha. From Cave 26 at Ajanta.

102 BUDDHIST ARCHITECTURE. BOOK I.

CHAPTER IV.

RAILS.
CONTENTS.

Rails at Bharaut, Mathura, Sanchi, and Amaravati.

IT is only within the last forty years or so, that our rapidly-
increasing knowledge has enabled us to appreciate the important
part which Rails play in the history of Buddhist architecture.
The rail of the great Tope at Sanchi has, it is true, been long
known ; but it is the plainest of those yet discovered, and
without the inscriptions which are found on it, and the gate-
ways that were subsequently added to it, presents few features
to interest any one. There is a second rail at Sanchi which
is more ornamented and more interesting, but it has not yet
been published in such a manner as to render its features
or its history intelligible. The great rail at Bodh - Gaya
is one of the oldest and finest of its kind, but, though it
was examined and reported on by Rajendralal Mitra and by
General Cunningham, neither of them added much to our
previous information. When the Amaravati sculptures were
brought to light and pieced together,1 it was perceived that
the rail might, and in that instance did, become one of the
most elaborate and ornamental features of the style. In 1863
General Cunningham found two or three rail pillars at Mathura
(Muttra), of an early Jaina stupa, but his discovery, in 1874,
of the great rail at Bharaut made it clear that this was the feature
on which the early Buddhist architects lavished all the resources
of their art, and from the study of which we may consequently
expect to learn most.

The two oldest rails of which we have any knowledge in
India are those at Bodh-Gaya and at Bharaut. The former,
General Cunningham thought, cannot be of much later date
than Ajoka.2 The latter has been ascribed to the period of

1 ' Tree and Serpent Worship,' Preface to the First Edition.

2 ' Archaeological Reports,' vol. i. p. 10.

CHAP. IV.

RAILS.

103

the Sunga dynasty, or about two centuries before the Christian
Era. On the whole, I am inclined to believe that the Bodh-
Gaya rail was really erected by Asoka, or during his reign.
At all events, we know from the fifteenth chapter of the
' Mahawansa ' that even if he did not worship this tree, he is
said to have reverenced it to such an extent that, as tradition
reports, when he sent his daughter Sanghamitta to aid in the
conversion of Ceylon to the Buddhist creed, he cut off and
entrusted her with a branch of this tree planted in a golden
vessel. That tree was replanted with infinite ceremony at
Anuradhapura, and it, or its lineal descendant, remains the
principal numen of the island to this day. Hiuen Tsiang tells
us that Ajoka built a small vihara to the east of the tree on
the spot where the present temple stands ; l and nothing is
consequently more probable than that he should have added
this rail, which is concentric with his vihara, but not with the /
tree.

There certainly is no inherent improbability that he should
have done so, for it seems hardly doubtful that this was the
traditional tree under whose shade Sakyamuni attained
" complete enlightenment," or, in other words, reached Buddha-
hood ; and no spot consequently could be considered more
sacred in the eyes of a Buddhist, or was more likely to be
reverenced from the time forward.

The Bharaut rail, according to an inscription on it, was
erected by a Prince "Vatsi-putra Dhanabhuti, son of Gaupti-
putra Angaradyut, and grandson of Gargiputra VLsvadeva "
— " in the time of the .Sungas." 2 This helps us only to a small
extent, indicating, however, that this rail is of somewhat
later date than the time of Aroka. Some fragments of
another rail of early date were also found at Patna — the
ancient Pataliputra — in excavations conducted there by Dr.
L. A. Waddell, in 1895, which are much of the same type
as those of Bodh - Gaya.3 As already mentioned, we have
no complete set of photographs of the Bodh - Gaya rail.
It is true the drawings by Major Kittoe, in the India
Office Library, are very much better than those published
by General Cunningham in his report;4 but they do not

1 Griinwedel, ' Buddhist Art,' Eng. tr.
p. 7° j Julien, ' Memoires,' tome i. pp.
464-465; Deal, ' Records,' vol. ii. p. 118.

2 ' Indian Antiquary,' vol. xxi. p. 227.

3 Waddell's ' Report on the Excava-
tions at Pataliputra ' (Calcutta, 1903),
plates i and 3.

4 ' Archaeological Reports,' vol. i.
plates 8 to ii. In Rajendralal Mitra's
' Buddha-Gaya,' plates 34-38 and 44, 45,

a number of medallions from the rail-
pillars are given ; but the drawings in
the volume are so often inaccurate in
details that we cannot altogether trust
them. In Cunningham's ' Mahabodhi,'
plates 8 and 9, are given small photographs
of a few discs. A complete survey should
give the whole pillar in each case ; of eight
of them there are photographs in the
Calcutta Museum and India Office.

IO4

BUDDHIST ARCHITECTURE.

BOOK I.

suffice for this purpose. In so far, however, as the evidence
at present available enables us to judge, it seems nearly certain
that the Bharaut sculptures are somewhere between those of
the gateways at Sanchi and those at Bodh-Gaya ; and conse-
quently we may, for the present at least, assume the latter
rail to be B.C. 250, that at Bharaut B.C. 200, and the gateways
at Sanchi to range from about B.C. 160 to say B.C. 100.

The Bodh-Gaya rail is a rectangle, measuring 145 ft. by
108 ft, and is very much ruined. Its dimensions were, indeed,
only obtained by excavation.1 The pillars are apparently only
6 ft. 8 in. in height, standing on a plinth, and are generally
ornamented with a semi-disc top and bottom containing a single
figure, or a group of several. They have also a central circular
disc, with either an animal or bust in the centre of a lotus.
Portions of the coping of the rail have been recovered, the
inner faces of which are ornamented with long lines of animals
— elephants, deer, bulls, winged horses, makaras, centaurs, etc.;
and the outer faces are carved with bands of flowers. The
intermediate rails between the pillars are sculptured with
circular lotus flowers on both sides, some of them containing
busts or animals. As the most ancient sculptured monument
in India, it would be extremely interesting to have this rail
fully illustrated,2 not so much for its artistic merit as because
it is the earliest authentic monument representing manners
and mythology in India. Its religion, as might be expected,
is principally Tree and Serpent worship, mingled with venera-
tion for dagabas, wheels, and Buddhist emblems. The domestic
scenes represent love-making, and drinking — anything, in fact,
but Buddha or Buddhism, as we afterwards come to understand
the term.

BHARAUT OR BHARHUT.

Whatever interest may attach to the rail at Bodh-Gaya
it is surpassed ten times over by that of the rail at Bharaut,
which, taking it all in all, is perhaps the most interesting
monument — certainly in a historical point of view — known to
exist in India. The tope itself, which seems to have been

1 Gen. Cunningham assumed that
the original rail formed a rectangle,
about 74 ft. from north to south by 54
ft. from east to west ; but at a later
date it was reconstructed as an enclosure
for an enlarged temple, measuring 145 ft.
from east to west, and 108 from north to
south ( ' Archaeological Reports, ' vol. iii.
p. 90, and plate 25). In ' Archaeological

Reports,' vol i. plate 4, he gave the
dimensions of the enclosure as only 131
ft. by 98 ; and Rajendralal Mitra
(' Buddha-Gaya,' p. 73), contends that it
measured 154 ft. by 114 ft. 8 in.

2 Major Kittoe made careful drawings
of most of the medallions to be seen
seventy years ago. Two of them are
reproduced here, the first representing a

CHAP. IV.

RAIL AT BHARAUT.

105

about 68 ft. in diameter, had entirely disappeared, having been
utilised by the natives to build their villages ; but about one-
half of the rail, which was partly thrown down and buried in
the rubbish, had been preserved. Originally it was 88 ft. in
diameter, and consequently some 277 ft. in length. It was
divided into four quadrants by the four entrances, each of
which was guarded by statues 4^ ft. high, of Yakshas and
Yakshinis, and Nagarajas carved in relief on the corner pillars.
The eastern gateway only is known to have been adorned with
a Toran — or, as the Chinese would call it, a " P'ai-lu " — like
those at Sanchi. One pillar of it is shown in the woodcut
(No. 32) on next page, and sufficient fragments were found in
the excavations to enable General Cunningham to restore it with
considerable certainty. From his restoration it appears to have
been 22 ft. 6 in. in height from the ground to the top of the
chakra, or wheel, which was the central emblem on the top
of all, supported by a honeysuckle ornament of great beauty.
The beams had no human figures on them, like those at Sanchi.
The lower had a procession of elephants, bringing offerings
to a tree ; the middle beam, of lions similarly employed ; the
upper beam has not been recovered, but the beam-ends are
ornamented with conventional crocodiles or makaras, and show

man on his knees before an altar worshipping a tree, while a flying figure brings
a garland to adorn it. The other represents a relic casket, over which a seven-

30. Tree Worship ; Bodh-Gaya Rail.

31. Relic Casket : Bodh-Gaya Rail.

headed Naga spreads his hood, and over him an umbrella of state. There are,
besides, two trees in a sacred enclosure, and another casket with three umbrellas
(Woodcuts No. 30, 31).

io6

BUDDHIST ARCHITECTURE.

BOOK I.

elevations of buildings so correctly drawn as to enable us to
recognise all their features in the rock-cut edifices now existing.
The toran, most like this one, is that which surmounted
the southern entrance at Sanchi, which I believe to be the
oldest of the four found there, and to have been erected in the
middle of the 2nd century, before our era (B.C. 160-150). This

one, however, is so much more wooden and constructively so
inferior, that I would, on architectural grounds alone, be inclined
to affirm that it was the older. The age of the rail, however,
does not depend on this determination, as the toran may have
been added afterwards.

The rail was apparently 9 ft. in height, including the coping,

CHAP. IV.

RAIL AT BHARAUT.

107

and had three discs on intermediate rails. The inner side of
the upper rail was ornamented by a continuous series of bas-
reliefs, divided from each other by a beautiful flowing scroll.
The inside also of the discs was similarly ornamented — their
sculptures bearing an evident analogy to those on the Bodh-
Gaya rail, whilst some of the pillars had bas-reliefs in three
storeys on three of their sides. Altogether, I fancy not less
than one hundred separate bas-reliefs were recovered, all
representing some scene or legend of the time, and nearly all
inscribed l not only with the names of the principal persons
represented, but with the title of the jataka or legend, so that
they are easily recognised in the books now current in Buddhist
countries. It is the only monument in India that is so
inscribed,2 and it is this that consequently gives it such value
for the history not only of art but of Buddhist mythology.

If this work professed to be a history of Indian art, including
sculpture, it would be necessary to illustrate this rail to a much
greater extent than is here attempted ; but as architecturally it is
hardly more important than others, the reader who is interested
in it may be referred to the volume published by its discoverer.3
Meanwhile, however, it cannot be too strongly insisted upon
that the art here displayed is purely indigenous. There is
absolutely no trace of Egyptian influence. It is, indeed, in
every detail antagonistic to that art ; nor is there any trace of
classical art. The capitals of the pillars do resemble some-
what those at Persepolis, and the honeysuckle ornaments point
in the same direction ; 4 but, barring that, the art, especially the
figure-sculpture belonging to the rail, seems an art elaborated
on the spot by Indians, and by Indians only.5

Assuming these facts to be as stated, they give rise to one
or two inferences which have an important bearing on our
investigations. First, the architecture of this rail, with its toran,
are more essentially wooden than even those at Sanchi, and, so

1 ¥01 the translation of these inscrip-
tions by Dr. Hultzsch, see ' Indian
Antiquary,' vol. xxi. pp. 225-242. See
also vol. x. pp. iiSff, 2556°. ; vol. xi.
pp. 258". ; and vol. xiv. pp. I3?f.

2 The sculptures on the walls of the
old Papnath temple at Pattadakal are also
labelled ; but it is an almost exceptional
instance.

3 General Cunningham's 'Stiipa of
Bharhut,' London, 1879.

4 Whatever the early Hindus borrowed
from Persia or elsewhere they gradually
modified by varying the details, until it
became native in form.

5 The following outline (Woodcut No.
33, on the next page) of one of the bas-
reliefs on a pillar at Bharaut may serve to
convey an idea of the style of art and of
the quaint way in which the stories are
there told. On the left, a king with
a five-headed snake-hood is represented
kneeling before an altar strewn with
flowers, behind which is a tree (Sirisa
dcasia?) hung with garlands. Behind
him is an inscription to this effect : —
"Erapato (Air^vata) the Nagaraja wor-
ships the blessed one (Bhagavat)." Above
him is the great five-headed Naga him-
self, rising from a lake. To its right a

io8

BUDDHIST ARCHITECTURE.

BOOK I.

far as it goes, tends to confirm the conclusion that, at the period
they were erected, the style was passing from wood to stone.

man in the robes of a priest standing up to his middle in the water, and above
the Naga a female genius, apparently floating in the air. Below is another
Nagaraja, with his quintuple snake-hood, and behind him two females with a single

33. Tree and Serpent Worship at Bharaut, (From a Photograph.)

snake at the back of their heads — an arrangement which is universal in all
Naga sculpture. They are standing up to their waists in water. If we may
depend on the inscription below him, this is Erapato twice over, and the
females are his two wives. The inscription up the side states that this is " the gift of
Rishidatta, a preacher." — Cunningham's ' Stiipa of Bharhut,' plate 14.

This bas-relief is further interesting as being an epitome of my work on ' Tree and
Serpent Worship.' As expressing in the shortest possible compass nearly all that is
said there at length, it will also serve to explain much that is advanced in the
following pages. As it is years older than anything that was known when that
book was written, it is a confirmation of its theories, as satisfactory as it is complete.

CHAP. IV. RAIL AT BHARAUT.

109

On the other hand, however, the sculpture is so sharp and clean,
and every detail so well and so cleverly expressed in the hard
sandstone in which it is cut, that it is equally evident the carvers
were perfectly familiar with the material they were using. It
is far from being a first attempt. They must have had chisels
and tools quite equal to carving the hardest stone, and must
have been perfectly familiar with their use. How long it may
have taken them to acquire this degree of perfection in stone
carving, it is of course impossible to guess, without further data ;
but it must have been centuries. Though, therefore, we may
despair of finding any architectural buildings older than the
time of A.roka, it is by no means improbable that we may find
images or bas-reliefs, and inscriptions of a much earlier date,
and for the history of India and her arts they would be as useful
as the larger examples.

For the present we must be content with the knowledge
that we now know perfectly what the state of the arts was
in India when the Greeks first visited it. Neither the Bodh-
Gaya nor the Bharaut rails were, it is true, in existence in
Alexander's time ; but both were erected within the limits of
the century in which Megasthenes visited the country, as
ambassador from Seleukos, and it is principally from him that
we know what India was at that time. If he did not see these
monuments he may have seen others like them, and at all
events saw carvings executed in the same style, and wooden
chaityas and temples similar to those depicted in these
sculptures. But one of the curious points they bring out is,
that the religious observances he witnessed at the courts of the
Brahmanical king, Chandragupta, are not those he would have
witnessed had he been deputed to his Buddhist grandson the
great A^oka. Here, as everything else at this age, everything
is Buddhist, but it is Buddhism without Buddha. He nowhere
appears, either as a heavenly person to be worshipped, or even
as an ascetic. The nearest indication of his presence is in a
scene where Ajata^atru — the king in whose reign he attained
Nirvana — kneels before an altar in front of which are impressions
of his feet. His feet, too, seem impressed on the step of the
triple ladder, by which he descended from Heaven at Sankisa ;
Maya's dream, and the descent of the white elephant, can be
recognised, and other indications sufficient to convince an
expert that Buddhism is the religion indicated. But, as at
Sanchi, by far the most numerous objects to which worship is
addressed in these sculptures, are trees, one of which, the
inscription tells us, is the Bodhi-tree of Sakyamuni. Besides
this, the Bo-trees of six or seven of his predecessors are
represented in these sculptures, and both by their foliage and

no

BUDDHIST ARCHITECTURE.

BOOK I,

their inscriptions we can easily recognise them as those known
at the present day as ascribed to these previous Buddhas.1

Naga people, and kings with their five-headed serpent-hoods,
are common ; but only one instance has yet been brought to
light in which the serpent can be said to be worshipped.
Making love and drinking are not represented here as at
Sanchi — nor are females represented nude as they are on the
Jaina sculptures at Mathura. All are decently clothed, from the
waist downwards at least, and altogether the manners and
customs at Bharaut are as much purer as the art is better than it
is in the example of Sanchi.

MATHURA (MUTTRA).

When excavating at Mathura, General Cunningham found
several pillars of a rail, which, judging from the style, is pro-
bably later than that at Bharaut, but still certainly anterior to
the Christian Era. The pillars, however, are only 4^ ft. high,
and few traces of the top rail or of the intermediate discs have
been found. Each pillar is adorned by a figure of a nude
female in high relief, singularly well executed, richly adorned
with necklaces and bangles, and a bead belt or truss around
their middles. Each stands on a crouching dwarf or demon,
and above each, in a separate compartment, are the busts of two
figures, a male and female, on a somewhat smaller scale, either
making love to each other, or drinking something stronger than
water.2

Though the sculptures at Sanchi and Katak have made us
familiar with some strange scenes, of what might be supposed
an anti-Buddhistic tendency, this rail, we cannot now doubt,
belonged to a group of Jaina temples and monastic buildings of a
very early age. We do not, indeed, know if the rail was straight
or circular, or to what class of building it was attached ; but it
is pretty certain that these pillars belonged to one or more Jaina
stupas such as that of which the remains were excavated by
Dr. Fiihrer in 1888-89. Jaina tradition had always claimed
Mathura as one of the centres of their sect, and an inscription
found there and dated in the year 79 of the Kushan kings, records
the consecration of a statue to the stupa of Suparyva.8 This
confirms what had previously been anticipated — that the Jains,
as well as the Buddhists, erected stupas in honour of their

1 Tumour's 'Mahawansa,' Introd. p.
32 ; Griinwedel, ' Buddhist Art in India '
(Eng. tr.), p. 74.

2 Outlines of these sculptures are given
in Gen. Cunningham's ' Archaeological

Reports,' vol. iii. plate 6 ; also in
Fiihrer's plates 60-64, in V. A. Smith's
'Jaina Stupa, etc., at Mathura.'

3 'Epigraphia Indica,'vol. ii. pp. 204,
32lf.

CHAP. IV.

RAIL AT SANCHI.

prophets, and that their art did not differ materially from that
of the latter.1

£

tftlfVta.tlfl.fJ:

#fe

' '•'_•

J

J

J

SANCHI-KANAKHEDA.

Though the rails surrounding the topes at Sanchi are not,
in themselves, so interesting as those at Bodh-Gaya and
Bharaut, still they are useful in exhibiting the various steps
by which the modes of decorating rails were arrived at, and
the torans or gateways of the great rail are quite unequalled
by any other examples known
to exist in India. The rail
that surrounds the great tope
may be described as a circular
enclosure 140 ft. in diameter,
but not quite regular, being
oval on one side, to admit of
the ramp or stairs leading to
the berm or procession - path
surrounding the monument.
As will be seen from the
annexed woodcut (No. 34), it
consists of octagonal pillars 8 34
ft. in height, and spaced 2 ft.
apart. These are joined to-
gether at the top by a rail 2 ft. 3 in. deep, held in its position by a
tenon cut on the top of the pillars, as at Stonehenge ; between
the pillars are three intermediate rails, which are slipped into
lens - shaped holes, on either side, the whole showing how
essentially wooden the construction is. The pillars, for instance,
could not have been put up first, and the rails added afterwards.
They must have been inserted into the right or left hand posts,
and supported while the next pillar was pushed laterally, so
as to take their ends, and when the top rail was shut down
the whole became mortised together as a piece of carpentry,
but not as any stone - work was done either before or
afterwards.

The rail of the No. 2 Stupa at Sanchi is of special interest as
being more ornamented with sculptures which, with many of the
inscriptions, appear to belong to a period distinctly antecedent
to those of the gateways of the great stupa (Woodcut No. 35);
there circular discs are added in the centre of each pillar,

Rail at Sanchi. (From a Drawing by
Gen. Cunningham.)

1 Biihler, ' Legend of the Jaina Stupa
at Mathura ' ; and ' Epigraphia Indica,'
vol. ii. pp. 311 et seqq. and plates; see
also the curious tale about Kanishka and

a Jaina stupa that he worshipped by
mistake. — 'Journal Asiatique,' JXe ser.
tome viii. ( 1896), pp. 458ff. Conf. Arite,
p. 54, note I.

112

BUDDHIST ARCHITECTURE.

BOOK I.

and semicircular plates at top and bottom.1 In carpentry the
circular ones would represent a great nail meant to keep the
centre bar in its place ; the half discs top and bottom, metal

plates to strengthen
the junctions — and
this it seems most
probably may really
have been the origin
of these forms.

If from this we
attempt Ifo follow the
progress made in the
ornamentation of
these rails, it seems to
have been arrived at
by placing a circular
disc in each of the
intermediate rails, as
shown in the woodcut
(No. 36), copied from
a representation of
the outer face of the
Amaravati rail, carved
upon it. In the actual
rail the pillars are
proportionally taller and the spaces somewhat wider, but in all
other respects it is the same — it has the same zoophorus below,
and the same conventional figures bearing a roll above, both of
which features are met with almost everywhere.

35-

Rail, No. 2 Tope, Sanchi.
(From a Drawing by Col. Maisey.)

36. Representation of Rail, (From a Bas-relief at Amara vat!,)

A fourth stage was reached in that shown in the next
woodcut (No. 37), from a representation of a rail in the

1 The sculptures on this rail have not | drawn by Col. Maisey. See his ' Sanchi,'
been fully illustrated ; some of them were | plates 29-31, and pp. 67-70.

CHAP. IV.

RAIL AT SANCHI.

Gautamiputra cave at Nasik, dr. A.D. 160 to 175, where there
are three full discs on the pillars as well as on the rails, and
no doubt other variations may yet be found ; but these are
sufficient to show how the discs were multiplied till the pillars
almost become evanescent quantities in the composition.

Rail in Gautamiputra Cave, Nasik,

The greatest innovation, however, that took place, was the
substitution of figure - sculpture for the lotus or water leaves
of the discs, if that can be called an innovation, which certainly
took place in the wooden age of architecture, before it was
thought of translating these things into stone. The earliest
rails we know, those at Bodh-Gaya and Bharaut, show these
changes already completed in the manner above described.
The plainness of the rail, or the absence of figure-sculpture, is
consequently no test of its greater or less antiquity, though
the extreme multiplication of discs, as shown in the last
example, seems only to have taken place just before their
discontinuance.

To return, however, from this digression. The rail that
surrounds the great tope at Sanchi was probably commenced
immediately after its erection, which, as explained above, was
probably in A^oka's time, B.C. 250; but as each rail, as shown
by the inscription on it, was the gift of a different individual,1

1 Gen. Cunningham collected and
translated 196 inscriptions from this
tope, in his work on the Bhilsa Topes,
pp. 235 ef seqq., plates 16-19 ; but the
more accurate versions of a larger collec-

VOL. I.

tion of 378 from this rail and 78 from
that of Stupa No. 2, are those by
Professor Buhler, published in ' Epi-
graphia Indica,' vol. ii. pp. 87 et seqq.,
and 366 et seqq.

H

BUDDHIST ARCHITECTURE.

BOOK I.

it may have taken a hundred years to erect. The age of the
torans is more easily ascertained. There is an inscription on
the south gateway, which is certainly integral, which states that
the gateway was erected during the reign of King Satakarni
of the Andhra dynasty, and it is nearly certain that this applies
to a king of that name who reigned about B.C. 155. If we
assume that this gateway — which leads to the steps by which
the berm is ascended — is probably the oldest of the four, it gives
us a starting-point from which to determine the age of the
others. The next to be erected would be the northern. That
may have been followed by the eastern — the one of which there
is a cast at South Kensington — and the last erected was perhaps
the western.1 The style and details of all these show a succes-
sion and a progress that could hardly have taken place in much
less than a century, and, with other reasons, enable us to assert
without much hesitation, that the four gateways were added
to the rail of the great tope during the 2nd century before the
Christian Era.2 The northern gateway is shown in the general
view of the building (Woodcut No. 12), but more in detail in
the cut (No. 38) on the following page.

In design and dimensions these four gateways are all
very similar to one another. The northern is the finest,3 as
well as somewhat larger than the others. Its pillars, to the
underside of the lower beam, measure 18 ft, including the
elephant capitals, and the total height to the top of the
emblem is 35 ft. The extreme width across the lower beam
is 20 ft. The other gateways are somewhat less in dimen-
sions, the eastern being only 33 ft. in height. The other
two having fallen, and — though re-erected by Government —
we cannot be sure what their exact dimensions may originally
have been.

All these four gateways, or torans as they are properly
called, were covered with the most elaborate sculptures both
in front and rear — wherever, in fact, their surface was not
hidden by being attached to the rail behind them. Generally
the sculptures represent scenes from the life or legend of

1 In later stupas, the west side seems
to have been frequently regarded as the
front ; but generally it was on the side
facing the monastic buildings.

2 The details from which these de-
terminations are arrived at will be found
in ' Tree and Serpent Worship,' pp. 98
et seqq. Since that work was published,
however, the discovery of .Satakarni's
name in the Hathigumpha inscription, at
Khandagiri, the re - adjustment of the

chronology of the Andhras, and other
epigraphical results, have considerably
altered the actual dates ascribed to
monuments of this early period ; their
relative ages, however, are not materially
affected by this.

* It is very much to be regretted that
when Lieut. Cole had the opportunity he
did not take a cast of this one instead of
the eastern gateway. It is far more com-
plete, and its sculptures more interesting.

•

.- £*s

Northern Gateway of Tope at Sanchi. (From a Photograph.

n6

BUDDHIST ARCHITECTURE.

BOOK I.

Buddha,1 but nowhere is he represented in the conventional
forms either standing or seated cross-legged, which afterwards
became universal. In addition to these are scenes from the
jatakas or legends, narrating events or actions that took place
during the five hundred births through which Sakyamuni had
passed before he became so purified as to reach perfect Buddha-
hood. One of these, the Wessantara, or "alms-giving Jataka,"
occupies the whole of the lower beam of the northern gateway,
and reproduces all the events of that wonderful tale, exactly as
it is narrated in Sinhalese and Pali books at the present day.
Besides these legendary scenes, the worship of trees is repre-
sented at least seventy-six times ; of dagabas or relic shrines,
thirty-eight times ; of the chakra, or wheel, the emblem of
Dharma — the law — ten times ; and of Sri, the goddess of
fortune, who afterwards, in the Hindu Pantheon, became
Lakshmi the consort of Vishnu, ten times. The Triratna or
trident emblem which crowns the gateways may, and I am
inclined to believe does, represent the Buddhist creed. On
the left-hand pillar of the north gateway it crowns a pillar,
hung with wreaths and emblems, at the bottom of which
are the sacred feet (Woodcut No. 39) : — the whole looking like
a mystic emblem of a divinity, it was forbidden to represent
under a human form. The corresponding face of the opposite
pillar is adorned with architectural scrolls, wholly without any
esoteric meaning so far as can be detected, but of great beauty
of design (Woodcut No. 40).

Other sculptures represent sieges and fighting, and con-
sequent triumphs. Others portray men and women eating and
drinking and making love, and otherwise occupied, in a manner
as unlike anything we have been accustomed to connect with
Buddhism as can well be imagined. Be this as it may, the
sculptures of these gateways form a perfect picture Bible of
Buddhism as it existed in India in the 2nd century before the
Christian Era, and as such are as important historically as they
are interesting artistically.2

The small tope (No. 3) on the same platform, and about
40 yards north-north-east from the great tope at Sanchi, was
surrounded by a rail which has now almost entirely disappeared.
It had, however, at least one toran, the pillars and one beam of

1 See Griinwedel's ' Buddhist Art in
India,' Eng. transl., pp. 58-74.

2 For details of these sculptures and
references, the reader must be referred
to ' Tree and Serpent Worship,' where
a great number of them are represented
and described in great detail. Sculptures
do not, strictly speaking, belong to this

work, and, except for historical purposes,
are not generally alluded to.

The sculptures were all photographed
to scale some years ago by Mr. H. Cousens
of the Archaeological Survey, but as yet
there is no prospect of their publication.
— 'Journal of the Royal Asiatic Society,'
1902, pp. 29-45.

CHAP. IV.

RAIL AT SANCHI.

117

which were still standing twenty years ago, when the upper two
beams were also pieced
together and replaced.
It is only about half the
size of those of the great
tope, measuring about
17 ft. to the top of the
upper beam, and 13 ft.
across its lower beam. It
is apparently somewhat
more modern than the
great gateways, and its
sculptures may have re-
ference to the acts of
vSariputra and Mogga-
lana, whose relics, as
above mentioned, were
deposited in it.

This tope was only
40 ft. in diameter, which
is about the same dimen-
sion as No. 2 Tope, con-
taining the relics of
some of the apostles
who took part in the
third convocation under
Asoka., and afterwards
in the diffusion of the
Buddhist religion in the
countries bordering on
India.

As above pointed
out, the rails at Bodh-
Gaya and Bharaut afford
a similar picture of
Buddhism at a time
perhaps a century earlier.
The difference is not
striking, but on a close
examination it is evident
that the art, if not also
the morals, had degener-
ated during the interval.
There is a precision and
a sharpness about the
Bharaut sculptures which is not found here, and drinking and

\

Bas-relief on left-hand

Pillar,
Northern Gateway,

Ornament on right-hand

Pillar,
Northern Gateway.

n8

BUDDHIST ARCHITECTURE.

BOOK I.

love-making do not occur in what remains of the sculptures — they
do, however, occur at Bodh-Gaya — to anything like the extent
they do at Sanchi. There is no instance at Bharaut of any figure
entirely nude ; at Sanchi apparent nudity among the females
is rather the rule than the exception. The objects of worship
are nearly the same in both instances, but are better expressed
in the earlier than in the later examples. We may, however,
make some allowance for differences of material and race or
locality — possibly also to peculiarities in the different sects
for which the two monuments were executed.1 The Mathura
sculptures may suggest that the Digambara Jains regarded
female as well as male nudity as a mark of sanctity.

Before leaving these torans, it may be well to draw attention
again to the fact of their being, even more evidently than the rails,
so little removed from the wooden originals out of which they
were elaborated. No one can look at them, however carelessly,
without perceiving that their forms are such as a carpenter
would imagine, and could construct, but which could not be
invented by any process of stone or brick masonry with which
we are familiar. The real wonder is that, when the new fashion
was introduced of repeating in stone what had previously been
executed only in wood, any one had the hardihood to attempt
such an erection in stone ; and still more wonderful is it that,
having been done, three of them should have stood during
eighteen centuries, till one was knocked down by some clumsy
Englishmen, and that only one — probably the earliest, and
consequently the slightest and most wooden — should have
fallen from natural causes.

Although these Sanchi torans are not the earliest specimens
of their class executed wholly in stone, neither are they the
last. We have, it is true, no means of knowing whether those
represented at Amaravati2 were in stone or in wood, but, from
their different appearances, some of them most probably were
in the more permanent material. At all events, in China and
Japan their descendants are counted by thousands. The "p'ai-lus"
in the former country, and the " tori-is " in the latter, are copies
more or less correct of these Sanchi gateways, and like their

1 The difference of style may be com-
pared with that which prevails among
Musalman monuments in different parts
of India in the I5th and i6th centuries.

2 They must certainly have been very
common in India, for, though only one
representation of them has been detected
among the sculptures at Sanchi ('Tree
and Serpent Worship,' plate 27, fig. 2),
at least ten representations of them are

found at Amaravati, plates 59 (fig. 2),
60 (fig. I), 63 (fig. 3), 64 (fig. i) 69, 83
(fig. 2), 85 (figs, i and 2), 96 (fig. 3),
98 (fig. 2), and no doubt many more may
yet be found. In Cave 10 at Ajanta,
containing the oldest paintings there,
two were represented on the left wall.
' Notes on the Bauddha Rock-Temples
of Ajanta, 'p. 51, and plate II.

CHAP. IV.

RAIL AT AMARAVATI.

119

Indian prototypes are sometimes in stone, sometimes in wood
and frequently compounded of both materials, in varying pro-
portions. What is still more curious, a toran with five bars
was erected in front of the Temple at Jerusalem, to bear the
sacred golden vine. It, however, was partly in wood, partly in
stone, and was erected to replace one that adorned Solomon's
Temple, which was wholly in bronze, and supported by the
celebrated pillars Jachin and Boaz.1

AMARAVATI.

Although the rail at Bharaut is the most interesting and
important in India in an historical sense, it is far from being
equal to that at Amaravati, either in elaboration or in artistic
merit. Indeed, in these respects, the Amaravati rail is probably
the most remarkable monument in India. In the first place, it
is more than twice the dimensions of the rail at Bharaut, the
great rail being 192 ft. in diameter, the base i6a| ft.,2 or almost
exactly twice the dimensions of that at Bharaut ; between these
two was the procession-path — 13 ft. wide. The inner wall of
sculptures was the facing of the base that supported the plat-
form on which the dome stood. Externally, the total height
of the great rail was about 14 ft. ; internally, it was 2 ft. less,
while the sculptured facing of the base was, perhaps, only 6 ft.
in height, with a frieze and cornice over it.

The external appearance of the great rail may be judged
of from the annexed woodcut (No. 41), representing a small
section of it. The lower part, or plinth, was ornamented by
a frieze of animals and boys or dwarfs, generally in ludicrous
and comic attitudes. The pillars, as usual, were rectangular
with the corners splayed off, ornamented with full discs in the
centre, and half discs top and bottom, between which were
figure sculptures of more or less importance. On the three
rail-bars were full discs, all most elaborately carved, and all
different. Above runs the usual undulating roll moulding,
which was universal in all ages,3 but is here richly inter-
spersed with figures and emblems. The inside of the rail
was very much more richly ornamented than the outside

' Tree and Serpent Worship,' Appen-
dix I. p. 270; 'Temples of the Jews,'
pp. 152, 155.

1 From some misunderstanding of the
first accounts, it was supposed the Amara-
vati stupa had an inner rail ; this was a
mistake — the inner circle of sculptures
was the facing of the base of the stupa.

3 In Burma at the present day a roll
precisely similar to this, formed of coloured
muslin, distended by light bamboo hoops,
is borne on men's shoulders in the same
manner as shown here, on each side of
the procession that accompanies a high
priest or other ecclesiastical dignitary to
the grave.

120

BUDDHIST ARCHITECTURE.

BOOK I.

shown in the woodcut. All the central range of discs, both
on the pillars and on the rails, were carved with figured

External Elevation of Great Rail at Amaravati.

subjects, generally of very great elaboration and beauty of
detail ; and the coping was one continuous bas-relief upwards
of 600 ft. in length. At the returns of the gateways another
system was adopted, as shown in next woodcut (No. 42). The
pillars being narrower, and the discs smaller, the principal
sculpture was on the intermediate space : in this instance a
king on his throne receives a messenger, while his army in
front defends the walls ; lower down the infantry, cavalry, and
elephants sally forth in battle array, while one of the enemy
sues for peace, which is probably the information being com-
municated to the king.

The sculptured base, though, perhaps, lower than the rail,
was even more richly ornamented than it, generally with
figures of dagabas — apparently twelve in each quadrant —

CHAP. IV.

RAIL AT AMARAVATI.

121

most elaborately carved with scenes from the life of Buddha

or from the legends. One of
these dagabas has already been
given (Woodcut No. 20, page
Si).1 Between these were pillars
and slabs ornamented, either as
shown in Woodcuts Nos. 43
and 44, or with other Buddhist

%

42. Angle pillar at Amaravat!. 43. Slab from Base of the Stupa, Amaravati.

designs or emblems, but all as rich, at least, as these ; whilst

1 For other examples, see 'Amaravati and Jaggayyapeta Stiipas,' plates i., xxxi.
figs- 6, 7, and xxxiii, xxxix.

122

BUDDHIST ARCHITECTURE.

BOOK I.

over these slabs was an architrave, carved like ivory, with
scenes from the life of Buddha ; the whole making up a series
of pictures of Buddhism, as it was understood in the ist and

2nd centuries, unsurpassed by
anything now known to exist
in India. The slab represented
in Woodcut No. 43 (p. 121),
though now much ruined, is
interesting as showing the three
great objects of Buddhist
worship at once. At the top
is the dagaba with its rail,
but with the five-headed Naga
in the place usually occupied
by Buddha. In the central
compartment is the chakra or
wheel, now generally acknow-
ledged to be the emblem of
Dharma, the second member
of the Buddhist Triad ; below
that the tree, possibly repre-
senting the Sangha or the con-
gregation ; and in front of all
a throne, on which is placed
what I believe to be a relic,
wrapt up in a silken cloth.

This combination is repeated
again and again in the earlier
of these sculptures, and may be almost designated as the
shorter Buddhist catechism, or rather the creed — Buddha,
Dharma, Sangha. The last woodcut (No. 44) is also interest-
ing, as showing, besides the three emblems, the form of
pillars with double animal capitals so common in structures
of this and an earlier period.

The age of this monument can hardly be fixed with
certainty ; the sculptures on the rail and on some of the slabs
that probably belonged to the stupa itself, are sufficiently
analogous in general style to those left of the Bharaut stupa
to suggest that they may be of scarcely more than a century
later. It must have been commenced at least before figures
of Buddha were represented in sculpture — the relic casket and
pair of footmarks being the symbols employed to represent
him. But among the figures that appear to have belonged to
the base of the stupa, there are many that can hardly be earlier
than the beginning of the 2nd century A.D. From an inscrip-
tion of the reign of Pulumavi Vasishthiputra, it would seem

D^gaba (from a Slab), Amaravati.

CHAP. IV.

RAIL AT AMARAVATI.

123

that the stupa — called a "great chaitya" — belonged to the
Chaitika or Purva^aila school, and must have undergone a
great restoration about A.D. 1 50 ; l but we find pieces of very
ancient sculpture, some of which have been reworked on the
back to fit into new places ; and at least one inscription was
found that is engraved in pure Maurya characters of about
B.C. 200. It seems probable then that an early stupa existed
here, of which only a few archaic sculptures have survived.
This was restored — perhaps reconstructed and much enlarged
— a century or more B.C., the sculptures of the rail represent-
ing the veneration of relics, the Triratna, domestic and other
scenes, etc., but without the figure of Buddha. Then, in the
2nd century A.D., the stupa itself seems to have been restored
when the sculptures wainscotting its base were added, picturing
scenes from the legend of Buddha.2

When Hiuen Tsiang visited this place in the year 639 it had
already been deserted for more than a century, but he speaks
of its magnificence and the beauty of its site in more glowing
terms than he applies to almost any other monument in India.
Among other expressions, he uses one not easily understood
at first sight, for he says, " it was ornamented with all the
magnificence of the palaces of Baktria " 3 (Tahia). Now, how-
ever, that we know what the native art of India was from
the sculptures at Bharaut and Sanchi, and as we also know
nearly what the art of Baktria was from those dug up
near Peshawar, especially at Jamalgarhi, we see at once that
it was by a marriage of these two arts that the Amaravati
school of sculpture was produced, but with a stronger classical
influence than anything of its kind found elsewhere in India.4

With this, which is certainly the most splendid specimen of
its class, we must conclude our history of Buddhist rails. No
later example is known to exist ; and the Gandhara topes,
which generally seem to be of this age or later, have all
their rails attached to their sides in the shape of a row
of pilasters. If they had any figured illustrations, they were
either in the form of paintings on plaster on the panels or

1 There is no record of the positions of
the sculptures belonging to the basement ;
and the Government excavation of the
whole area in 1881 destroyed the last
chance of any further determination.

2 For a full account of the Amaravati
stupa, see thevolume of the Archaeological
Survey, ' The Buddhist Stupas of Amara-
vati and Jaggayyapeta,' 1887.

3 ' Histoire de Hiouen Thsang traduite
par Julien,' vol. i. p. 188 ; Beal, ' Life of

Hiuen Tsiang,' pp. i^6f.

4 In some of the technical details of the
sculptures, notably in the treatment of
drapery, the influence of classic art is
perceptible ; and it is perhaps here alone
in India Proper that this foreign impress is
seen. The age of the later sculptures
nearly coinciding with the same influence
in Gandhara may account for this. Conf.
Dr. Le Bon, ' Les Monuments de 1'Inde,'
p. 14.

124

BUDDHIST ARCHITECTURE.

BOOK I.

rilievos between the pilasters ; for we cannot understand any
Buddhist monument existing anywhere, without the jatakas
or legends being portrayed on its walls in some shape or other.

At Sarnath all reminiscences of a rail had disappeared,
and a new mode of ornamentation introduced, which bore no
resemblance to anything found on the earlier topes.

Although, therefore, our history of the rails may finish about
A.D. 200, it by no means follows that examples may not yet
be brought to light belonging to the five centuries that elapsed
between that date and the age of A^oka. As they all certainly
were sculptured to a greater or less extent, when they are
examined and published, we may hope to have an ancient
pictorial history of India for those ages nearly as complete
as that possessed by any other country in the world. At
present, however, we only know of ten or twelve examples,
but they are so easily thrown down and buried that we may
hope to find more whenever they are looked for, and from
them to learn the whole story of Buddhist art.

NOTE. — The central crowning orna-
ment in Woodcut No. 38, page 115, is
a chakra or wheel in the centre, with
Triratna emblems right and left. These
triratna symbols represent the three
"Jewels" of Buddhism, — Buddha,
the Dharma and Sangha. On the upper
beam five dagabas and two trees are
worshipped; on the intermediate blocks,
.Sri and a chakra ; on the middle beam
are seven sacred trees, with altars ; on

the intermediate blocks, Sri and the
chakra again. The lower beam is wholly
occupied by the early scenes in the
Wessantara jataka, which is continued in
the rear. The subjects on the pillars have
been described in ' Tree and Serpent
Worship,' but are on too small a scale
to be distinguishable in the woodcut.
See also Griinwedel's ' Buddhist Art in
India,' Eng. trans., pp 19, 74, 145. The
Triratna is also a Jaina symbol.

45. Triratna Emblem. (From a sculpture at 46. Triratna Symbol from

Amaravatt.) Sanchi.

CHAP. V.

CHAITYA HALLS

125

CHAPTER V.

CHAITYA HALLS.

CONTENTS.
Structural Chaityas— Bihar Caves — Western Chaitya Halls, etc.

ALTHOUGH, if looked at from a merely artistic point of view,
it will probably be found that the rails are the most interesting
Buddhist remains that have come down to our time, still, in
an historical or architectural sense, they are certainly surpassed
by the chaitya halls. These are the temples of the religion,
properly so called, and the exact counterpart of the churches
of the Christians, not only in form, but in use.

Some twenty or thirty of these are known still to exist
in a state of greater or less preservation, but — with very few
exceptions — all cut in the rock. In so far as the interior is
concerned this is of little or no consequence, but, were it not
for one or two recent discoveries, we should not have been
able to judge of their external form or effect,1 and, what is
worse, we should not have known how their roofs were con-
structed. We know that, generally at least, they were formed
with semicircular ribs of timber, and it is also nearly certain
that on these ribs planks in two or three thicknesses were laid,
but we could hardly have guessed what covered the planks
externally.

Till recently the only structural one known was that at
Sanchi, which is shown in plan in the accompanying woodcut
(No. 47). It did not, however, suffice to show us how the
roofs of the aisles were supported externally. What it does
show, which the caves do not, is that when the aisle which

1 It had previously been considered
probable that a tolerably correct idea of
the general exterior appearance of the
buildings from which these caves were
copied may be obtained from the Raths
(as they are called) of Mamallapuram
or Mahavellipore (described further on).
These are monuments of a later date,
and belonging to a different religion, but

they correspond so nearly in all their
parts with the temples and monasteries
now under consideration, that we could
scarcely doubt their being, in most re-
spects, close copies of them, as we now
discover that they really are. Curiously
enough, the best illustrations of some of
them are to be found among the sculptures
of the Bharaut Rail.

126

BUDDHIST ARCHITECTURE.

BOOK I.

surrounded the apse could be lighted from the exterior, the
apse was carried up solid. In all the caves the pillars sur-
rounding the dagaba are of different form and
plainer than those of the nave. They are, in fact,
kept as subdued as possible, as if it was thought
they had no business there, but were necessary
to admit light into the circumambient aisle of
the apse.

The discovery of two old structural chaityas
enables us now to realise the formation of their
roofs. One of these, at Ter in the Naldrug
district of Haidarabad,1 has been appropriated
to Vaishnava worship, and to some extent altered, [>ian=of chait a
but its arrangements will be understood from the Hail, Sanchi.
accompanying view and plan (Woodcuts Nos. 48 Scale 3° ft> to T '"•
and 49). The mandap on the east may possibly have been an

addition,butifso,ofanearly
date, as the mouldings and
pilasters testify. The doors
of this and of the shrine
have been altered : that on
the north side has been
cut out, and the partition
wall across the apse added,

48.

Ancient Buddhist Chaitya at Ter.
(From a photograph by H, Cousens,)

fully laid in a cement of clay.

49. Plan of Ancient Buddhist Chaitya
at Ter. Scale 50 ft. to i in.
(From a Plan by H. Cousens.)

whilst the dagaba — pro-
bably of marble — has been
removed to make room for
two Vaishnava images.
The shrine or chaitya is
wholly built of bricks of
large size, measuring 17 in.
by 9, and 3 in. thick, care-
This apartment measures 26 ft.

1 Ter was indentified by Dr. Fleet as
the site of the ancient Tagara ( ' Journal
of the Royal Asiatic Society,' 1901,
pp. 537ff.); and, led by this, Mr. H.
Cousens visited the place in November

1901, and discovered this interesting
monument. The plan (Illustration No.
49) and the substance of this account is
based on his paper in the ' Archaeological
Survey Annual, 1902-03,' pp. I95f.

CHAP. V.

CHAITYA HALLS.

127

in length by 12 ft. wide, the apsidal end and barrel-vaulted
roof rising about 30 ft. from the floor ; it is of brick carefully
corbelled inwards to form the vault, and rising to a ridge, all
carefully plastered. The mandap in front is about 13 ft. high
inside, and has a flat roof of wooden beams supported by wooden
pillars, and over-laid by brick and plastered outside. Slender
pilasters are formed on the outside walls, supporting thick roll
mouldings of an early form. The facade of the chaitya rises
about 1 8 ft. above the hall roof (Woodcut No. 48), and is of
special interest : the recess in the middle of it now contains
a Hindu image in plaster, but originally it must have been a
window to admit light into the chaitya. And if we compare
this facade with that of the Buddhist rock-cut chaitya at
Elura, the close resemblance in style, and even in details, derived
from earlier wood constructions, is very apparent.

The early chaitya, discovered about eighteen years ago at
Chezarla, in the Kistna district of the Madras presidency, has
been preserved, like the pre-
ceding, by being appropri-
ated as a vSaiva temple of

Kapotejvara — "the pigeon • iUbii-ii- ii- fTF-i!
god " ; for the legend of
.Sivi-Ujinara, — who offered
his own flesh to feed a hawk
rather than surrender a
dove (kapotd) that had fled
to him for protection — is well known from the Mahabharata and

Elevation of Chezarla Chaitya Temple.
Scale 25 ft. to i in.

52-

Section on A B,
of No, 50.

other works;1 the story has also Buddhist forms in the Jataka
book.2 Like the Ter example, this chaitya is built of bricks

1 Benfey's ' Panchatantra,' vol. i. p.
388 ; ' Kalhasaritsagara,' ch. 7 ; ' Maha-
bharata,' Vana-parva, sect. 197.

2 The ' Kapota Jataka,' is No. 42, and
the Sivi Jataka, No. 499 ; see also Beal's

' Buddhist Records,' vol. ii. pp. 182-183,
and 'Life of Hiuen Tsiang,' p. 125;
Burgess, ' Notes on the Bauddha Rock
Temples of Ajanta: their Paintings, etc,'
pp. 47, and 75-76.

128 BUDDHIST ARCHITECTURE. BOOK I.

of the same large size, and a mandapa 25 ft. in length by 9 ft.
9 in. wide has been added at some later date, at the east or front
end of it. The original chapel is about 23 ft. in length inside
and 8 ft. 9 in. wide in front and somewhat less at the apse. The
walls are 3 ft. 6 in. thick, but inside pilasters have been built up
to support a flat roof about 7^- ft. from the floor, hiding entirely
the inside of the vaulting ; outside it is identical with that at Ter.
The illustrations Nos. 50, 51, and 52, will enable the reader to
form a fairly distinct conception of this interesting monument.

These two examples, though of so small dimensions, fully
confirm the inductions arrived at before their discovery, of the
form of roof and arrangement of these chapels. They must at
one time have been very numerous all over India, and further
important discoveries may still reward careful research. At
Guntupalle in the Godavari district is the ruin of one, measuring
53 ft. 6 in. in length by 14 ft. 6 in. wide ; but only a few feet of
the walls remain. And at Vidyadharapuram, near Bezvvada,
the foundations have been traced by Mr. Rea of still another.

As so much of our information regarding the chaityas, as
well as the viharas, which form the next group to be described,
was first derived from the rock-cut examples in western India,
it would be convenient, if it were possible, to present something
like a statistical account of the number and distribution of the
groups of caves found there. From what we know of their
numbers and distribution we are warranted in assuming that
there are at least fifty groups of caves in India proper.

Some of these groups contain as many as 100 different and
distinct excavations, many not more than ten or a dozen ; but
altogether we may fairly assume that not less than 1 200 distinct
specimens are to be found. Of these probably 300 may be of
Brahmanical or Jaina origin ; the remaining 900 are Buddhist —
either monasteries or temples, the former being incomparably
the more numerous class ; for of the latter not more than
twenty or thirty are known to exist. This difference arose, no
doubt, from the greater number of the viharas being grouped
around structural topes, as was always the case in Afghanistan
and Ceylon ; and, consequently, they did not require any rock-
cut place of worship while possessed of the more usual and
appropriate edifice.

The facades of the caves are generally perfect, and form an
exception to what has been said of our ignorance of the external
appearance of Indian temples and monasteries, since they are
executed in the rock with all the detail that could have graced
the buildings of which they are copies. In the investigation of
these objects, the perfect immutability of a temple once hewn

CHAP. V. CHAITYA HALLS. iag

out of the living rock is a very important advantage. No repair
can add to, or indeed scarcely alter, the general features of what
is once so executed ; and there can be no doubt that we see
them now, in all essentials, exactly as originally designed. This
advantage will easily be appreciated by any one who has tried to
grope for the evidence of a date in the design, afforded by our
much-altered and often reconstructed cathedrals of the Middle
Ages.

The geographical distribution of the caves is somewhat
singular, more than nine-tenths of those now known being
found within the limits of the Bombay Presidency. The
remainder consist of two groups in Bengal — those of Bihar and
Katak, neither of which is important in extent ; those of
Dhamnar, Kholvi, Besnagar, and Bagh in Rajputana ; in Madras,
the groups at Mamallapuram, Bezwada and Guntupalle ; and
two or three small groups in the Panjab and Afghanistan.1

This remarkable local distribution may be accounted for by
the greater prevalence in western than in eastern India of
rocks perfectly adapted to such works. The great cave district
of western India is composed of horizontal strata of amygdaloid
and other cognate trap formations, generally speaking of very
considerable thickness and great uniformity of texture, and
possessing besides the advantage that their edges are generally
exposed in nearly perpendicular cliffs. No rock in any part
of the world could either be more suited for the purpose or more
favourably situated than these formations. They were easily
accessible and easily worked. In the rarest possible instances
are there any flaws or faults to disturb the uniformity of the
design ; and, when complete, they afford a perfectly dry temple
or abode, singularly uniform in temperature, and more durable
than any class of temple found in any other part of the world.

From the time of Asoka, B.C. 250, when the first cave was
excavated at Rajagriha, till about the 8th century, or later, the
series is uninterrupted ; and, if completely examined and drawn,
the caves would furnish us with a complete religious and artistic
history of the greater part of India during ten or eleven centuries,
the darkest and most perplexing of her existence. But, although
during this long period the practice was common to Buddhists,
Hindus, and Jains, it ceased before the Muhammadan conquest.
Hardly any excavations have been made or attempted since
that period, except, perhaps, some rude Jaina monoliths in the
rock at Gwaliar, and three in southern India.

1 For the Afghanistan caves, see W.
Simpson's paper in ' Transactions of the
Royal Institute of British Architects,'

VOL. I.

N.S., 1891, pp. 2S4ff, and 'Journal,'
vol. vii. p. 244.

BUDDHIST ARCHITECTURE.

BOOK I.

BIHAR CAVES.

As might be expected from what we know of the history of
the localities, the oldest caves in India are situated in Bihar, in
the neighbourhood of Rajagriha — now Rajgir— which was the
capital of Bengal at the time of the advent of Buddha. Bihar,
however, was also one of the earliest provinces in which the Jaina
doctrines were propagated, and their great Tirthankara Mahavira
was a native of Vai^ali,1 and a contemporary of Gautama Buddha.
He preached in Tirhut, Bihar, and neighbouring districts, and is
said to have died at Pawapuri, about 10 miles to the north-east
of Rajgir, where his sarnosaran or stupa stands, marking one of
the most sacred places of pilgrimage of the sect.2 They have
several temples about Rajgir, and in early times they would
have bhikshugrihas or residences for their ascetics hewn out in
the rocks — just as other sects had. And on the wall of the
Sonbhandar cave is an inscription dating perhaps from about A.D.
200, ascribing it to a Muni Vairadeva and as " fit for the residence
of Arahants " — indicating that it then belonged to the Jains.3

The most interesting group is situated at a place called
Barabar, 16 miles north of Gaya. One there, called the
Kama Chaupar, bears an inscription which records the excava-
tion of the cave in the nineteenth year after the coronation of
A.yoka (B.C. 244).* It is simply a rectangular hall measuring
33 ft. 6 in. by 14, and except in an arched roof rising 4 ft. 8 in.
above walls, 6 ft. I in. in height, it has no architectural feature
of importance. At the right, or west end,
is a low platform as if for an image, and
the walls are polished quite smooth. A
second, called the Sudama or Nyagrodha
cave (Woodcut No. 53), bears an inscrip-
tion of Ajoka's twelfth year, the same
year in which most of his edicts are dated,
B.C. 250, and, consequently, is the oldest
architectural example in India. It dedi-
cates the cave to the mendicants of the
Ajivika sect.5 The cave consists of two
apartments : an outer, 32 ft. 9 in. in length, and 19 ft. 6 in. in

53. Sudama Cave.

1 Near Chapra and Cherand, on the
north bank of the Ganges, some 20 to 25
miles above Patna. — 'Journal of the Asiatic
Society of Bengal,' vol. Ixix. pp. 77ff.

2 Buhler ' On the Indian Sect of the
Jainas,' Eng. trans., pp. 25<f. ; Cunning-
ham, ' Archaeological Survey Reports,'
vol. xi. pp. I7of. ; ante, p. 54, note I.

3 Cunningham, ' Archaeological Survey
Reports,' vol i. p. 25, and plate 13.

4 ' Indian Antiquary,' vol. xx. pp. i68ff.
and 36 iff.

5 The Ajivikas were followers of
Makhali Gosala, a contemporary and
opponent of Mahavira and of Buddha.
They were naked recluse devotees and
fatalists, and were often ranked with the
Digambara Jains. The Vijva cave here,
and the three caves in Nagarjuni hill,

i excavated in the reign of Dajaratha,

CHAP. V.

BIHAR CAVES.

breadth, and beyond this a nearly circular apartment 19 ft. u
in. by 19 ft., in the place usually occupied by the solid
dagaba ; l in front of which the roof hangs down and projects
in a manner very much as if it were intended to represent
thatch. The most interesting of the group is that called
Lomas Rishi, which, though bearing no contemporary in-
scription, certainly belongs to the same age. The frontispiece
is singularly interesting as representing in the rock the form of
the structural chaityas of the age. These, as will be seen from
the woodcut (No 55), were apparently constructed with strong

55. Facade of Lomas Rishi Cave. (From a Photograph by Mr Peppe, C. EX)

wooden posts, sloping slightly inwards, supporting a longitudinal
purlin morticed into their heads, while three smaller purlins on
each side are employed to keep the roof in form. Between the
pillars was a framework of wood, above which are shown five

were all made for the Ajivika ascetics.
1 Indian Antiquary,' vol. xx. p. 362 ;
' Epigraphia Indica,' vol. ii. pp. 272.
274 ; Buhler, ' Sect of the Jainas,'
(English version) p. 39.

1 At Kondivte, in Salsette, near
Bombay, there is a chaitya cave of
more modern date, which possesses a
circular chamber like this, except that it
is sunk perpendicularly into the hill side.
In the older examples it is probable a
relic or some sacred symbol occupied the

cell ; in the later it may have been an

54. Kondivte Cave, Salsette.
Scale, 50 ft. to i in.

image — though we know too little of the
Ajivikas to say of whom or what.

132

BUDDHIST ARCHITECTURE.

BOOK I-

56. Lomas Rishi Cave-

smaller purlins. Over these lies the roof, apparently formed of
three thicknesses of plank, or probably two of timber planks
laid reverse ways, and one of metal or some other substance
externally. The form of the roof is something of a pointed
arch, with a slight ogee point on the summit to form a weather-
ing. The door, like all those of this series, has sloping jambs l
— a peculiarity arising, as we shall afterwards see, from the
lines of the openings following, as in this instance, those of
the supports of the roof.

The interior, as will be seen from the annexed plan (No 56),
is quite plain in form, and does not seem to have been ever
quite completed. It consists of a hall 33
ft. by 19 ft, beyond which is an apartment
of nearly circular form, forming a shrine,
as in the Sudama cave.

Three quarters of a mile to the north-
east of these, in a ridge of granite, known
as the Nagarjuni hill, are three more
caves, each bearing an inscription of Da^aratha, the supposed
grandson and successor of A^oka, about B.C. 220, dedicating
them also to the Ajivikas. The largest is the Gopika or
" Milkmaid's Cave," which is a hall 46 ft. 5 in. long, with circular
ends, and 19 ft. 2 in. wide, with one door in the centre of
the south side. The walls are 6| ft. high, and the vault of
the roof rises to io| ft. in the middle. The whole interior is
polished, but quite plain, and we can only conjecture that it
was formed for a refectory or Dharmas-ala.

The remaining two, called the Vahiyaka 2 and the Vadathika
caves, are so small as hardly to deserve notice. They are on
the north side of the ridge. The first is entered from the end,
and consists of a single chamber 16 ft. 9 in. in length by 11 ft.
3 in. wide, and io£ ft. high at the middle of the vault, with
highly polished walls. The Vadathika cave is also entered
from the end, and is of about the same size.3

Judging from the inscriptions on these caves, the whole

1 Gen. Cunningham ('Archaeological
Reports,' vol. i. p. 45) and others have
called this an Egyptian form. This it
certainly is not, as no Egyptian doorway
had sloping jambs. Nor can it properly
be called Pelasgic. The Pelasgi did use
that form but derived it from stone con-
structions. The Indians only obtained it
from wood. — Conf. Foucher, ' L'Art
Greco-Bouddhique du Gandhara,' tome i.
pp. 107, 108, 130 ; Simpson, in ' Trans-
actions of the Royal Institute of British

Architects,' vol. xxx. (1879-1880), p. 56.

2 By an error in reading the inscription,
this cave was formerly called Vapiyaka
or " Well Cave " ; but the epigraph reads
" Vahiyaka." — ' Indian Antiquary,' vol.
xx. p. 364 note.

3 Gen. Cunningham, ' Archaeological
Reports,' vol. i. p. 50, states the width
of this cave as 4 ft. 3 in., but the small
drawing on plate 19 measures about
ioi ft. ; it is it ft. 3 in. — 'Jour. Asiat.
Soc. Bengal,' vol. xvi., plate at p. 408.

CHAP. V. WESTERN CHAITYA HALLS. 133

were excavated between the date of the Sudama and that of
the " Milkmaid's Cave," so called (which was excavated by
Dajaratha), probably within forty years of that date. They
appear to range, therefore, from B.C. 250 to about 220, and
the Lomas Rishi is probably the most modern — it certainly
is the most richly ornamented. No great amount of elabora-
tion, however, is found in these examples, inasmuch as the
material in which they are excavated is the hardest and
most close-grained granite ; and it was hardly to be expected
that a people, who so recently had been using chiefly wood
as a building material, would have patience sufficient for labours
like these. They have polished them like glass in the interior,
and with that they have been content.

There is yet another small cave of this class — called Sita-
marhi — about 13 miles south of Rajagriha, and 25 miles east
from Gaya. It consists of a chamber rectangular
in plan, and measuring 15 ft. 9 in., by 1 1 ft 3 in.,
which is hollowed out of an isolated granite
boulder lying detached by itself, and not near
any rocks. Inside it is as carefully polished as
any of those at Barabar.1 Its principal interest,
however, is in its section (Woodcut No. 57),
which is that of a pointed arch rising from the
floor level to a height of 6 ft. 7 in., without any
perpendicular sides, which are found in the other
caves here. The jambs of the doorway also slope
inwards from the bottom to the top, about I in.
each. From its peculiarities we might infer that
it is possibly the oldest in the district ; but we
must have a more extended series before we can
form a reliable sequence in this direction. In 57- P^n and Section

, r , . , of Sita-marm. Scale

the meantime, however, we may feel sure that 2o ft. to i in.
this hermitage belongs to the great Mauryan
age, but whether before or after A^oka's time must be left at
present undetermined.2

WESTERN CHAITYA HALLS.

There are in the Western Ghats and elsewhere in the
Bombay Presidency six or seven important chaitya caves whose
dates can be made out, either from inscriptions, or from internal
evidence, with very fair approximate certainty, and all of which
were excavated, if I am not very much mistaken, before the
Christian Era. The oldest of these is situated at a place

1 Beglar in ' Archaeological Survey of India Reports,' vol. viii. pp. 106-107.
* ' Cave Temples,' pp. 52-53.

134

BUDDHIST ARCHITECTURE.

BOOK I.

called
B6r Ghat.

4 miles
There

is

58. Chaitya Cave, Bhaja. (From

a Plan by J. Burgess.)

Scale 50 ft. to i in.

south of the great Karle cave in the
no inscription upon it, but from the
plan (Woodcut No. 58), it will be
perceived that it is a chaitya hall of
the usual plan, but of no great dimen-
sions, being only 60 ft. from the back
of the apse to the mortices (a a), in
which the supports of the wooden
screen once stood. From the wood-
cut (No. 60), it will be perceived that
the pillars of the interior slope in-
wards at a considerable and most un-
pleasing angle. The rood - screen
which closes the front of other caves
of this class is gone, as it is also in
the case of the Kondane and Pital-
khora examples, and in Cave No. 10
at Ajanta. In other examples it is in
stone, and consequently remains, but in those instances, where it
was originally in wood, it has disappeared, though the holes to
receive its posts and the mortices by which it was attached to

the walls are still there. The
ogee fronton was covered with
wooden ornaments, which have
disappeared ; though the pin-
holes remain by which they
were fastened to the stone.
The framework, or truss, that
filled the upper part of the
great front opening, no longer
exists, but what its appearance
was may be judged of by the
numerous representations of
itself with which it is covered,
or from the representation of a
chaitya facade from the con-
temporary rail at Bodh-
Gaya (Woodcut No. 59). and
there are several others on the
rail at Bharaut, which are not

59. Front of a Chaitya Hall.
(From a Bas-relief at Bodh-Gaya.)

only correct elevations of such a facade as this, but represent the
wooden carved ornaments which — according to that authority —
invariably adorned these facades. The only existing example
of this wooden screen is that at Karle,1 but the innumerable

1 Little more than thirty years ago, the screen in Kondane chaitya cave was also
tolerably entire. — 'Archaeological Survey of Western India,' vol. iv., frontispiece,

CHAP. V.

WESTERN CHAITYA HALLS.

135

small repetitions of it, not only here but in all these caves, shows
not only its form, but how universal its employment was. The
rafters of the roof were of wood, and many of them, as may be
seen in the woodcut, remain to the present day. Everything,
fact, that could be made in wood remained in wood, and

in

only the constructive parts necessary for stability were executed
in the rock.

60.

Fa9ade of the Cave at

(From a Photograph.

It is easy to understand that, the first time men undertook
to repeat in stone forms they had only been accustomed to
erect in wood, they should have done so literally. The sloping
inwards of the pillars was requisite to resist the thrust of the
circular roof in the wooden building, but it must have appeared
so awkward in stone that it would hardly be often repeated.
As, however, it was probably almost universal in structural
buildings, the doorways and openings naturally followed the
same lines, hence the sloping jambs. Though these were by

136 BUDDHIST ARCHITECTURE. BOOK I.

no means so objectionable in practice, they varied with the
lines of the supports, and, as these became upright, the jambs
became parallel. In like manner, when it was done, the
architects could hardly fail to perceive that they had wasted
both time and labour in cutting away the rock to make way
for their wooden screen in front. Had they left it standing,
with far less expense they could have got a more ornamental
and more durable feature. This was so self-evident that it
never, so far as is known, was repeated, but it was some time
before the pillars of the interior got quite perpendicular, and
the jambs of the doors quite parallel.

There is very little figure sculpture about this cave ; none
in the interior, and what there is on the facade seems to be of
a very domestic character. But on the pillars in the interior
at g and h in the plan (Woodcut No. 58), we find two emblems,
and at #, £, and /"three others are found somewhat rudely formed,
but which occur again so frequently that it may be worth while
to quote them here (Woodcut No. 6I).1 They are known as the
triratna, or trident — the central point being usually more im-
portant than here shown — the shield, and the chakra, or wheel.
The two first are generally found in combination, as in Woodcuts
Nos. 39 and 45, and the wheel is frequently found edged with
triratna ornaments, as in the central compartment of Woodcut
No. 43 from Amaravati. The fourth emblem here is the triratna,

Triratna. Shield. Chakra. Triratna.

i

in combination with a face, and the fifth is one which is
fan-shaped and frequently repeated on coins and elsewhere,
but to which no name has yet been given.

At a short distance along the face of the scarp, is an excava-
tion of some interest, containing a group of fourteen monolithic
dagabas of various sizes. All of them have the Buddhist-rail
pattern cut round the upper margin of the drum. Five under
the overhanging rock vary in diameter from 4 ft. 8 in. to 6 ft.
3 in., and the front two have the square box only on the dome,
as in the cave, and without the cornice, while the three behind

See ' Cave Temples of India,' plate vij,

CHAP. V.

WESTERN CHAITYA HALLS.

137

62. Capital of a rock-cut Dagaba at Bhaja.
(From a Photograph.)

have also the heavy cornice : that upon the largest being con-
nected with the roof by the stone shaft of the " chhattra " or
umbrella, whilst the other
two had been provided with
wooden shafts. Of the nine
in front, the first from the
north has a handsome
capital 3 ft. 8 in. high, very
elaborately carved (Wood-
cut No. 62) ; most of the
others have been broken.
One or more had only the
box form without abacus,
and on four of them are
holes on the top as if for
relics. The names of the
Theras or priests, still
legible on several of them,
indicate that they served
the place of monuments in
a cemetery, though they may also have been reverenced as
altars to saints.

About 10 miles north-west from Karle, in a ravine of the
Western Ghats, are the Kondane chaitya cave and vihara.
This chaitya is interesting, inasmuch as its facade is even a
more literal reproduction of the wooden forms from which
it was derived than that at Bhaja. Nothing could be more
literal than the copying of the overhanging forms of the con-
structive parts of the facade, which shows no trace of stone
construction in any feature, and which it would be hardly
possible to construct in masonry. Its dimensions differ but
little from those of the Bhaja chaitya, being 66J ft. from the
line of the front pillars to the extremity of the apse, 26 ft.
8 in. wide, and 28 ft. 5 in. high to the crown of the arch ;
the nave was 14 ft. 8 in. wide, surrounded by thirty pillars —
most of which have rotted away, but which incline inwards
as do the side walls of the aisles. The dagaba was 9 ft. in
diameter with a capital, like that at Bhaja, of about double
the usual height. These two chaityas may be considered as
contemporary or nearly so, and they are the finest among
the four which are the very oldest specimens of their class in
the west of India.1

Pitalkhora is a ravine among the Indhyadri hills, about

1 'Cave Temples,' pp. 220-222 and plate 8 ; ' Archaeological Survey of Western
India,' vol. iv. pp. 8-10, and plate ir

138

BUDDHIST ARCHITECTURE.

BOOK I.

12 miles south of Chalisgam in Khandesh, where there are
several caves — Buddhist, Brahmanical, and Jaina. Unfortunately
the chaitya is entirely ruined by the decay of the rock, the front
half of the temple having quite disappeared. From the style
of the viharas, and a few epigraphs, we can only conclude that
it must have ranked quite as early as the preceding.1

The fourth of this series will be treated of among the four
AjantA chaitya temples. The next group of caves, however,
that at Bedsa, 10 or 11 miles south of Karle, shows con-
siderable progress towards lithic construction. The screen is
in stone ; the pillars are more upright though still sloping
slightly inwards, the jambs more nearly parallel ; and, in fact,
we have nearly all the features of a well-designed chaitya cave.
The two pillars in front, however, as will be seen from the plan
(Woodcut No. 63), are so much too large in proportion to the

63. Plan of Cave at Bedsa. (From a Plan by J. Burgess.) Scale 50 ft, to i in.

rest, that they are evidently stambhas, and ought to stand free
instead of supporting a verandah. Their capitals (Woodcut No.

64, next page) are more like the Persepolitan type than almost
any others in India, and are each surmounted by horses and
elephants bearing men and women of bold and free execution.2
From the view (Woodcut No. 65 on page 140) it will be seen

1 ' Cave Temples,' pp. 242-246, and
plate 15; 'Archaeological Survey of
Western India,' vol. iv. pp. 11-12.

2 In the PitalkhorS. vihara, we find the
Persepolitan capital repeated with a
variety of animals over it ; for the Hindu
artists, from their natural aptitude for

modifying and adapting forms, very soon
replaced the bicephalous bull and ram of
the Persian columns by a great variety
of animals, sphinxes, and even human
figures in the most grotesque attitudes. —
Dr. Le Bon, ' Les Monuments de 1'Inde,'
P- I5-

CHAP. V.

WESTERN CHAITYA HALLS.

139

how much the surface is covered with the rail decoration, a
repetition on a small scale of the rails described in the last
section, and which it may here be mentioned is a fair test of

64. Capital of Pillar in front of Cave at Bedsa (From a Photograph.)

the age of any building. It gradually becomes less and less
used after the date of these two chaitya caves, and disappears
wholly in the 4th or 5th centuries ; but during that period its
greater or less prevalence in any building is one of the surest
indications we have of the relative age of any two examples.
In this cave, as will be observed, nearly the whole of the
ornamentation is made up of miniature rails, and repetitions of
window fronts or facades. It has also a semicircular open-
work moulding, like basket-work, which is only found in the
very oldest caves, and is evidently so unsuited for stone-work
that it is no wonder it was dropped very early. No example
of it is known after the Christian Era. There is an inscription
in this cave in an ancient form of letters, but without date, and
this alone is not sufficient to fix its age absolutely without
further evidence.

140

BUDDHIST ARCHITECTURE.

BOOK I.

65-

View on Verandah of Cave at Bedsa. (From a Photograph.)

The next of these early caves is the chaitya at Nasik. Its
pillars internally are so nearly perpendicular that their inclina-
tion might escape detection, and the door jambs are nearly
parallel.

The facade, as seen in the woodcut (No. 66, p. 141), is a
very perfect and complete design, but all its details are copied
from wooden forms, and nothing was executed in wood in this

66. Chaitya Cave at Nasik, (From a Photograph.)

i42 BUDDHIST ARCHITECTURE. BOOK I.

cave but the rafters of the roof internally, and these have fallen
down.

Outside this cave, over the doorway, there is an inscription,
stating that " the villagers of Dhambika gave the carving over
it " ; and another — though imperfect — on the projecting ledge
over the guardian to the left of the entrance, reads that " the
rail-pattern and the Yaksha were made by Nadasiri." A third
inscription, on two of the pillars of the nave, ascribes the com-
pletion of the " Chaityagriha " to Bhatapalika granddaughter (?)
of Maha-Hakusiri. The first two are in nearly pure Maurya
characters, and appear to be about coeval with the inscription of
Krishna-raja in the small vihara close by, which we ascribe to
about B.C. 160 ; and the third can hardly be much later.1
Taking these inscriptions in conjunction with the architecture,
the age of this cave hardly seems doubtful. We may accept
B.C. 1 60 as approximately the date of its inception, though its
completion may be a quarter of a century later, and, if this is so,
it carries back the caves of Bhaja, Bedsa, and the others, to a
period considerably before that time, while, on the other hand,
it as certainly is older than the Karle cave, which appears to
come after it in age, whilst Cave No. 9 at Ajanta may be
quite as old.

KARLfi.

The last of the caves mentioned above, known as that at
Karle, is situated near the railway between Bombay and Poona,
and is the finest of all — the finest, indeed, of its class. It is
certainly the largest as well as the most complete chaitya cave
known in India, and was excavated at a time when the style
was^in its greatest purity. In it all the architectural defects
of the previous examples are removed : the pillars of the nave
are quite perpendicular. The screen is ornamented with
sculpture — its first appearance apparently in such a position —
and the style had reached a perfection that was never after-
wards surpassed.

In this cave there is an inscription in the left end of the
porch, and another on the lion - pillar in front, which are
certainly integral, and the first ascribes its completion to " Setthi
Bhutapala of Vaijayanti " (or Banavasi), whilst the second states
that the lion-pillar was the gift of the " Maharathi Agnimitranaka
son of Goti " ; but neither of these, nor others on pillars, door-
way, and arch, help us to a date. We are thus thrown back
on style, without any help from examples closely akin in

1 ' Epigraphia Indica,' vol. viii. pp. 91-93.

CHAP. V.

KARLE.

details : we have also later inscriptions, of Ushabhadata, the son-
in-law of the Kshatrapa Nahapana,1 and, as the latter belongs
to the beginning of the 2nd century, Ushabhadata cannot be
placed earlier than about A.D. 120. But the cave had been
completed long before this, and we may pretty safely place
it in the century B.C., and possibly early in that century.

67-68.

Section and Plan of Chaitya Cave at Karle. Scale 50 ft. to i in.

The building, as will be seen by the annexed woodcuts
(Nos. 67, 68, 69), resembles, to a very great extent, an early
Christian church in its arrangements : consisting of a nave
and side-aisles, terminating in an apse or semidome, round
which the aisle is carried. The general dimensions of the
interior are 124 ft. 3 in. from the entrance to the back wall, by
45 ft. 6 in. in width. The side-aisles, however, are very much
narrower than in Christian churches, the central one being
25 ft. 7 in., so that the others are only 10 ft. wide, including
the thickness of the pillars. As a scale for comparison, it
may be mentioned that its arrangement and dimensions are
very similar to those of the choir of Norwich Cathedral, or

1 For the K&rle inscriptions, see ' Archaeological Survey of Western India,' vol. iv.
pp. 90-92, 112-113 ; ' Epigraphia Indica,' vol. vii. pp. 47ff.

144

BUDDHIST ARCHITECTURE.

BOOK I.

of the Abbaye aux Hommes at Caen, omitting the outer aisles
in the latter buildings. The thickness of the piers at Norwich

and Caen nearly corresponds to the breadth of the aisles in
the Indian temple. In height, however, Karle is very inferior,
being only 45 ft. from the floor to the apex.

Fifteen pillars on each side separate the nave from the

CHAP. V. KARLE. 145

aisles ; each pillar has a tall base, an octagonal shaft, and
richly ornamented capital, on the inner front of which kneel
two elephants, each bearing two figures, generally a man and
a woman, but sometimes two females, all very much better
executed than such ornaments usually are ; behind are horses
and tigers, each bearing a single figure.1 The seven pillars
behind the altar are plain octagonal piers, without either base
or capital, and the four under the entrance gallery differ con-
siderably from those at the sides. The sculptures on the
capitals supply the place usually occupied by frieze and cornice
in Grecian architecture; and in other examples plain painted
surfaces occupy the same space. Above this springs the roof,
semicircular in general section, but somewhat stilted at the
sides, so as to make its height greater than the semi-diameter.
It is ornamented even at this day by a series of wooden ribs,
probably coeval with the excavation, which prove beyond the
shadow of a doubt that the roof is not a copy of a masonry
arch, but of some sort of timber construction which we cannot
now very well understand.

Immediately under the semidome of the apse, and nearly
where the altar stands in Christian churches, is placed the
Dagaba, in this instance a plain dome, on a two-storeyed circular
drum, the upper margins of each section being carved with rail
ornaments (Woodcut No. 70). Just under the lower of these are
holes or mortices for woodwork, which may have been adorned
with hangings, which some of the sculptured representations
would lead us to suppose was the usual mode of ornamenting
these altars. It is surmounted by a capital or Tee, the base of
which is similar to the one shown on Woodcut No. 15, and
on this still stand the remains of an umbrella in wood, some-
what decayed and distorted by age. This canopy was circular
and minutely carved on the under surface.2

Opposite this is the entrance, consisting of three doorways
under a gallery exactly corresponding with our roodloft, one lead-
ing to the centre, and one to each of the side-aisles ; and over
the gallery the whole end of the hall is open, as in all these
chaitya halls, forming one great window, through which all
the light is admitted. This great window is formed in the
shape of a horseshoe, and exactly resembles those used as
ornaments on the facade of this cave, as well as on those of
Bhaja, Bedsa, and at Nasik described above, and which are
met with everywhere at this age. Within the arch is a frame-

Drawings of some of the pillars are Western India,' vol. iv. plate 13.
given in ' Cave Temples of India,' plates ! ' Cave Temples of India,' p. 235, and

12 and 14, and ' Archaeological Survey of
VOL. I.

plate 13.

146

BUDDHIST ARCHITECTURE.

BOOK I.

work or centering of wood standing free (Woodcut No. 6g\
This, so far as we can judge, is, like the ribs of the interior,
coeval with the building ; l at all events, if it has been renewed,
it is an exact copy of the original form, for it is found repeated
in stone in all the niches of the facade, over the doorways,
and generally as an ornament everywhere, and, with the

70.

View of Interior of Cave at Karle. (From a Photograph.)

Buddhist "rail," copied from Sanchi, forms the most usual
ornament of the style.

The sculpture on the screen wall between the doors is mostly
of much later date than the cave itself. All the figures of
Buddha there represented are of late date and belong to the

1 About 1870 it was reported that this
screen was in danger of falling outwards,
and I wrote repeatedly to India begging
that something might be done to preserve
it ; this was eventually effected, but by
"restoration" rather than by judicious
repair. Only a small portion of the
original ribbing of the Bhaja cave now
remains. That of the Bedsa cave was
destroyed about 1 86 1 ('Journal Bombay
Branch of the Royal Asiatic Society,'

vol. viii. p. 223) ; and it would be a
thousand pities if this, which is the only
original screen in India, were allowed to
perish when a very small outlay would
save it. Like the Iron pillar at Meharaull
close to Delhi, which never rusts, teak
wood that does not decay, though exposed
to the atmosphere for 2000 years, is a
phenomenon worth the attention not only
of antiquaries, but of natural philosphers.

CHAP. V. GREAT CHAITYA AT KARLE. 147

Mahayana school ; the larger pairs of figures, however, are
earlier and may be original. The later inscriptions are of
the time of the Andhra king Pulumavi (dr. A.D. 150).

The presence of the woodwork is an additional proof, if
any were wanted, that there were no arches of construction
in any of these Buddhist buildings. There neither were nor
are any in any Indian building anterior to the Muhammadan
conquest, and very few indeed in any Hindu building afterwards.

To return, however, to Karle, the outer porch is considerably
wider than the body of the building, being 52 ft. wide by 15 ft.
deep, and is closed in front by a screen composed of two stout
octagonal pillars, without either base or capital, supporting what
is now a plain mass of rock, but which was once ornamented by
a wooden gallery forming the principal ornament of the facade.
Above this a dwarf colonnade or attic of four columns between
pilasters admitted light to the great window, and this again was
surmounted by a wooden cornice or ornament of some sort,
though we cannot now restore it, since only the mortices remain
that attached it to the rock.

In advance of this screen stands the lion-pillar, in this
instance a plain shaft with sixteen flutes, or rather faces,
surmounted by a capital not unlike that at Kesariya (Woodcut
No. 7), but at Karle supporting four lions instead of one, and,
for reasons given above (p. 60), they seem almost certainly
to have supported a chakra or Buddhist wheel (Woodcut
No. 8). A similar pillar probably stood on the opposite side,
but it had either fallen or been removed to make way for
the little Hindu temple that now occupies its place.

The absence of the wooden ornaments of the external porch,
as well as our ignorance of the mode in which this temple
was finished laterally, and the porch joined to the main temple,
prevents us from judging what the effect of the front would
have been if belonging to a free-standing building. But the
proportions of such parts as remain are so good, and the effect
of the whole so pleasing, that there can be little hesitation
in ascribing to such a design a tolerably high rank among
architectural compositions.

Of the interior we can judge perfectly, and it certainly
is as solemn and grand as any interior can well be, and the
mode of lighting the most perfect — one undivided volume
of light coming through a single opening overhead at a very
favourable angle, and falling directly on the dagaba or principal
object in the building, leaving the rest in comparative obscurity.
The effect is considerably heightened by the closely set thick
columns that divide the aisles from the nave, as they suffice
to prevent the boundary walls from ever being seen, and,

148 BUDDHIST ARCHITECTURE. BOOK I.

as there are no openings in the walls, the view between the
pillars is practically unlimited.

These peculiarities are found more or less developed in
all the other caves of the same class in India, varying only
with the age and the gradual change that took place from the
more purely wooden forms of these caves to the lithic or stone
architecture of the more modern ones. This is the principal
test by which their relative ages can be determined, and it
proves incontestably that the Karle cave was excavated not
so very long after stone came to be generally used as a
building material in India.

There are caves at AjantS. and probably at Junnar which
are as old as those just described, and supply details that
are wanting in the examples just mentioned.1 Meanwhile, how-
ever, their forms are sufficient to place the history on a firm
basis, and to explain the origin and early progress of the
style with sufficient distinctness.

From the inscriptions and literary evidence, it seems hardly
doubtful that the date of the Karle cave is about B.C. 80, and
that at Nasik about B.C. 150. We have no literary authority
for the date of the two earlier ones, but the archaeological
evidence appears irresistible. The Bhaja and Kondane caves
are so absolutely identical in style with the Lomas Rishi cave
in Bihar (Woodcut No. 55) that they must be of very nearly the
same age. Their pillars and their doorways slope so nearly at
the same angle, and the essential woodenness — if the expression
may be used — of each is so exactly the same, that, the one being
of the age of Ajoka, the others cannot be far removed from the
date of his reign. The Bedsa cave exhibits a degree of progress
so nearly half-way between the Bhaja and Nasik examples, that
it may be dated about B.C. 120. The Pitalkhora cave must also
range about the same age, and the whole six, with one or two to
be described at Ajanta, thus exhibit the progress of the style
during nearly two centuries, and form a basis from which we
may proceed to reason with little hesitation or doubt.

AJANTA.

There are four chaitya caves in the Ajanta series which,
though not so magnificent as some of the four just mentioned,
are nearly as important for the purposes of our history.2 The
oldest there is the chaitya, No. 10, which is situated very near

1 A much fuller account of the Rock- 2 For further particulars regarding the

Temples of India will be found in ' The j Ajanta caves, the reader is referred to

Cave - Temples of India,' 1880, and 'Cave Temples,' pp. 280-346, and

' Archaeological Survey of Western ' Archaeological Survey of Western

India,' vols. iv. and v. India,' vol. iv. pp. 43-59.

CHAP. V.

CHAITYA HALLS AT AJANTA.

149

to the next to be referred to, a little higher up in the rock,
however, and is of nearly twice its dimensions. It is 96 ft. 6 in.

71. Interior of Chaitya Cave No, 10 at Ajanta, (From a Sketch by the Author,)

in depth by 41 ft. 3 in. in width internally, and 36 ft. high. As
may be seen from the annexed view (Woodcut No. 71), the

nave is separated from the
aisles by a range of thirty-
nine plain octagonal shafts,
very slightly inclined inwards,
but without capitals or bases.
The triforium belt is of un-
usual height, and was origin-
ally plastered and painted.
Traces of this can still be
seen, though the design cannot
be made out (Woodcut No.
72). Like Bhaja and Kondand
it had only a wooden or
brick front ; but one of the
most remarkable characteris-
tics of the cave is that it

shows what may be regarded as a sign of transition from wood
to stone in its architectural details. As at Pitalkhora, the ribs
of the aisle are in stone cut in the rock, but copied from the
wooden forms of previous examples ; but too much stress should
not be laid on this feature. The vault of the nave was adorned

72. Cross-section of Cave No. ip at Ajanta,
Scale, about 26 ft. to i in.

BUDDHIST ARCHITECTURE.

BOOK I.

with wooden ribs, the mortices for which are still there, and their
marks can still be traced in the roof, but the wood itself is gone.

There is a short Pali inscription on this cave, at the right
side of the facade, which seems to be integral, but unfortunately
it does not contain names that can be identified ; l but from the
form of the characters a palaeographist would almost certainly
place it considerably anterior to the Christian Era.

Next to this, the second chaitya here (No. 9), and probably
not much later in point of age, is the lowest down on the cliff,
and is of the smallest class, being only 45 ft. by 22 ft. 9 in. in
width, and 23 ft. high. All its woodwork has perished, though
it would not be difficult to restore it from the mortices left and
the representations of itself on the facade. There are several
inscriptions, but they are not integral : they are painted on the
walls, and belong, from the form of their characters, to about
the 6th or 7th century of our era, when the frescoes seem to
have been renewed, so that the real tests of its age are, — first,
its position in the series, which make it, with a neighbouring
vihara(No. 12), undoubtedly one of the oldest there; the other
test is the architecture of its facade, which so much resembles
that of the Nasik chaitya that it cannot be very far off in
date. It may, however, be somewhat earlier, as the pillars in
the interior slope inwards at an angle somewhere between that
found at Bhaja and that at Bedsa ; and, in so far as that is a
test of age, it is in favour of a greater antiquity in the Ajanta
example. Such a criterion, however, dependent on the choice
of the superintendent of the excavation, is far too delicate to
use with much confidence as a chronometrical test.

The facades of both these caves are so much ruined by the
rock falling away that it is impossible to assert that there was
no sculpture on the lower parts. None, certainly, exists in the
interior, where everything depends on painting ; and it is, to
say the least of it, very improbable that any figure-sculpture
adorned No. 9 — the figures of Buddha on the sides of the court
being of much later date — while it seems likely that No. 10
also depended wholly on conventional architectural forms for
its adornment.2

The next chaitya cave in this series (No. 19) is separated
from these two by a very long interval of time. Unfortunately,

1 It may be read — "A facade or
entrance, the gift of Katahadi Vasishthi-
putra." — ' Archaeological Survey of
Western India,' vol. iv. p. 116.

2 What fragments of painting remain
in this cave (No. 10) differ markedly in
the costumes of the figures and their
physiognomy from what we find in the

other caves, and are evidently of much
earlier date. — 'Notes on the Bauddha
Rock-cut Temples of Ajantii,' plates 8-1 1 ;
' Cave Temples,' plate 29 ; ' Inscriptions
from the Cave Temples of Western India,'
plate 34, p. 67 ; Workman's ' Through
Town and Jungle,' p. 159.

CHAP. V.

CHAITYA HALLS AT AjANTA.

no inscription exists upon it which would assist in assigning it
any precise date ; but it belongs to a group of viharas, Nos. 16
and 17, whose date, as we shall after-
wards see, can be fixed with tolerable
certainty as belonging to the end of the
5th century A.D. The cave itself, as will
be seen from the plan (Woodcut No. 73),
is of the smallest size, nearly the same
as No. 9, or 46 ft. 4 in. in length, by
23 ft. 7 in. wide and 24 ft. 4 in. high,
and its arrangements do not differ
much, but its details belong to a totally
different school of art. All trace of wood-
work has disappeared, but the wooden
forms are everywhere repeated in stone,
like the triglyphs and mutules of the

Doric order, long after their original 73- Chaitya No. 19 at Ajanta.

i T\/T it. .LI.- Scale so ft. to i in.

meaning was lost. More than this,

painting in the interval had to a great extent ceased to be the
chief means of decoration, both internally and externally, and
sculpture substituted for it in monumental works ; but the
greatest change of all is that Buddha, in all his attitudes, is
introduced everywhere. In the next woodcut (No. 74) — the
view of the facade — it will be seen how completely figure-
sculpture had superseded the plainer architectural forms of
the earlier caves. The rail ornament, too, has entirely dis-
appeared ; the window heads have been dwarfed down to mere
framings for masks ; but, what is even more significant than
these, is that from a pure atheism we have passed to an over-
whelming idolatry. At Karle the eight figures that originally
adorned the porch are chiefs or donors with their wives, in pairs.
All the figures of Buddha that appear there now are long
subsequent additions. None but mortals were sculptured in
the earlier caves, and among these 5akyamuni nowhere appears.
Here, on the contrary, he is Bhagavat — the Holy One — the
object of worship, and occupies a position in the front of the
dagaba or altar itself (Woodcut No. 75, p. 153), surmounted
by the triple umbrella and as the Numen of the place.

We may be able, in the near future, to fix more nearly
the time in which this portentous change took place in Buddhist
ritual. For the present it is sufficient to remark that images
of Buddha, and their worship, were not known in India much
before the commencement of our era, and that the revolution
was complete by the 4th century, — if not earlier.

Before leaving this cave, however, it may be well to remark
on the change that had taken place in the form of the dagaba

'52

BUDDHIST ARCHITECTURE.

BOOK I.

during these 500 years. If Woodcut No. 75 is compared
with the dagabas in Nos. 70 and 71, it will be seen how much
the low rounded form of the early examples had been conven-
tionalised into a tall steeple-like object. The drum had become
more important than the dome, and was ornamented with archi-
tectural features that have no meaning as applied. But more

74. View of Fa9ade, Chaitya Cave No, 19 at Ajanta. (From a Photograph,

curious still is the form the triple umbrella had assumed. It
had now become a steeple reaching almost to the roof of the
cave, and its original form and meaning would hardly be suspected
by those who were not familiar with the intermediate steps.

I am not aware of more than three umbrellas being found
surmounting any dagaba in the caves, but the following repre-
sentation of a model of one found at Sultanpur, near Jalalabad
(Woodcut No. 76), probably of about the same age, has six
such discs ; and in Bihar numerous models are found with
seven, making with the base and finial nine storeys,1 which

1 Kittoe in 'Journal of the Asiatic | No. 20, p. 80; and Foucher, ' L'Art
Society of Bengal,' vol. xvii. (1847), Greco-Bouddhique du Gandhara,' tome i.
pp. I72ff., plate 6 ; conf. ante, Woodcut p. 79.

CHAP. V.

CHAITYA HALLS AT AJANTA.

'53

afterwards in China became the conventional number for the
nine-storeyed towers of that land.

The last chaitya at Ajanta (No. 26) is of a medium size,
67 ft. 10 in. by 36 ft. 3 in., and 31 ft. 3 in. high, and has a
long inscription and three short ones, but, unfortunately, they
contain nothing to enable us to fix its date with certainty.1
It is certainly more modern than the last-named, its sculptures
are coarser, and their meaning more mythological. We shall

probably not err in assuming
that it was excavated towards
the end of the 6th or beginning
of the 7th century ; and that

75. Rock-cut Dagaba at Ajanta.
(From a Drawing by the Author.)

76.

Small Model found in the Tope at Sul-
tanpur. (From Wilson's 'Ariana
Antiqua,' plate 3.)

the year 600 is not far from its true date.2 An idea of the

1 ' Archaeological Survey of Western
India,' vol. iv. pp. 132-136. In the
longer inscription, the Arhat or Sthavira
Achala is mentioned, who is also spoken
of by Iliuen Tsiang.— Beal, 'Buddhist

Records,' vol. ii. pp. 257-258.

2 'Cave Temples,' pp. 341-345, and
plates 37, 38, 50, and 51; 'Archaeo-
logical Survey of Western India,' vol. iv.
PP- 58, 59, and plates 3 and 36.

BUDDHIST ARCHITECTURE.

BOOK I.

richness of the sculpture on the pillars and triforium of this
chaitya, and of the style of the interior, will be better conveyed
by the accompanying woodcut, No. 77, than by any attempt
at description. The pillars are much in the same style as

77-

Pillars on the left side of the Nave, in Cave No. 26 at Ajanta.
(From a Photograph.)

those in Cave No. i, but even richer and more minute in
detail, closely resembling those in No. 25, immediately pre-
ceding it in position, if not also in age. The triforium, it will
be observed, is overloaded with ornament, and thereby loses

CHAP. V. JUNNAR. 155

seriously in architectural dignity and effect. Historically its
chief interest is in showing how idolatrous Buddhism was
becoming when Brahmanism was about to expel the former
from the country of its birth.

JUNNAR.

Around the old town of Junnar, about 48 miles north from
Poona, are some five separate groups of caves, consisting
altogether of fully a hundred and fifty different excavations —
the majority of them being small. Like other early caves, they
are mostly devoid of figure ornament, and notwithstanding
ten chapel or chaitya caves, scattered among the different
groups, it might perhaps be questioned whether they should all
be classed as Buddhist, or whether some of them at least did
not belong to the Jains or other sects. Fuller illustration and
study of what figure ornament there is must settle this ; but
the inscriptions on certain of the caves indicate that they were
for followers of certain Buddhist schools. These inscriptions
seem to range palaeographically, from about B.C. 100 to A.D. 3OO.1

There are not, it is true, any chaityas among them so magnifi-
cent as that at Karle, nor any probably quite so old as those
at Bhaja and Bedsa ; but there is, in the Gane^a group, a chaitya,
both in plan and dimensions, very like that at Nasik, and a
vihara, quite equal to the finest at that place. The great interest
of the series, however, consists in its possessing examples of forms
not known elsewhere.2 There are, for instance, among others,
six chaitya caves, with square terminations, flat roofs and with-
out internal pillars, and one circular cave which was quite unique
until the discovery of another of the same form at Guntupalle,
near the east coast.

The great peculiarity of the series is the extreme simplicity
of the caves composing it. They are too early to have any

1 The Junnar inscriptions have been as to be almost unintelligible; in 1850,

translated by Dr. Kern, ' Indian Anti- Dr. Wilson described them in the

quary,' vol. vi. pp. 391". ; and by Drs. 'Bombay Journal'; and in 1857 Dr.

Bhagwanlal Indraji and Biihler, ' Cave Stevenson republished their inscriptions,

Temple Inscriptions,' pp. 41 - 55 ;
' Archaeological Survey of Western India,'
vol. iv. pp. 92-98, and 103.

2 These caves have long been known
to antiquarians. In 1833 Colonel Sykes
published a series of inscriptions copied
from them, but without any description
of the caves themselves ('Journal of

with translations, in the eighth volume
of the same journal ; and Mr. W. F.
Sinclair, C.S. , wrote a short account of
them in the ' Indian Antiquary' vol. iii.
pp. 33ff. In November 1874, a hurried
survey was made, the results of which
are given in ' Cave Temples,' pp. 248-262
and plates 17, 18 ; and in the ' Archseo

the Royal Asiatic Society,' vol. iv. pp. logical Survey of Western India,' vol. iv.

287-291). In 1847, Dr. Bird noticed pp. 26-36. Photographs, however, are

them in his ' Historical Researches,' needed to make them more clearly in-

with some wretched lithographs, so bad | telligible.

'56

BUDDHIST ARCHITECTURE.

BOOK I.

figures of Buddha himself,1 but there are not even any of those
figures of men and women which we meet with at Karle and
elsewhere. Everything at Junnar wears an aspect of simplicity
and severity, due partly to the antiquity of the caves of course,
but, so far as known, unequalled elsewhere.

There are evidences in several places that plaster and
painting were employed for the decoration of these caves ;
and among them we find ten chaityas — some unfinished —
mostly without side aisles or the arched facades we find else-
where. Only one — that in Manmoda hill 2 — has any sculpture
on the facade: and as will be seen from Woodcut No. 78,
this is quite unique in style. In a fan -shaped sculpture is
represented the goddess Sri or Gaja Lakshmi, standing with
her two elephants pouring water over her, and behind them,
on each side are two worshippers in attitudes of adoration.
Though so ubiquitous and continuous through all ages, it is
seldom this goddess occupies so important a position as she
does here ; but her history has still to be written. On the
edge of the small semicircular centre of the sculpture on this
facade is an inscription in an alphabet of about the century
H.C., stating that " the hall-front was the gift of Chanda a
Yavana." Above the fronton, on each side of the finial, are
figures : — on the right, of a Nagaraja with a fly flap, and on the
left a figure with a bird on his shoulders, and behind each is
a dagaba in high relief. The interior of this cave is 30 ft. in
length and 12 ft. 6 in. wide between the pillars of the nave,
of which three are blocked out on the right side and two just
commenced on the left ; and the entrance, nearly the width of
the nave, is destroyed by the breaking of the lintel. The
dagaba is plain, but, with the whole interior of the cave, it
has been left quite unfinished, owing apparently to a soft
stratum of rock being met with by the excavators.

A little to the east of this is a series of four cells with a
neatly carved facade ; each door has the chaitya-window arch
over it projecting about 15 in., and is carved below in the
same style as over the doorway of the Nasik chaitya, whilst
between the arches are sculptured dagabas, rail-pattern and
smaller arches — indicative of an early date.3

The Tulja Lena group, about 2 miles west from Junnar,
consists of some dozen excavations, among which one is a
vihara with five cells, another was, perhaps a refection hall,

1 There are some defaced images
probably of late date in a cell in the
Manmoda group, that are probably late
Jaina. — ' Cave Temples,' p. 261; 'Archae-
ological Survey of Western India,' vol.
iv. p. 36.

2 There is a representation of this cave
in Dr. Bird's book, plate 16, but so badly
done that it requires being told what is
intended in order to find it out.

3 A part of this fa$ade is represented
in ' Cave Temples,' plate 17, fig. 5.

•

Munmoda Chaitya Cave at Junnar,
(From a Photograph.)

i58

BUDDHIST ARCHITECTURE.

BOOK I.

and a third is the circular chaitya cave. They are all very
much dilapidated — the fronts having mostly fallen away — but
the carving that remains on two of the facades, consisting of
chaitya-window, rail-pattern and dagaba ornamentation is so
like to what we find at Bhaja, Kondane, Bedsa and Nasik,
that we cannot be far wrong in ascribing this group to the
like early period.1

The plan and section (Woodcuts Xos. 79, 80) will explain

79. Circular Cave, Junnar.

(From a Plan by J. Burgess.)
Scale 50 ft. to i in.

80. Section of Circular Cave, Junnar,

(From a Drawing by J. Burgess.)

Scale 2? ft. to i in.

the form of the circular cave above alluded to. It is not
large, only 25 ft. 6 in. across, while its roof is supported by
twelve plain . octagonal pillars which surround the dagaba.2
The tee has been removed from the dagaba to convert it
into a lingam of Siva, in which form it is now worshipped.
The interest of the arrangement of this cave will be more
apparent when we come to describe the dagabas in Ceylon,
which were encircled with pillars in the same manner as this
one. Meanwhile the following representation (Woodcut No. 81)
of a circular temple from the Buddhist sculptures at Bharaut
may enable us to realise, to some extent at least, the external
form of these temples, which perhaps were much more common
in ancient times than any remains we now possess would
justify us in assuming.

Among the other Manmoda caves are two small unfinished
chaityas and a small vihara beside one of them that have
all octagonal pillars with the water-pot bases and capitals in
their verandahs, but with a square block between the abacus
and architrave. Near the more southerly is an excavation
with an inscription by the minister of Nahapana of A.D. 124,
which must be about the date of these caves.3 In the 5ivaner

1 'Cave Temples/ plate 17, fig. 4.

2 A cave at Guntupalle, in the Goda-
vari district, so far resembles this, that
it is circular, 1 8 ft. in diameter, with a
domed roof, but without pillars, and
contains a dagaba, now converted into a
Linga. — Infra, p. 167.

3 These caves are all pretty fully
described in ' Cave Temples ' and
' Archaeological Survey Western India,'
vol. iv., the two Manmoda chaityas are
No. 18 — 'Cave Temples,' p 260, and
No. 31— at pp. 261-262, and plan on
plate 1 8, fig. 8.

CHAP. V.

CHAITYA AT ELURA.

and Gane-ya Lena groups are four other small chaityas, with
verandah pillars of the same Nasik type;1 and in the latter

81. Round Temple and part of Palace or Monastery.
(From a bas-relief at Bharaut.)

series is one that may justly be regarded as the best example
we have of the chaitya of the ist or 2nd century of our era.2
Its proportions are good, and all the details well understood
and properly applied. The vihara 3 near it, now converted into
a Hindu shrine of Gane^a, measures 50^ ft. wide by 56^ ft. deep,
without pillars, the facade of its verandah being almost a
complete copy of that of the Gautamiputra cave (No. 3) at
Nasik, with six pillars and two antae standing on a bench,
the outside of which is carved with rail-pattern.

ELtiRA.

The celebrated VLrwakarma cave at Elura is a chaitya
of the first class, intermediate, perhaps, in age between the
two last described caves at Ajanta, or it may be as modern

1 The chaityas in the scarp of 5ivaner
are No. 3— 'Cave Temple,' p. 249, and
pillar in ' Cave Temple Inscriptions,' plate
17; No. 51— 'Cave Temples,' p. 251,
plan and section, plate 18, figs. I, 2, and
pillar in ' Cave Temple Inscriptions,'
plate 19; and No. 69 — 'Archaeological
Survey,' vol. iv. p. 30. The Ganera Lena
example is No. 15— 'Cave Temples,'
pp. 256-257 and plan and section, plate
1 8, figs. 6, 7 ; and ' Archaeological Survey

Western India,' vol. iv. p. 32, No. 15.

2 No. 6 — for details see ' Cave
Temples,' pp. 254-256, plan, plate 18, fig.
9 to scale 25 ft. to I in. , and section, fig.
10, to double that scale ; pillars in ' Cave
Temple Inscriptions,' plate 29, p. 54 ;
'Archaeological Survey Western India,'
vol. iv. p. 31. The other, also a vaulted
chaitya here, is No. 32 — ' Cave Temples,'
pp. 257-258.

3 No. 7) ' Cave Temples,' p. 256.

:6o

BUDDHIST ARCHITECTURE.

BOOK I.

as the last. There are unfortunately no inscriptions nor any
traditions that would assist in fixing its age, which must
consequently depend wholly on its position in the series and
its architectural peculiarities.

The dimensions of this cave are considerable, 85 ft. 10 in.
by 43 ft, and 34 ft. high ; the inner end is entirely blocked
up by the dagaba which, instead of being circular as in all

82. Interior of Vuwakarma Buddhist Cave at Klura. (From a Photograph.)

the older examples, has a frontispiece attached to it larger
than that in Cave No. 19 at Ajanta, which, as shown in Woodcut
No. 73, makes it square in front. On this addition is a figure
of Buddha seated with his feet down, and surrounded by
attendants and flying figures in the later style of Mahayana
Buddhist art (Woodcut No. 82). In the roof, all the ribs and
ornaments are cut in the rock, though still copied from wooden
prototypes, and the triforium is sculptured with groups, as in Nos.
19 and 26 of Ajanta. Its most marked characteristic, however,
is the facade, where for the first time we miss the great horse-
shoe opening, which is the most marked feature in all previous

CHAP. V.

WESTERN CHAITYA HALLS.

161

examples. We can still trace a reminiscence of it in the upper
part of the window in the centre (Woodcut No. 83) ; but
it was evidently considered necessary, in this instance, to
reduce the size of the opening, and it is easy to see why

83. Fafade of the Ymvakarma Cave at Eltira. (From a Photograph.)

this was the case. At Bedsa, Karle, Kanheri, and elsewhere,
there was a verandah or porch with a screen in front of the
great window, which prevented the direct rays of the sun from
reaching it, and all the older caves had wooden screens, as
at Karle from which curtains could be hung so as to modify
VOL. I. L

l62

BUDDtflST ARCHITECTURE

BOOK I.

the light to any desired extent. At Eliira, no screen could
ever have existed in front, and wooden additions had long
ceased to be used, so that it consequently became necessary
to reduce the size of the opening. In the two later chaityas
at Ajanta, this is effected by simply reducing their size. At
Elura it was done by dividing it. If we had the structural
examples in which this change was probably first introduced,
we might trace its progress ; l but, as this one is the only rock
example we have of a divided window, we must accept it as
one of the latest modifications of the facades of these chaityas.
Practically, it may be an improvement, as it is still sufficiently
large to light the interior in a satisfactory manner ; but artisti-
cally it seems rather to be regretted. There is a character
and a grandeur about the older design which we miss in this
more domestic-looking arrangement, though it is still a form
of opening not destitute of beauty.

Owing to the sloping nature of the ground in which it
is excavated this cave possesses a forecourt of considerable
extent and of great elegance of design, which gives its fagade
an importance it is not entitled to from any intrinsic merit
of its own.2

KANHERI.

One of the best known and most frequently described
chaityas in India, is that on the island of Salsette, about
1 6 miles north of Bombay, and 6 north-west from Thana,
known as the great Kanheri cave. In dimensions it belongs
to the first rank, being 86 ft. 6 in. in length by 39 ft. 10 in.
wide, and about 38 ft. high. In the verandah there is
an inscription recording that a monk named Buddhaghosha
dedicated one of the middle-sized statues in the porch to the
honour of Bhagavat, i.e. Buddha.3 This does not fix the age
of the cave, but on the two front pillars of the same porch
are inscriptions — or rather fragments of such — from which it
is gathered that the chaitya was begun by two brothers, "the
merchants Gajasena and Gajamitra," in the reign of Gautami-
putra Siriyana 5atakarni, that is, about A.D. i8o.4 This fixes
its date some centuries before Nos. 19 and 26 at Ajanta, but
much later than the great Karle chaitya, of which it is a
literal copy, but in so inferior a style of art that the architecture

1 See the chaitya at Ter, ante, p. 126.

2 'Cave Temples,' pp. 377-379, and
plates 62, 63 ; ' Archaeological Survey
of Western India,' vol. v. pp. 9-13, and
plates 3, and 16-18.

3 This and three other short inscrip-
tions on the same verandah are in

characters of about the 4th or 5th century
A.D. — 'Archaeological Survey of Western
India,' vol. v. p. 77.

4 Loc. cit. pp. 75, 76 ; other inscrip-
tions of the later Andhra kings occur
at Kanheri in caves Nos. 5, 36, and 81,
with more of the same age.

CHAP. V.

WESTERN CHAITYA HALLS

163

of this cave can only be looked upon as an exceptional anomaly,
the principles of whose design are unlike anything else to
be found in India, emanating probably from some individual
caprice, the . origin of which we may never now be able to
recover. The fact, however, that it was undertaken by two
laymen, who may have found the undertaking more expensive
than they had anticipated, or who may have died before it was
finished, may account for its degraded style: the inferior
quality of the rock, and the inexperience of the workmen
of the period, under direction of " the reverend Bodhika " as
overseer, may also have contributed to the result.

Internally the roof was ornamented with timber rafters,
and though these have fallen away, the wooden pins by which
they were fastened to the rock still remain ; and the screen
in front has all the mortices and other indications, as at Karle,
proving that it was intended to be covered with wooden galleries
and framework. What is still more curious, the figures of the
donors with their wives, which adorn the front of the screen
at Karle, are here repeated with necessary variations ; and
the rock at this spot being pretty close-grained, they are the
best carved figures in these caves. They are probably also
the only sculptures of the age of the cave. The occurrence
of such figures here is the more strange as it belonged to an
age when their place was reserved for figures of Buddha, and
when, perhaps at Karle itself, they were cutting away the old
sculptures and old inscriptions, to introduce figures of Buddha,

84. Rail in front of the Chaitya Cave, Kanheri. (From a Dra wing by Mr. H. Cousens.)

either seated cross-legged, or borne on the lotus, supported
by Naga figures at its base.1

1 A tolerably correct representation of
these sculptures is engraved in Langles's
'Monuments anciens et modernes de
l'Hindostan,'tom. ii. p. 81, after Niebuhr.
The curious part of the thing is, that the

Buddhist figures of the Karle fa?ade are
not copied here also. — ' Archaeological
Survey of Western India,' vol. iv., plates
4, and 39-41.

164

BUDDHIST ARCHITECTURE.

BOOK I.

In front of this cave is a dwarf rail (Woodcut No. 84) ;
unfortunately it is so weather-worn that it is difficult to make
out all its details ; but comparing it with the Gautamiputra

rail (Woodcut No. 37) and
the Amaravati rail (Wood-
cut No. 41), it will be seen
that it contains all those
complications that were in-
troduced in the ist and 2nd
centuries, but which were
discontinued in the 4th and
5th, when the rail in any
shape fell into disuse as an
architectural ornament.

If again we compare the
annexed woodcut No. 85,
representing one of the capi-
tals in this cave, with those
in the Karle chaitya, we
find the same degradation
of style as is exhibited in
woodcuts Nos. 103 and 104
(p. 1 85) illustrating the styles
of the Nahapana and Gauta-
miputra viharas at Nasik.

The evidence in fact
seems complete that this
cave was excavated in the
last years of the 2nd century;
but, admitting this, it re-

Capital of a Pillar from the Chaitya Cave
at Kanheri.

mains an anomaly, the like
of which only occurs once
again so far as I know in the

history of Indian architecture, and that in a vihara at Nasik of

the same age, to be described hereafter.1

DHAMNAR.

Near the village of Chandwas, about half-way between Kota
and Ujjain, and 48 miles south-west from Jhalrapathan, in
Rajputana, there exists a series of caves at a place called Dhamnar
which are of considerable extent, but the interest that might

1 For further particulars regarding this
cave the reader is referred to Fergusson's
' Rock-cut Temples, of India,' plates n
and 12 ; to 'Cave Temples of India,' pp.

350-353, and plate 53 ; and to ' Archaeo-
logical Survey of Western India,' vol. iv.
pp. 6 if.

CHAP. V.

WESTERN CHAITYA HALLS.

165

be felt in them is considerably diminished, by their being cut
in a coarse laterite conglomerate, so coarse that all the finer
architectural details had to be worked out in plaster, and that,
having perished with time, only their plans and outlines can
now be recognised. Among the sixty or seventy excavations
here found, one in the principal group, which is entered from
the east end of a broad terrace that still exists tolerably
entire, has a dagaba in front — 9^ ft. in diameter and 14 ft. 3 in.
high — standing in the centre of a small court, in the open air.
Immediately behind it is the cell or sanctuary, in which is a
large figure of Buddha seated cross-legged, and on each side
of the entrance are two large dwarpals, as at Aurangabad. The
cell is isolated by a covered passage running round it, on the

Caves at Dhamnar. (Corrected from a Plan by Gen. Cunningham.)
Scale 50 ft. to i in.

east side of which is a recumbent figure of the dying Buddha,
about 15 ft. long; in the passage behind the cell are three
seated figures of Buddhas with smaller standing figures between ;
and three more stand in the west passage with a sitting one
(Woodcut No. 86).

The next is an excavation 25 ft. wide with a curved inner-end
— the whole length being 26| ft — and containing a circular dagaba
8i ft. in diameter and 16 ft. 3 in. high, which supports the roof.
To the west of it is another chaitya cave of some extent, and
presenting peculiarities of plan not found elsewhere. It is
practically a chaitya cella situated in the midst of a vihara.
The cell in which the dagaba is situated is only 31 ft. 8 in.

i66

BUDDHIST ARCHITECTURE.

BOOK I.

by 1 3 ft. 6 in., but to this must be added the porch or ante-
chapel that extended about 20 ft. further, making the whole
about 52 ft. On two sides, and on half the third, it is sur-
rounded by a verandah leading to cells. The third side never
was finished, but in one of the side cells, measuring 1 5 ft. 8 in.
by 10 ft. 10 in., is a smaller dagaba ; at the back are four cells —
one of them, 17 ft. 4 in. by 1 1 ft. 9 in., with an arched and
ribbed roof; and on the west side are six cells, of which the
third measures loj ft. by 8 ft. 4 in., and has two statues on the
back wall, — the whole making a confused mass of chambers
and chaityas in which all the original parts are confounded,
and all the primitive simplicity of design and arrangement is
lost, to such an extent that, without previous knowledge, they
would hardly be recognisable.1

There are no exact data for determining the age of this
cave, but like all of the series it is late, probably between A.D.
600 and 700, and its great interest is that, on comparing it with
the chaitya and vihara at Bhaja or Bedsa (Woodcuts Nos. 58
and 63), we are enabled to realise the progress and changes that
took place in designing these monuments during the eight or
nine centuries that elapsed between them.2

KHOLVI.

Twenty-two miles south-east from Dhamnar is another
series of caves not so extensive, but interesting as being
probably the most modern group of Buddhist caves in India.
No complete account of them has yet been published,3 but
enough is known to enable us to feel sure how modern they are.
There are between forty and fifty excavations here, mostly
small, and in three groups on the south, east, and north sides
of the hill — the principal caves being on the south face. The
most marked feature about them is the presence of some seven
stupas or dagabas with square bases, in all the larger of which
shrines have been hollowed out for images. One, called Arjun's
House, is a highly ornamented dagaba, originally apparently
some 20 ft. in height, but the upper part being in masonry has

1 In Gen. Cunningham's ' Archaeo-
logical Reports,' vol. ii. the plates 78-83,
for Dhamnar, and 84 for Kholvi, are on
too small a scale to be of much use. There
are errors also in the plan, as in repre-
senting nine cells on the west of the larger
chaitya instead of six ; the facade on
plate 80 belongs, not to " Bhim's bazar,"
but to the " Great Kacheri," as on plate
79 ; the pillars of the verandah in No. 1 1

(p. 273) are not " 3 ft." in height, but
about 5 ft. 8 in.

- A complete survey of the Dhamnar
and Kholvi excavations has not been
published, and they present peculiarities
that only a fully detailed survey would
enable us to understand.

3 ' Journal of Bombay Branch Royal
Asiatic Society,' vol. v. pp. 336-349.

CHAP. V. WESTERN CHAITYA HALLS. 167

fallen away. Inside this is a cell open to the front, in which is
a cross-legged seated figure of Buddha, showing an approach to
the Hindu mode of treating images in their temples, which looks
as if Buddhism was on the verge of disappearing.

The same arrangement is repeated in the only excavation
here which can be called a chaitya hall. It is only 26 ft. by
1 3 ft. internally ; but the whole of the dagaba, which is 8 ft. in
diameter, has been hollowed out to make a cell, in which an
image of Buddha is enshrined. The dagabas, in fact, — here
there are three standing by themselves — have become temples,
and only distinguishable from those of the Hindus by their
circular forms.1

It is probably hardly necessary to say more on this subject
now, as most of the questions, both of art and chronology, will
be again touched upon in the next chapter when describing
the viharas which were attached to the chaityas, and were, in
fact, parts of the same establishments. As mere residences, the
viharas may be deficient in that dignity and unity which
characterises the chaityas, but their number and variety make
up to a great extent for their other deficiencies ; and altogether
their description forms one of the most interesting chapters in
our history.

GUNTUPALLE.

At Jilligerigudem near Guntupalle in the Godavari district,
about 20 miles north from Elor, are a dozen or more
buried stupas, and there are caves at five or six different
places. These were surveyed by Mr A. Rea in 1887, when
he partly excavated one of the mounds which contained a
stone stupa having a drum 18 ft. in diameter and about 7 ft.
high, with a dome of about 1 5 ft. diameter, of which the upper
part only seemed to have been disturbed. Near by were
the broken shafts of what probably had been a large pillared
hall or mandapa.

Two groups of the caves here have been destroyed by hewing
away the partition walls between them. Of the others the
largest group contains a number of cells of quite limited
dimensions — 5 to 6 ft. by 7 or 8 ft. They face south-east,
and at the south-west end are four cells opening from a
verandah v/ith a vaulted roof, one cell being at the left end
and three behind — the central one being set 4^ ft. farther
back than the other two. Close to these is another verandah
with a vaulted roof and two cells opening off it and a vaulted

1 The particulars of the architecture of these caves are taken from Gen. Cunning-
ham's report above alluded to, vol. ii. pp. 280-288.

1 68

BUDDHIST ARCHITECTURE.

BOOK I.

passage between them leading to the third cell. Next are
three more cells grouped by the sides of a vaulted room
about 8 ft. 6 in. by 9 ft 6 in. ; but beyond this most of the
cells are almost destroyed.

The only ornamentation on these caves is the "chaitya
window" over the doors and some of the windows as we
find it in the verandah at Bedsa (Woodcut No. 65), with three
curved lattices, and the terminal above is a circular knob. But
the door jambs, curiously enough, splay outwards or make
the openings widest above ; they appear also to have been
fitted with wooden posts and lintels. Over the windows and
inside doorways only, the projecting horseshoe arch was carved,
whilst two of the principal doors had arched heads.

About a hundred yards south from these monks' cells,
on the brow of the hill in which they are excavated, is a
circular cave, like that at Junnar (Woodcut 80, p. 158), con-

87. Fa9ade of Chaitya Cave at Guntupalle. (From a
Drawing by Mr. A. Rea. ) Scale 10 ft. to i in.

taining a dagaba. This cave is 18 ft. in diameter and 14 ft.
9 in. high, and the drum of the dagaba is 3 ft. 9 in. high
with a diameter of 12 ft. at the floor ; its dome is hemispherical
and of 4 ft. 7 in. radius. Upon it is left a knob, as if part
of the staff of an umbrella, but, perhaps, it is only a fragment
of the original capital which has been hewn away to convert
it into the Saiva Lingam for which it is now worshipped.
The dome of the roof is elliptical, rising 7 ft. 3 in. in the
centre, and is carved with sixteen radiating curved ribs on
which four concentric circular rafters are represented as resting.
At Junnar these were probably of wood, and have long ago

CHAP. V.

WESTERN CHAITYA HALLS.

169

disappeared. In front of the cave is a vestibule about four
yards long with vaulted roof — a thin wall separating it from
the shrine, and in this a drip or representation of part of the
roof of the cave comes down above as in the Sudama cave
(Woodcut No. 53). The entrance of the vestibule is 8 ft. 6 in.
wide below and quite a foot more at the arched head, whilst
the facade over it as will be seen in the woodcut (No. 87)
is carved with a horse-shoe arch, bearing a rude resemblance
to that of the Lomas Rishi cave, and projecting about 14 in.1
The structural stupas and remains of Buddhist sculptures
found about them warrant us in attributing these rock excava-
tions to that religious body ; and the style of the apartments
and of the chaitya cave, compared with the similar remains
at Junnar, seem to refer us to an early period, before large
vihara halls were required. Such considerations lead us to
ascribe these remains to a date about 200 B.C.

1 The editor owes to Mr Rea the use

sf the drawings he made of these Guntu-

ille or Jilligerigudem caves. The above

ccount is based on his reports in Madras

G. O., Nos. 181 and 457 of 1888.— Sketch
plans and sections were published in the
' Journal of the R. Asiatic Society.' New
ser. vol. xix., pp. 508-511.

88. Sphinxes from Buddhist Vihara at Pitalkhorft.

1 7o

BUDDHIST ARCHITECTURE.

BOOK I.

CHAPTER VI.

VIHARAS, OR MONASTERIES.

CONTENTS.

Structural Viharas — Bengal Caves — Western Vihara Caves — Nasik, Ajanta,
Bagh, Dhamnar and Kholvi, Elura, Aurangabad and Kuda Viharas.

STRUCTURAL VIHARAS.

A VlHARA,1 properly speaking is a residence or dwelling,
whether for a monk or an image ; and a group of apartments
for a community of monks is, strictly speaking a Sangharama
or monastery. The word Vihara, however, like Dagaba, came
from Ceylon, where it was used to designate not only a cell but
also any monastic establishment,2 and this extended application
has come to be generally understood by us ; and with this
explanation we employ it.

We are almost more dependent on rock-cut examples for
our knowledge of the Viharas or monasteries of the Buddhists
than we are for that of their Chaityas or churches ; a circum-
stance more to be regretted in this instance than in the other.
In a chaitya hall the interior is naturally the principal object,
and where the art of the architect would be principally lavished.
Next would come the facade. The sides and apse are com-
paratively insignificant and incapable of ornament. The fagades
and the interior can be as well expressed in the rock as when

1 As immediately to be explained,
" Viharas "is applied only to monasteries,
the abodes of monks or hermits. It was
not, however, used in former times
in that restricted sense only. Hiuen
Tsiang calls the Great Tower at Bodh-
Gaya a vihara, and describes similar
towers at Nalanda, 200 and 300 ft. high,
as viharas. The ' Mahawansa ' also
applies the term indiscriminately to
temples of a certain class, and to resi-
dences. The name was used to dis-
tinguish them from stupas or towers,

which were relic shrines, or erected as
memorials of persons or events, and never
were residences or simulated to be such,
or contained images, till the last gasp of
the style as at Kholvi. Strictly speak-
ing, the monasteries ought to be called
Sangharamas, but, to avoid multiplica-
tion of terms, vihara is used in this work
as the synonym of monastery.

2 So also in Nepal it is applied to
monasteries. — Oldfield's 'Sketches from
Nepal,' vol. ii. pp. 275ff.

CHAP. VI.

VIHARAS.

171

standing free ; but the case is different with the viharas. A
court or hall surrounded with cells is not an imposing archi-
tectural object. Where the court has galleries two or three
storeys in height, and the pillars that support these are richly
carved, it may attain an amount of picturesqueness we find in
our old hostelries, or of that class of beauty that prevails in the
courts of Spanish monasteries. Such was, I believe, the form
many of the Indian structural viharas may have taken, but
which could hardly be repeated in the rock ; and, unless some
representations are discovered among the paintings or sculptures,
we shall probably never know, though we may guess, what the
original appearances may have been.

There was, however, I believe, another form of Vihara even
less capable of being repeated in the rock. It was pyramidal,
and is the original of all the temples of southern India. Take,
for instance, a description of the Sangharama mentioned both
by Fah Hian and Hiuen Tsiang,1 though neither of them, it must
be confessed, ever saw it, which accounts in part for some
absurdities in the description : — "The building," says Fah-Hian,
" has altogether five storeys. The lowest is shaped with
elephant figures, and has 500 stone cells in it ; the second is
made with lion shapes, and has 400 chambers ; the third is
made with horse shapes, and has 300 chambers ; the fourth
is made with ox shapes, and has 200 chambers ; and the fifth
is made with dove shapes, and has 100 chambers in it" — and
the account given of it by Hiuen Tsiang is practically the
same. At first sight, and especially in the earlier translations,
this looks wild enough ; but if we understand by it that the
several storeys were adorned with elephants, lions, horses, etc.,
we get a mode of decoration which began at Karle where a
range of elephants adorns the lower storey, and was continued
with variations to Halebid, where, as we shall see further on,
all these five animals are, in the I3th century, superimposed
upon one another exactly as here recounted.

The woodcut (No. 89) on next page, taken from one of the
raths at Mamallapuram or " Seven Pagodas," probably correctly
represents such a structure, and I believe also the form of a
great many ancient viharas in India. The diagram (No. 90)
is intended to explain what probably were the internal arrange-
ments of such a structure. As far as it can be understood
from the rock-cut examples we have, the centre was occupied
by halls of varying dimensions according to height, supported

1 Beal's ' Buddhist Records,' vol. i.
Introd. p. 69, and vol. ii. pp. 2141". ;
conf, Julien's ' Memoires,' tome ii.
p. IO2. This monastery was probably at

.Sri .Sailam, on the Krishna river. The
present Hindu temple will be noticed at
the end of Book III. , chapter iv., p. 408.

172

BUDDHIST ARCHITECTURE. BOOK I.

89. Dharmaraja Rath at Mamallapuram. (From a Photograph.

90. Diagram explanatory of the arrangement of a Buddhist Vihara of Four Storeys in

Height.

CHAP. VI.

VIHARAS.

91- 92-

Square and oblong Cells from a Bas-relief

at Bharaut.

by wooden posts above the ground - floor, and used as the
common day - rooms of the monks. The sleeping - cells
(Woodcuts Nos. 91, 92) were apparently on the terraces,
and may have been such as
are frequently represented in
the bas-reliefs at Bharaut and
from Gandhara. Alternately
they seem to have been square
and oblong, and with smaller
apartments between. Of
course we must not take too
literally a representation of a
monastery, carried out solidly
in the rock for a different pur-
pose, as an absolutely correct
representation of its original.

The importance, however, of this form, as explaining the
peculiarities of subsequent Buddhist and Dravidian archi-
tecture, is so great that it is well worth quoting here, though
this will be more evident in the sequel than it can be at
present. In construction the breadth, in a structural building,
would probably have been greater in proportion to the height
than in this example, but that is of little consequence for
our present purposes.

It is, of course, always difficult, sometimes impossible, to
realise the form of buildings from verbal descriptions only, and
the Chinese Pilgrims were not adepts at architectural defini-
tions. Still Hiuen Tsiang's description of the great Nalanda
monastery is important, and so germane to our present subject
that it cannot well be passed over.

This celebrated monastery, which was the Monte Casino
of India for the first five centuries of our era, was situated at
the modern Baragaon, 34 miles south-south-west of Patna, and
7 miles north of the old capital of Rajagriha. If not founded
under the auspices of the celebrated Nagarjuna about the
ist century A.D., he at all events resided there, introducing the
Mahayana or "great translation," and making it the seat of
that school for Central India. After his time six successive
kings had built as many viharas on this spot, when one of
them surrounded the whole with a high wall, which can still
be traced, measuring 1600 ft. north and south, by 400 ft,
and enclosing eight separate courts. Externally to this en-
closure were numerous stupas or tower-like viharas, ten or
twelve of which are easily recognised, and have been identified,
with more or less certainty, by General Cunningham, from
the Pilgrim's description. The general appearance of the

I74 BUDDHIST ARCHITECTURE. BOOK I.

place may be gathered from the following : — " In the different
courts the houses of the monks were each four storeys in
height. The pavilions had pillars ornamented with dragons,
and had beams resplendent with all the colours of the rainbow
— rafters richly carved — columns ornamented with jade, painted
red and richly chiselled, and balustrades of carved open work.
The lintels of the doors were decorated with elegance, and
the roofs covered with glazed tiles of brilliant colours, which
multiplied themselves by reflection, and varied the effect at
every moment in a thousand manners." Or as he enthusiasti-
cally sums up : — " The Sangharamas of India are counted by
thousands, but there are none equal to this in majesty or
richness, or the height of their construction."1

From what we know of the effects of Burmese monasteries at
the present day this is probably no exaggeration ; and with its
groves of Mango-trees, and its immense tanks, which still remain,
it must have been, as he says, " an enchanting abode." Here
there resided in his time — within and without the walls —
10,000 priests and neophytes, and religion and philosophy were
taught from a hundred chairs, and here consequently our Pilgrim
sojourned for five years, imbibing the doctrines of the Law
of Buddha. What Cluny and Clairvaux were to France in
the Middle Ages, Nalanda was to Central India, the depository
of all true learning, and the foundation from which it spread
over all the other Buddhist lands ; but still, as in all instances
connected with that strange parallelism which existed between
the two religions, the Buddhists kept five centuries in advance
of the Roman Church in the invention and use of all the
ceremonies and forms common to both religions.

It would indeed be satisfactory if the architecture of this
celebrated monastery could be restored and its arrangements
made clear. Something has been done by Cunningham 2
towards this, and excavations were made by Mr. Broadley
and Captain Marshall. The former, it is feared, destroyed
more than he restored, and his drawings are so imperfect as
to be utterly unintelligible. The latter did not publish his
discoveries. Nothing, however, would probably better repay
a systematic exploration than this celebrated spot, if under-
/ taken by some one experienced in such researches, and qualified
v to make detailed architectural drawings of what is found.

If, however, it should turn out, as hinted above, that the
whole of the superstructure of these viharas was in wood,
either fire or natural decay may have made such havoc among

1 ' Iliouen Thsang,' tome i. p. 151 ; or
Seal's ' Life of Hiuen Tsiang,' p. in.

2 ' Archseological Reports,' vol. i. pp.
28-36, plate 16.

CHAP. VI.

BENGAL CAVES.

all that remains of them, as to leave little to reward the labours
of the explorer. What has been done in this direction certainly
affords no great encouragement to hope for much. At Sultan-
ganj, near Monghyr, a large vihara was cut through by the
railway, but except one remarkable bronze statue of Buddha l
nothing was found of importance. The monastery apparently
consisted of two large courtyards surrounded by cells. What
was found, however, could only have been the foundations,
as there were no doorways to the apartments or means of
communication between each other or with the exterior.2

The vihara excavated by Captain Kittoe and Mr. Thomas,
at Sarnath, seems certainly to have been destroyed by fire.
All that remained was a series of some twenty cells and four
larger halls surrounding a pillared court 50 ft. square. On
one side were three cells evidently forming a sanctuary, as
is frequently found in the later rock-cut examples.3

The excavations conducted by General Cunningham, at the
same place, were hardly more satisfactory in their result. The
two buildings he explored seem to bear the relation to one
another of a vihara 60 ft. square over all, and the temple of
little more than half these dimensions with a projecting porch
on each face.4 Only the foundation of these buildings now
remains, and nothing to indicate how they were originally
finished. The like is doubtless the condition of the extensive
excavations made by the Archaeological Survey during the last
four years at Kasia in Gorakhpur district as well as at Sarnath.
But no detailed account of the results has yet been made avail-
able, and short notices, without plans, are very unsatisfactory,
if not occasionally unintelligible. Foundations of numerous
temples, stupas and other buildings, we learn, were laid bare,
and interesting sculptures and inscriptions discovered.5

We may eventually hit on some representation which may
enable us to form definite ideas on this subject, but till we
do this we probably must be content with the interiors as seen
in the rock-cut examples.

BENGAL CAVES.

None of the Bihar caves can, properly speaking, be called
viharas, in the sense in which the word is generally used, except
perhaps the Son-bhandar cave, which was probably a Jaina or

1 In private hands in Birmingham in
1876.

a 'Journal of the Asiatic Society of
Bengal,' vol. xxxiii. pp. 361 et seqq.

J 'Journal of the Asiatic Society of
Bengal,' vol. xxiii. pp. 469 et seqq.

4 For this and the other Sarnath re-
mains see Cunningham's ' Archaeological
Reports,' vol. i. pp. 114 et seqq., plates
32-34.

5 'Journal of the Royal Asiatic Society,'
1907,' pp. 99Sff. See Note below, p. 207.

I76

BUDDHIST ARCHITECTURE.

BOOK I.

Ajivika excavation. It is a plain rectangular cave, 33 ft. 9 in.
long by 17 ft. wide, and u ft. 7 in. to the springing of the curved
roof (Woodcuts Nos. 93 and 94). It has one door and one

93. Plan of Son-bhandar Caves.
Scale 50 ft. to i in.

94. Section of Son-bhandar Cave.
Scale 25 ft. to i in.

window, but both — like the rest of the cave — without mouldings
or any architectural features that would assist in determining its
age. The jambs of the doorway slope slightly inwards, but not
sufficiently to give an idea of great antiquty. In front there was

a wooden verandah, the mortice holes for which are still visible
in the front wall, as shown in the woodcut No. 95. Such
wooden verandahs were probably common, as they were attached
to many of the caves at Kanheri. As mentioned above, if the
inscription is as early as the excavation, it may be several
centuries later than the Barabar caves ; the cave may, however,
be earlier than the inscription. l

The other Son - bhandar cave is about 30 ft. to the right

1 A detailed account of these Bengal
caves is given in Gen. Sir A. Cunningham's
'Archaeological Reports,' vol. i. pp. 25-27,
40-53, and vol. iii. pp. 140-144; but his

drawings are on too small a scale, how-
ever, and too rough to show all that is
wanted. — 'Cave Temples of India,' pp.
37-46.

CHAP. VI.

WESTERN VIHARA CAVES.

177

of the larger one, and is in all respects similar, except that
its dimensions are only 22 ft. by 17 ft., and the roof has fallen
in. Between the two a mass of rock is left in order to admit
of a stair being cut in it leading to the surface of the rock
above ; but what stood on the platform there has not been
investigated.

The other caves, at Barabar and Nagarjuni, if not exactly
chaityas in the sense in which that term is applied to the
western caves, were at least oratories^ places of prayer and
worship, rather than residences. One Ajivika ascetic may have
resided in them, but for the purpose of performing the necessary
services. There are no separate cells in them, nor any division
that can be considered as separating the ceremonial from the
domestic uses of the cave, and they must consequently, for the
present at least, be classed as chaityas rather than viharas.

The case is widely different when we turn to the caves in
Orissa, which are among the most interesting, though at the
same time the most anomalous, of all the caves in India.
With possibly one or two exceptions belonging to other sects,
they were evidently excavated for the Jains. Till comparatively
recently, however, they were mistaken for Buddhist, but this
they clearly never were ; hence they must be described in a
subsequent section of this work (Book V. chapter ii., in Vol. ii.).

WESTERN VIHARA CAVES.

The oldest cave in western India is probably a small vih^ra
to the west of the Bhaja group, which is unique of its kind. It
faces north and consists of a
verandah i/| ft. in length
by 7 ft. wide at the east end
and 9^ ft. at the west, with a
hall, also of somewhat irregular
form, 1 6 ft. deep by 16 ft 7 in.
across, exclusive of a bench
21 in. broad along the east
side. The accompanying plan
(Woodcut No. 96) will show
the arrangement of the four
cells entering from the hall
and one from the verandah,
in three of which are stone
beds ; besides, there are three
cells, or cubicles, with a separate entrance outside the verandah
to the left, each with its stone bed — usually an indication of
early date.

VOL. I. M

96. Plan of small Vihara at Bhaja.

(From a plan by Mr. H. Cousens.)

Scale 25 ft. to i in.

i78

BUDDHIST ARCHITECTURE.

BOOK I.

The notable feature, however, are the sculptures of the cave :
a pillar and pilaster in the east end of the verandah have the
bell-shaped quasi-Persepolitan capitals we find on some of the
A.yoka lats, and which became more Indianised in detail at
Bedsa, Karl£, and elsewhere ; the figures that surmount these
are here of exceptional form, being human female busts on
bovine bodies (Woodcut No. 97). The jambs of the cell
doorways slope inwards, and the walls above and between
them are ornamented with the chaitya window pattern. On
each wall are three deep niches roofed by three chaitya arches ;
and between those on the east side, corresponding to the cell
doors opposite, are two panels containing standing figures or
guards, each with a staff or spear and
peculiar head-dresses. In the verandah
are three others in very unusual and elabo-
rate costumes and head-dresses ; one hold-
ing two arrows and another a bow. And
on the west end of the verandah is a large
sculpture, divided by the cell door. On
one side we have Surya and his two wives,
in his chariot drawn (as in the Greek
mythology) by four horses, and apparently
accompanied by two guards on horseback,
with what may be meant for demons of
darkness below the horses ; the other half
represents two large figures on an elephant,
with many smaller ones below and around,
97. Capital of Pilaster at and two sacred trees. In the verandah of
Bhaja. Scale about i/i6th. the Ananta Gumpha at Khandagiri and at
Bodh-Gaya are also found representations
of Surya, and on the Lahul lota both parts of this sculpture
are represented with but slight differences.1 Are these traces of
the Suryopasakas or Saurapatas, the sun-worshippers, who were
long an influential sect in India,2 or how are we to account for
such sculptures?

Besides this, among the Buddhist caves of western India
there are at least six or seven viharas which we know for
certain were excavated before the Christian Era. There are

1 This cave has been pretty fully
illustrated in 'Cave Temples,' pp. 513-
517, and plates 96-98; 'Archaeological
Survey of Western India,' vol. iv. pp. 3-6,
and plate 6; and in Le Bon's c Les Monu-
ments de 1'Inde,' p. 42.

8 The five orthodox divisions of Hindus
are the 5aivas, Vaishnavas, Jiaktis,
Saurapatas, and Ganapatyas, but the

last two are of very limited extent, though
sun-worship is still found among the
Kathis and other tribes in Gujarat. —
' Archaeological Survey Western India,'
vol. ix. pp. 72ff ; Beal, ' Buddhist
Records,' vol. i. p. 223 ; vol. ii. pp. 188,
274; Al-Beruni's 'India,' Sachau's trans,
vol. i. pp. 1 1 6, 121, 298; and 'Bombay
Gazetteer,' vol. ix. part i. pp. 257, 393.

CHAP. VI. WESTERN VIHARA CAVES. 179

possibly forty, but the evidences are not so clear as to enable
us to feel confident in affixing dates to them. The few that
are known are those attached to the chaityas at Bhaja and
Bedsa (Woodcuts Nos. 58, 63), the two oldest at Ajanta, Nos. 12
and 13, and those at Nasik, Pitalkhora, and Kondane. Those
at Karle also, but they have been altered and enlarged, and are
much destroyed by the rock falling away, so that it would be
difficult to describe them ; l they are excavated in two or three
storeys, and the earlier portions are without ornamentation,
but almost certainly coeval with the chaitya itself. At Junnar
there are several, which are very old, and at Sana, Junagadh and
Talaja, in Gujarat, there are numbers of very ancient date.2

One of the oldest of these is that attached to the chaitya at
Bhaja (ante, Woodcut No. 58). It is five-celled ; three of these
have single stone beds in them, one is double-bedded, and one is
without that uncomfortable piece of furniture. In front of these
are two long stone benches at either end of a hall 33 ft. in length.
It is not clear whether this hall was always open as at present,
but, if it was closed, it was by a wooden screen like the
chaitya beside it, which is undoubtedly of the same age.
They are indeed parts of one design. The same may be said
of the Bedsa vihara, though placed a little further apart. In
this case, however, there are three cells with stone beds in the
verandah of the chaitya, and a fourth was commenced, when
apparently it was determined to remove the residence a little
further off, and no instance, I believe, occurs afterwards in
which they were so conjoined, till at least a very late date,
as, perhaps, at Dhamnar (Woodcut No. 86), all the parts got
again confounded together. As will be seen from the plan
(Woodcut No. 63) it is exceptional in form, being apsidal like
the chaitya itself. It is not clear whether this is a copy of any
existing wooden erection, or whether it was that, being the first
attempt at an independent vihara in the rock, they thought it
ought to resemble a chaitya in plan. My impression is that the
latter is the true explanation ; such an arrangement in a free-
standing structure intended for a residence would be absurd,
but we are here assisting at the " incunabula " of the style, and
must not be surprised at anomalies.

No. 12 at Ajanta is a square hall, measuring 36 ft. 8 in. each
way. It has no pillars, and its only ornament consists of horse-

1 For a detailed account and plans of three are chaitya temples ; and at Talaja',
these viharas, see 'Cave Temples,' 30 miles south of Gogha, are about thirty
pp. 24of., and plate 9; and 'Archseo- more. They are all very plain, with
logical Survey of Western India,' vol. iv. J scarcely any sculpture, and are probably
p. 25, and plate 13. i as early as any in western India.

2 At Sana, about 20 miles north-east i — ' Archaeological Survey of Western
of Diu, are sixty-two caves, of which India,' vol. ii. pp. 147-150.

i8o BUDDHIST ARCHITECTURE. BOOK I.

shoe arches over the doors of the twelve cells, with a band of
connecting rail-pattern and pairs of smaller arches over recesses
between. The cells have each two stone beds, and altogether
this vihara bears so close a resemblance to the one at Bhaja, as
also to the smaller one, No. 14 at Nasik, and to that at Kondane
as to assign it to a very early place among those here, and
coeval with the chaitya No. 10. Unfortunately, the rock over
its front has given way, and carried with it the facade, which
probably was the most ornamental part of the design.

Close to No. 12 is cave 13, which may be as old as anything
at Ajanta, but its front also has fallen away, and we have left
only a hall 13^ ft wide by i6| ft. surrounded by seven cells — in
all of which are the stone couches or beds characteristic of the
cells of an early age. No decorative feature appears on its walls.

The most marked characteristic of the early viharas on the
western side of India — if we except the Surya cave at Bhaja,
which is not Buddhist — is that, unlike their eastern Jaina
contemporaries, they are wholly devoid of figure-sculpture : no
bassi-rilievi, not even an emblem, relieves the severity of their
simplicity. Over the doorways of the cells there are the usual
horse-shoe arches, copied from the windows of the great chaityas,
and the invariable Buddhist rail repeated everywhere as a
stringcourse, with an occasional pillar or pilaster to relieve the
monotony.

The curious difference between the exuberance of figure-
sculpture in the east and its total absence in the west in the
pre-Christian Era caves, can only be ascribed to the different
religions to which they respectively belong. Looking, however,
at the progress made of late years in these subjects, there may
possibly be further reasons for this difference which, when
analysed, will throw fresh light on the early history of Jainism
and Buddhism. Meanwhile, it may be worthy of remark, that
the only living representation that is common to both sides of
India, is the presence of the three-headed Naga on the facade
of the Nasik chaitya (Woodcut No. 66), and its appearance in a
similar position on the Chulakama or Sarpa and Ananta caves
at Udayagiri in Orissa. It points to a prominent feature in early
Buddhist and Jaina mythology, which was probably encouraged
by cognate or identical legends respecting Sakyamuni and
Parrvanath. Besides this the three, five, or seven-headed Naga
occurs so frequently at Bharaut, Sanchi and elsewhere, that his
presence here cannot be called a distinctive peculiarity.

Cave No. n at Ajanta is much in advance of Nos. 12 and
13, there being four pillars in its centre (Woodcut No. 98). It
has seven cells inside, but the sanctuary is so arranged as to
suggest that a cell in the back had been cut through to make

CHAP. VI.

WESTERN VIHARA CAVES.

181

3. Cave n at Ajanta.
Scale 50 ft. to i in.

room for it — and thus it may be of much later date than the

cave ; the sculptures in the verandah also are not of an early

age. A small cell is excavated high

up in the side wall of the shrine — an

arrangement not found in any other of

the caves here. The hall is 37 ft. by

28, and 10 ft. high ; the four octagonal

columns are of unusually clumsy style,

from which it has been inferred that this

was one of the earliest examples of the

introduction of pillars into viharas ; and

the close proximity to the early chaitya

No. 10 was thought to countenance this.

On further consideration, however, it

appears that this cave, in its present

form, must be placed considerably later

than caves 12 and 13, and dating after

the Christian Era.

The next step after the introduction of four pillars to
support the roof,1 as in cave No. 1 1 at Ajanta (Woodcut No. 98),
was to introduce twelve pillars
for this purpose, there being no
intermediate number which would
divide by four, and admit of an
opening in the centre of every
side. This arrangement is shown
in the Woodcut (No. 99), repre-
senting the plan of the cave No. 2
at Ajanta. Before this stage of
cave architecture had been reached,
the worship had degenerated con-
siderably from its original form ;
and these larger caves always
possess a sanctuary containing
an image of Buddha. There are
sometimes, besides this, as in the
instance under consideration, two
side chapels, like those in Catholic churches, containing images
of subordinate saints, or probably of donors or benefactors.

The next and most extensive arrangement of these square
monastery-caves is that in which twenty pillars are placed in the
floor, so as to support the roof, six on each side, counting the
corner pillars twice. There are several of these large caves at

99. Cave 2 at Ajanta.
(From a Plan by the Author.)
Scale 50 ft. to i in.

1 Among the very early caves (prob-
ably Jaina) at Junagadh are examples of
a cave with four, and of another with one

pillar supporting the roof. — ' Archaeo-
logical Survey of Western India,' vol. ii.
pp. 139-140, and plate 17.

183

BUDDHIST ARCHITECTURE.

BOOK I.

Ajanta and elsewhere ; one at Bagh, in Malwa, measuring 82 ft.

by 80, represented in the woodcut (No. 100), has, besides the

ordinary comple-
ment, four additional
pillars in the centre;
these were intro-
duced evidently in
consequence of the
rock not being
sufficiently homo-
geneous and perfect
to support itself
without this addi-
tional precaution ;
and there is a Jaina
cave at Dharasimva
in the Dekhan, which
has, in the centre, a
square of twelve
additional pillars.1

These — which
might be classed,
according to the

ioo. Cave 3 at B^gh. (From a Plan by the Editor.) terms Used in Greek

Scale 50 ft. to i in. architecture, as

a style, when having no pillars ; distyle, when with two pillars

in each face ; tetrastyle,
with four; and hexastyle
with six — form the lead-
ing and most character-
istic division of these
excavations, and with
slight modification are
to be found in all the
later series.

The forms, however,
of many are so various
and so abnormal that it
would require a far more
extended classification
to enable us to describe
and include them all.
In many instances the
great depth of the cave which this square arrangement required

1 'Cave Temples,' pp. 5036°. and plate 93; 'Archaeological Survey of Western
India,' vol. iii. pp. 4-9, and plate 2.

rot. Darbar Cave, Salsette. (From a Plan by
Mr. A. A. West.) Scale 50 ft. to i in.

CHAP. VI.

NASIK VIHARAS.

183

was felt to be inconvenient; and a more oblong form was
adopted, as in the so-called Darbar cave at Kanheri (Woodcut
No. 101), where, besides, the sanctuary is projected forward, and
assists, with the pillars, to support the roof. In some examples
this is carried even further, and the sanctuary, standing boldly
forward to the centre of the hall, forms in reality the only
support. This, however, is a late and Brahmanical arrange-
ment, and must be considered more as an economical than an
architectural improvement. Indeed by it the dignity and beauty
of the whole composition are almost entirely destroyed.

NASIK VIHARAS.

The two most interesting series of caves for the investigation
of the history of the later developments of the Vihara system,
are those at Nasik and Ajanta. The latter is by far the most
extensive, consisting of twenty-six first-class caves, four of
which are chaityas. The former group numbers, it is true,
seventeen excavations, but only six or seven of these can be
called first-class, and it possesses only one chaitya. The others
are small excavations of no particular merit or interest. Ajanta
has also the advantage of retaining a considerable portion of
the paintings which once adorned the walls of all viharas erected
subsequently to the Christian Era, while these have almost
entirely disappeared at Nasik, though there seems very little
doubt that the walls of all the greater viharas there were once
so ornamented. This indeed was one of the great distinctions
between them and the earlier primitive cells of the monks before
the Christian Era. The Buddhist church between Asoka and
Kanishka was in the same position as that of Christianity
between Constantine and Gregory the Great. It was the last-
named pontiff who inaugurated the pomp and ceremonial
of the Middle Ages. It might, therefore, under certain circum-
stances be expedient to describe the Ajanta viharas first ; but
they are singularly deficient in well-preserved inscriptions con-
taining recognisable names. Nasik, on the other hand, is
peculiarly rich in this respect, and the history of the series can
be made out with very tolerable approximative certainty.1

1 These inscriptions were first copied
by Lieut. Brett, and published with
translations by Dr. Stevenson, in the fifth
volume of the ' Journal Bombay Branch
of the R. Asiatic Society,' pp. 39 et seqq.,
plates i to 1 6. They were afterwards
revised by Messrs E. W. and A. A. West,
in the seventh volume of the same journal,
pp. 37, etscqq., and translated by Professor
Bhandarkar in the ' Transactions of the

International Congress of Orientalists,'
1874. A revised translation was made
by Professor G. J. Biihler, and published
in the ' Archaeological Survey of Western
India,' vol. iv. pp. 98-116; and they
have lastly been revised by Mons. E.
Senart in ' Epigraphia Indica,' vol. viii.
(1905)) PP- 59-96, in which, however,
he has adopted a different numeration
of the caves from that in common use.

1 84

BUDDHIST ARCHITECTURE.

BOOK I.

The chaitya cave was, as above stated, commenced about
1 60 years before Christ, and the vihara of the same age attached
to it, is the small one (No. 14) close to it, and on a lower level
than those now on each side of it, and consequently more likely
to be what we are looking for than they are. It is a simple square
hall measuring 14 ft. each way, with two square cells in three of
its sides, the fourth opening on a verandah with two octagon
pillars in front. The only ornament of the interior is a horse-
shoe arch over each cell door, connected by a simple Buddhist rail.
In every detail it is in fact identical with the two old viharas
Nos. 13 and 12 at Ajanta, and it bears an inscription of Krishna
Raja, who seems almost certainly to be the second of the
Andhrabhritya race, and who probably ascended the throne about
B.C. I/O, and ruled for 18 years. The architectural details accord
perfectly with those of the chaitya, and the age ascribed to it.

Turning from these, which practically belong to the last
chapter rather than to this, the interest is centred in three great
viharas, the oldest of which (No. 8) bears the name of Nahapana

(Woodcut No. 102), the second
(No. 3) that of Gautamiputra, and
the third (No. 1 5) that of Sri Yajna,
— if our chronology is correct,
their dates are thus fixed as about
A.D. 100, 130, and 1 80.

The two principal viharas at
Nasik, Nos. 3 and 8, are so similar
in dimensions and in all their
arrangements, that it is almost
impossible to distinguish between
their plans on paper. They are
both square halls measuring more
than 40 ft. each side, without any
pillars in the centre, and are sur-
rounded on three sides by sixteen cells of nearly the same dimen-
sions. On the fourth or front side is a six-pillared verandah, in
the one case with a cell at each end, in the other with only one
cell, which is the most marked distinction between the two plans.
The architecture, too, is in some respects so similar that we can
hardly hesitate in assuming that the one is an intentional copy of
the other. It is in fact the problem of the great cave at Kanheri,
being a copy of that at Karle repeated here.1 Only the differ-
ence in age between the two chaityas being greater, the degrada-
tion in style is much greater than here, where it appears to be

102. Nahap&na Vihira, Nasik.
(From a Plan by J. Burgess.)
Scale 50 ft. to i in.

1 Ante, p. 162. See also plate 11 of my folio work on the 'Rock-cut Temples,'
where the pillars of the two caves are contrasted as here.

CHAP. VI.

NASIK VIHARAS.

185

little more than a generation, and may be largely due to work-
men from a different province.

The pillars in the verandah of cave No. 8 (Woodcut No. 103),
are so similar to those in the great Karle chaitya, that we
might hesitate to ascribe any very lengthened interval between
them ; indeed we find inscriptions at Karle of the same
Ushavadata and his wife Dakhamitra, the daughter of Nahapana,

103.

Pillar in Nahapana Cave, Nasik.
(From a Photograph.)

104.

Pillar in GautamJputra Cave, Nasik.
(From a Photograph.)

who give two cells in the verandah of this cave in the years
A.D. 1 19-123 ;l and as the inscriptions chiefly record endowments,
this cave may have been excavated, for aught we know, a century
or more before these donations were made and recorded. There

1 The dates in Ushavadata's inscriptions
are 41, 42, and 45 ; and as Nahapana,
inajunnar inscription, gives the date 46,
the latter must have been alive at the

earlier dates ; and as the Kshatrapas use
the Saka. era, the Nasik dates correspond
to A.D. 119, 120 and 123.

186 BUDDHIST ARCHITECTURE. BOOK I.

are also minor differences in the proportions of the pillars and
the execution of the figure sculptures, that may be indications of
some difference of age. On the other hand, no vihara on this
side of India has a facade more richly ornamented than this.
Those at Bhaja and Bedsa are quite plain, and those around
Karle, though richer, are inferior to this, so that on the whole
the architectural evidence tends to confirm the date as sub-
sequent to the Christian Era ; and if so, then Karle may be of
somewhat later date than had been previously ascribed to it.

The pillars of the Gautamiputra cave No. 3, as will be seen
from the last woodcut (No. 104), have lost much of the elegance
of those last described. How far this difference is to be
ascribed to the first cave having been constructed under a
Malwa architect, whilst the latter was probably executed by a
Telugu or Dakhani, it is hard to say. Instead of the graceful
bell-shaped Persian capitals, we have the pudding forms that
afterwards became so prevalent. The shafts are straight posts,
and have no bases, and the whole shows an inferiority not to be
mistaken. The carved and sculptured doorway also belongs
to a much less elegant style. Besides this, there are three things
here which prove almost incontestably that it belongs to the
same age as the Amaravati tope erected in the 2nd century — the
rail in front, already given (Woodcut No. 37, p. 1 1 3), the pilaster
at the end of the verandah, and the bas-relief of a dagaba,
which occupies the same position on the back wall in this cave
that the Bhairava with the club now occupies in No. 8.1 It has
the same attendants, and the same superfluity of umbrellas, as
are found there, so that altogether the age of the excavation
can hardly be considered doubtful.

Cave No. 12 is a small vihara, the central hall being 32 ft. by
23 ft, and with only four cells on one side. It had never been
finished, and considerable alterations have been attempted in its
interior at some date long subsequent to its first excavation,
apparently to adapt it to Hindu worship. Its verandah, however,
consisting of two attached and two free-standing columns, is
apparently of the same age as the Nahapana cave No. 8.
An inscription upon it states that it was excavated by
" tndragnidatta, the Yavana, a northerner from Dattamitri." 2
None of these names can be recognised, but they point to
an age when foreigners, possibly of the Panjab or Arakhosia,
visited the Dekhan.

The great vihara (No. 15) beyond the chaitya cave, and 12 ft.
above its level, is one of the most important of the series, not

1 This is not original, but is a in cave 3. — ' Cave Temples,' p. 270, and
figure of Bhairava formed out of what plates 19-23.
was originally a dagaba similar to that 2 ' Epigraphia Indica,' vol. viii. p. 91.

CHAP. VI.

NASIK VIHARAS.

187

105. Sri Yajna Cave, No. 15 at Nasik.
(From a Plan by J. Burgess.)
Scale 50 ft. to i in.

only from its size, but from its ordinance and date (Woodcut No.
105). The hall is 61 ft. in depth by 37^ ft. wide at the outer end,
increasing to 44 ft. at the
inner, and with eight cells on
each side. Originally it seems
as if it had been only 40 ft. in
depth, but at a later period
was extended — perhaps by
the lady Vasu, mentioned in
the inscription in the veran-
dah. The addition forms its
most marked peculiarity,
which is that it has a regular
sanctuary at its inner end,
with two richly carved pillars
in front (Woodcut No. 106),
and within, a colossal figure
of Buddha, seated, with flying
and standing attendants,
dwarpals, dwarfs, and all the
accompaniments usually found
in the third and subsequent
centuries belonging to the Mahayana school of Buddhism.

Fortunately we have in this cave an inscription containing a
well-known name. It is said to have been completed by Vasu,
wife of the commander-in-chief of the king .Sriyajna ^"atakarni,
in that king's seventh year, after it had been excavated many
years before by Vopaki, an ascetic, but had remained unfinished.
We are not able to fix the exact year to which this date refers,
but it does not seem doubtful that this king reigned in the last
quarter of the 2nd century, and we consequently have in this
cave a fixed point (about A.D. 185) on which to base our
calculations for the period about that time. Further, over the
doorway of the first cell on the left side is a short inscription, in
letters of the 5th or 6th century, intimating the gift of a cave —
perhaps only the cell — by a lay devotee Mamma.1

Beyond this there is still another excavation, No. 17 — it
can hardly be called a vihara — of very irregular shape, and
covered with sculpture of a date perhaps four centuries more
modern than that of the cave last described. Buddha is there
represented in all his attitudes, standing or sitting, accompanied
by chauri bearers, flying figures, dwarfs, etc. On one side is
a colossal recumbent figure of him attaining Nirvana, which is a
sign of a comparatively modern date. Besides these, there are

1 'Archaeological Survey of Western India,' vol. iv. pp. 114 and Il6; ' Epi-
graphia Indica,' vol. viii. pp. 93, 94.

1 88

BUDDHIST ARCHITECTURE.

BOOK I.

Dnyani Buddhas, Bodhisattwas, and all the modern pantheon
of Buddhism, arranged in admired confusion, as in most

of the

There

caves.

modern

is no inscription,
but from its sculpture
and the form of its pillars
we may safely ascribe it
to the last age of Buddhist
art, say about the year
600 or later. The pillars
approximate closely in
style to those found at
Elephanta, and in the
Brahmanical caves at
Eliira, which, from other
evidence, have been
assigned to dates varying
from 600 to 800 years of
our era.

More has perhaps
been said about the Nasik
caves than their archi-
tectural importance would
seem at first sight to
justify, but they are one
of the most important of
the purely Buddhist
groups.1 Their great
merit, however, is that

they belong to one of the most important of the older Indian
dynasties, known as the Andhrabhrityas, Satakarnis, or Sata-
vahanas. Owing to their coinage being mostly of lead, this
dynasty was for long overlooked by numismatists and others,
and could only be rehabilitated by their inscriptions and their
architectural work, on which these are found inscribed. And
labour on these materials has been rewarded by very important
chronological results.2

AjANTA VlHARAS

As before mentioned, the central group of the four oldest
caves at Ajanta forms the nucleus from which the caves radiate
south-east and south-west — eight in one direction, and fourteen

106.

Pillar in Sri Yajna Cave.

1 For further details, see ' Cave

Temples,' pp. 263 to 279, and plates.

a ' Archaeological Survey of Western

India,' vol. iv. pp. 98-114 ; Bhandarkar,
' Early history of the Dekkan,' pp.
14-44.

CHAP. VI.

AJANTA VIHARAS.

189

in the other. It seems, however, that there was a pause in
the excavation of caves after the first great effort, and that
they were then extended, for some time at least, in a south-west
direction. Thus caves Nos. 14 to 20 form a tolerably consecutive
series, without any violent break. After that, or it may be
contemporaneously with the last named, may be grouped Nos.
8, 7, and 6 ; and, lastly, Nos. 21 to 26 at one end of the series,
and Nos. i to 5 at the other, form the latest and most ornate
group of the whole series.1

As above explained, four in the centre are certainly anterior
to the Christian Era. One, No. 10, is certainly the oldest
here, and may consequently be contemporary with the gateways
at Sanchi ; and with it are associated Nos. 12 and 13. After
this first effort, however, came the pause just alluded to, for
Nos. u, 14, and 15, which
are the only caves we can
safely assign to the next
three centuries, are com-
paratively insignificant,
either in extent or in rich-
ness of detail.

Leaving these, we come
to two viharas, Nos. 16
and 17, which are the most
beautiful here, and, taken
in conjunction with their
paintings, probably the
most interesting viharas in
India.

No. 1 6 is a twenty-
pillared cave, measuring
about 65 ft. each way I07-
(Woodcut No. 107), with
sixteen cells and a regular
sanctuary, in which is a figure of Buddha, seated, with his feet
down. The general appearance of the interior may be judged
of by the following woodcut (No. 108) in outline, but only a
coloured representation in much greater detail could give an
idea of the richness of effect produced by its decoration.2 All

Plan of Cave No. 16 at Ajanta. (From a
Plan by J. Burgess.) Scale 50 ft. to
i in.

1 The caves run in a semicircle along
the north side of the Waghora torrent,
which, after falling over the cliff here,
makes a bend to the north. They were
numbered consecutively, like houses in a
street, beginning at the south-east end,
the first cave there being No. I, the last
accessible cave at the western end being

No. 26. For a plan of the group, see
'Archaeological Survey of Western India,'
vol. iv. plate 21.

2 In Mr. Griffiths's ' Paintings in the
Buddhist Cave Temples of Ajanta,' plate
92, he gives a coloured view of the interior
also of cave i.

190

BUDDHIST ARCHITECTURE.

BOOK I.

the walls are covered with frescoes representing scenes from the
Buddhist jatakas, or from the legends of Buddha's life, and the
roof and pillars by arabesques and ornaments, generally of great
beauty of outline, heightened by the most harmonious colouring.

108. View of Interior of Vihara No. 16, at Ajanta. (From a Sketch by the Author.)

No. 17, which is very similar in plan, was long known as
the Zodiac cave, from the figure of a Buddhist Bhava-chakra
or ' wheel of life ' painted at the left end of its verandah, which
was mistaken by early visitors for a celestial emblem.1 The
general effect of its architecture internally may be gathered

1 'Journal of the Royal Asiatic Society,' 1894, p. 370 and plate; 'Man' for
January 1901. Dr. Bird peeled off many of the figures. — 'Cave Temples,' pp. jogffg.

CHAP. VI. AJANTA VIHARAS. 191

from Woodcut No. 109 (from a photograph), or from the
next woodcut (No. no) representing one of its pillars to
a larger scale, from which the curiously wooden construction
of the roof will be better observed than from the photograph.
It is, in fact, the usual mode of forming flat or terraced roofs
at the present day throughout India, and which consequently

109. View in Cave No 17, at Ajanta. (From a Photograph.)

does not seem to have varied from the 5th century at all
events. As may be gathered from these illustrations, the pillars
in these caves are almost indefinitely varied, generally in pairs,
but no pillars in any one cave are at all like those in any
other. In each cave, however, there is a general harmony of
design and of form, which prevents their variety from being
unpleasing. The effect, on the contrary, is singularly harmonious
and satisfactory. The great interest of these two caves lies,

192

BUDDHIST ARCHITECTURE.

BOOK 1.

however, in their frescoes, which represent Buddhist legends

on a scale and with
a distinctness found
nowhere
India.

else in
The sculp-
tures of Amaravati

— which must be
considerably earlier

— are what most
nearly approach
them; but, as in
most cases, painting
admits of greater
freedom and greater
variety of incident
than sculpture ever
does, and certainly
in this instance
vindicates its claim
to greater phonetic
power. Many of the
frets and architec -
tural details painted
on the roofs and
pillars of these
viharas are also of
great elegance and
appropriateness,1
and, when combined
with the architec-
ture, make up a
whole unrivalled in
India for its ethno-
graphic as well as
for its architectural
beauty.

Fortunately the
age of these two
caves is not altoge-
ther doubtful ; there
is a long inscription
on each, much muti-
lated, it must be con-
fessed, but of which

Pillar in Vihara No, 17, at Ajanta.
(From a Sketch by the Author.)

1 For excellent illustrations of these, mostly in colour, see Mr. Grimths's ' Paintings
in the Buddhist Cave Temples at Ajanta,' vol. ii.

CHAP. VI. AJANTA VIHARAS. 193

enough can be made out to show that they were excavated
under kings of the Vindhyajakti race, one of whom, Pravarasena,
whose name appears in the inscription on No. 16, married a
daughter of a Maharaja Devagupta. And though, as yet, we
cannot fix a definite date for these princes, we may place the
inscriptions epigraphically about A.D. 500, or possibly a little
earlier.1 Hence we may approximately date these two caves in
the end of the 5th century. They are thus considerably more
modern than the Sri Yajna cave, No. 15, at Nasik, which is the
result we would expect to arrive at from their architecture and
the form of their sanctuaries.

Their great interest, therefore, from a historical point of view,
consists in their being almost unique specimens of the
architecture and arts of India during the great Gupta period.

Nos. 1 8, 19, and 20 succeed this group, both in position
and in style, and probably occupied the first half of the 6th
century in construction, bringing down our history to about
A.D. 550.

Before proceeding further in this direction, the cave-diggers
seem to have turned back and excavated Nos. 8, 7, and
6. The last named is the only two-storeyed cave at Ajanta,
and would be very interesting if it were not so fearfully
ruined by damp and decay, owing to the faulty nature of the
rock in which it is excavated. No. 7 has a singularly elegant
verandah, broken by two projecting pavilions. Internally, it
is small, and occupied by a whole pantheon of Buddhas.2 It
resembles somewhat No. 1 5 at Nasik, with which it is perhaps
nearly contemporary.

There still remain the first five caves at the south-east
end, and the six last at the western : one of these is a chaitya,
the other ten are viharas of greater or less dimensions. Some
are only commenced — and two — Nos. 4 and 24 — which were
intended to have been the finest of the series, are left in a very
incomplete state : interesting, however, as showing the whole
process of an excavation from its commencement to its com-
pletion. Xo. 4 is a 28-pillared cave, of which the hall is about
87 ft. square, and except the cells it is nearly finished ; but
No. 24, though the next largest, is planned with 20 pillars
and a hall 73-^ ft. wide by 75 ft. deep — but inside, only the
front aisle has been advanced towards completion, the pillars
in the back and sides being only roughly blocked out. The
verandah, however, had been sculptured in a style showing

1 ' Archaeological Survey of Western
India,' vol. iv. pp. 53, 128.

2 ' Rock-cut Temples,' plate 8. For a
fuller account and illustrations, see ' Cave

Temples of India,' pp. 299 - 300, and
plate 31; 'Archaeological Survey of
Western India,' vol. iv. p. 52, and plates
27 and 28, fig. I.

VOL. I. N

194

BUDDHIST ARCHITECTURE.

BOOK I.

that it was intended to be one of the most highly finished
monasteries in the group. The pillars of the fa£ade, with one
exception, are all destroyed : only the capitals, still attached
to the roof, testify to the beauty of design and life and finish
of the work. The woodcut No. in represents one of these
bracket capitals, carved in much the same pattern as the others.
The true capital having overhanging leaves, analogous to the
Ionic volutes, and which forms so marked a feature in subse-
quent Indian architecture, seems first to have been perfected
about the time this cave was excavated. It is so like in details

to those in cave No. 3 at
Aurangabad1 that there
can be little hesitation
in assigning them to the
same age. The capitals,
pillars, and pilasters in
this and Nos. i, 2, and
2 1 to 26, with the very
similar ones at Auranga-
bad may be taken as the
types of the last and
most elaborate phase of
Buddhist architectural
decoration in Western
India.

Caves Nos. i and 2
are among the most
richly sculptured of the
caves. The facade, in-
deed, of No. i is the
most elaborate and beau-
tiful of its class at Ajanta,
and, with the corre-
sponding caves at the
opposite end, conveys a higher idea of the perfection to which
decorative sculpture had attained at that age than anything else
at Ajanta.2

Capital from Verandah ot Cave 24.
(From a Photograph.)

1 'Archaeological Survey of Western
India,' vol. iii. pp. 66, 69, and plates
44-47-

2 Curiously enough, on the roof of cave
i, there are four square compartments
representing the same scene in different
manners — a king, or very important
personage, drinking out of a cup, with
male and female attendants. What the
story is, is not known, but the persons
represented are not Indians, but Persians,

and the costumes those of the Sassanian
period. — See Mr. Fergusson ' On the
Identification of the Portrait of Chosroes
II. among the Paintings in the Caves at
Ajanta,' in 'Jour. R. Asiatic Society,'
vol. xi. (N.S. ), pp. 155-170. Copies of
these pictures by Mr Griffiths were among
those destroyed by fire in the India
Museum at Kensington. — Griffiths'
' Ajanta Paintings,' vol. ii. plates 94, 95.

CHAP. VI.

AJANTA VIHARAS.

J95

The woodcut No. 112, from a photograph of one of the
pillars of the verandah of Cave I, may help to illustrate
the mode in which decoration is applied to them. The square
base changes into an octagon, but the passage from the

112. Pillar in the \ erandah of Cave i, Ajanta. (From a Photograph.)

one to the other is broken by four little dwarf figures, who
reappear on the capital for the same purpose. Above the
octagon the shaft is adorned with spiral flutes of singularly
pleasing design, bound together with bands of jewelled orna-

196

BUDDHIST ARCHITECTURE.

BOOK I.

ments of great beauty. The capital is ornamented with an
oblong bas-relief in the centre, containing a religious scene, as
is the case with most of those in the cave, and is supported by
flying figures on the brackets, as is also the case in many of the
later caves. On the right front corner pillar in the hall the fluting
is also spiral, but the twist is reversed in the upper section.1

With the last chaitya, which belongs to this group, these
caves carry our history down certainly into the 7th century.
The work in the unfinished caves, I fancy, must have been
arrested by the troubles which took place in Central India
about the year 650, or shortly afterwards, and after which it
is hardly probable that any Buddhist community would have
leisure or means to carry out works, on such a scale at least,
as these Ajanta viharas.

It is, of course, impossible, without a much greater amount
of illustration than is compatible with the nature of this work,
to convey to those who have not seen them any idea of the
various points of interest found in these caves ; but the general
reader will find a more detailed account in the volume on the
'Cave Temples of India,' supplemented in the 4th volume of
the 'Archaeological Survey of Western India.'

The fairly complete series of illustrations of the paintings
as well as the architecture of these caves which we now
possess, forms a valuable contribution to our knowledge,
affording examples of Buddhist art, without admixture from
any other religion, extending from the second century B.C. till
the seventh, after our era ; and besides illustrating the arts
and feelings of those ages, they form a chronometric scale by
which to judge of and synchronise other known series, with
which, however, they differ in several important particulars.
For instance, at Ajanta, there is no single example of those
bell - shaped Persian capitals to pillars, with waterpot bases ;
nor is there any example of animals with riders crowning the
capitals, such as are found at Bedsa, Karle, Nasik, Salsette,
Pitalkhora, and elsewhere.

The earlier copies of the paintings were lost when the dis-
astrous fire at the Crystal Palace, in December 1866, destroyed
Major Gill's facsimiles of the paintings — some twenty-five of them
— many of large size.2 Between 1872 and 1885 a serious effort,

1 ' Archaeological Survey of Western
India,' vol. iv. p. 49, and plate 18, fig. z.

1 The fate of these very remarkable
remains of early Indian art has been most
unfortunate. The Royal Asiatic Society
memorialised the Court of Directors in
1844, that an artist might be engaged to
make accurate copies of the ancient

frescoes referred to in Mr. Fergusson's
account of the Ajanta Caves. This was
promptly and generously approved by the
Court, and Major Robert Gill spent about
twelve years making copies of them ;
these were sent home from time to time,
and were exhibited in the Sydenham
Crystal Palace, where they were unfor-

CHAP. VI.

BAGH.

197

however, was made by the Bombay Government to recover, as
far as possible, this loss, and the publication of Mr Griffiths'
work on a portion of the results is a splendid addition to
our materials.

BAGH.

At a distance of about 1 50 miles a little west of north from
Ajanta, and 30 miles west of Mandu, near a village of the
name of Bagh, in Malwa, there exists a series of viharas
only little less interesting than the later series at Ajanta
They are situated in a secluded ravine in the side of the range
of hills that bounds the valley of the Narbada on the north
and were first visited or at least first described by Lieutenant
Dangerfield, in the second volume of the ' Transactions of the
Literary Society of Bombay' (1818). They have since been
described more in detail by Dr. Impey in the fifth volume
of the ' Journal Bombay Branch of the Royal Asiatic Society '
(1854). Unfortunately the plates that were to accompany
that paper were not published with it, but from them and from
his paper the principal details that follow have been gleaned.

The series consists of eight or nine viharas, some of them
of the largest class, but no chaitya hall, nor does any excavation
of that class seem ever to have been attempted here. On the
other hand, the larger viharas seem to have had a Sala or
schoolroom attached to them, which may also have been
employed, as Dr. Impey suggested, for religious service ; but,
like the Darbar cave at Kanheri, it was more probably a
Dharmajala or refectory. The fact, however, that the sanctu-
aries of the viharas generally have a dagaba in them, instead
of an image of Buddha, points to a distinction which may
hereafter prove of value : possibly they belonged to a Hinayana

tunately destroyed by fire in 1866 — no
photographs or coloured copies of them
having been secured. Mr. Fergusson
and the editor then called the attention
of Government to the urgency of recopy-
ing what still remained — for visitors and
the bats had destroyed much during the
previous twenty or thirty years. Finally
in 1872 a modest subsidy was provided
to employ Mr. John Griffiths, of the
Bombay Art School, with some of his
students, to copy what was left. With a
break of three years, this grant was re-
newed till 1885, after which the publica-
tion of the results was urged, but delayed ;
and again, out of 335 copies, 163 were
destroyed and others damaged by a fire in
South Kensington Museum, where they

had been placed. Mr. Griffiths subse-
quently edited for Government a selection
of the results, — ' The Paintings in the
Buddhist Cave Temples at Ajanta,' 1896,
in two large folio volumes containing 156
plates, besides illustrations in the text.

A somewhat detailed account of the
paintings was first published in ' Notes
on the Bauddha Rock-Temples of Ajanta :
their Paintings and Sculptures' by the
editor (Bombay, 1879), which was repro-
duced with some trifling verbal changes,
in the ' Bombay Gazetteer of Khandesh'
(1880), pp. 496-574; and appeared again,
rearranged, in the 'Aurangabad Gazetteer'
(1884), pp. 430-506. See also 'Cave
Temples,' pp. 284-288, 291, 306-307,
310-315, and 326-336, and plates 29-43.

198

BUDDHIST ARCHITECTURE.

BOOK I.

sect. On the whole, they are ppurer and simpler than the latest
at Ajanta, though most probably of a slightly earlier age.

The plan of one has already been given (p. 182), but it
is neither so large nor architecturally so important as the
great vihara, shown in plan, Woodcut No. 113. Its great

"3.

Great Vihara Cave at Bagh. (From a Plan by Dr. Impey.)

hall is about 96 ft. square, and would at AjantS. rank as a
twenty-eight pillared cave, like No. 4 there, but inside this are
eight pillars ranged octagonally ; and at a later age, apparently
in consequence of some failure of the roof, the damaged portion
was hewn out, making the central area higher than the rest
of the hall, and four structural pillars — shaded lighter — were
introduced. The architraves forming the inner sides of the
octagon on the roof are carved with a double row of chaitya
window ornaments.

The jala connected with this vihara measures 94 ft. by
44 ft., and the two are joined together by a verandah measuring
220 ft. in length, adorned by twenty free-standing pillars. At
one time the whole of the back wall of this gallery and the
inner walls of the vihara were adorned with a series of frescoes,
equalling in beauty and in interest those of Ajanta. As in
those at Ajanta, the uninitiated would fail to trace among
them any symptoms of Buddhism as generally understood.
The principal subjects are processions on horseback, or on
elephants. In the latter the number of women exceeds that of
the men. Dancing and love-making are, as usual, prominently
introduced, and only one small picture, containing two men,
can be said to be appropriated to worship.

CHAP. VI.

SALSETTE.

199

With one exception, no man or woman has any covering
on their heads, and the men generally have the hair cropped
short, and with only very small moustaches on the face. Some
half-dozen are as dark as the Hindus of the present day.
The rest are very much fairer, many as fair as Spaniards, and
nearly all wear coloured dresses.1

We are not at present in a position to say, and may not
for a long time be able to feel sure, who the races are that
are represented in these frescoes or in those at Ajanta. Certain
of the figures are doubtless imaginary superhuman beings —
Rakshasas, Yakshas, and the like, and the scenes are more
or less ideal. The style of art, especially at Bagh, is very
similar to that of Persia at about the same date.

So far as the materials yet available indicate, the earliest
of this group of caves could not well have been commenced
much before A.D. 500 ; the date of the latest, if our chronology
is correct, could not well be carried down much beyond 600,
but a complete survey of them is required before we can decide
with confidence.

SALSETTE.

One of the most extensive of all the groups of Indian caves
is that generally known as the Kanheri Caves on the Island of
Salsette between Bombay and Thana. The great chaitya cave
there, as mentioned above, is only a bad copy of the Karle
cave, and was excavated in the end of the 2nd century, and none
of the viharas seem to be much earlier. It may have been be-
cause it was an island that it remained undisturbed by the
troubles of the mainland, and that the practice of excavating
caves lasted longer here than in any series above described.
Be this as it may, the caves here go straggling on till they fade
by almost imperceptible degrees into those of the Hindu religion.
The Hindu caves of Montpezir or Mandape^var and Joge^var,
and other Buddhist caves at Magathana and Kondivtd, are so
like them, and the change takes place so gradually, that it is
sometimes difficult to draw the line between the two religions.

Although, therefore, we have not at Salsette any viharas
that can compare with those of Nasik, Ajanta, or Bagh, yet
because they range from the 2nd century to far into the gih, and
fade so gradually into the next phase, are they worthy of con-
siderable attention.

As these caves are so near Bombay and Bassein, and so
easily accessible, they early attracted attention, and were

1 It is much to be regretted that no
attempt has yet been made to secure
faithful copies of these interesting wall-
paintings, which are fast perishing.

Some notes respecting them are given
in ' Notes on the Bauddha Rock-Temples
of Ajanta,' etc. (Bombay, 1879) pp. 94,
95-

200

BUDDHIST ARCHITECTURE.

BOOK I.

described by Portuguese visitors of the i6th century and by
numerous travellers during the iSth.1 Daniell's assistants made
a large number of drawings for him in 1795-1796, that were
never published. Careful measured drawings were made of
all of them by the brothers West in 1853-1859; but except
the inscriptions and an account of excavations at cave 12,
scarcely any of their work was published.2

A plan of one has already been given (Woodcut No. 101).
It is a two-storeyed vihara, and one of the finest here, though
it would not be considered remarkable anywhere else. Another,
of which a representation is given in my ' Illustrations of the
Rock-cut Temples,' plate 14, represents Avalokitej-wara with
eleven heads — the only instance I know of in India, though
it is common in Tibet in modern times.3 The others are
generally cells, though a monograph of these caves would be
a most valuable addition to our stock of knowledge of the
development of Mahayana mythology, which is largely illus-
trated in their sculptures. Traces of painting have also been
found in some of them.

DHAMNAR AND KHOLVI.

There are no viharas at either of these places, which can
at all compare, either in dimensions or in interest, with those
already described. The largest, at Dhamnar, is that already
given in combination with the chaitya, Woodcut No. 86, p. 165,
and, though important, is evidently transitional to another state
of matters. Next to this is one called the ' Great Kacheri ' ; but
it is only a six-celled vihara, with a hall about 25 ft. square,
encumbered by four pillars on its floor ; and near the chaitya
above alluded to is a similar hall, but smaller and without
cells. At Kholvi * there is nothing that can correctly be called
a vihara at all. There is, indeed, one large hall, called ' Bhfm's
house,' measuring 42 ft. by 22 ft; but it has no cells, and is much
more like what would be called a .Sala at Bagh than a vihara.
The others are mere cells, of no architectural importance.5

1 Niebuhr, 'Voyage en Arabic et
d'autres pays circonvoisins,' 1776-1780.
Most of the plates referring to these
caves were reproduced by Langles in his
1 Monuments d'Hindostan,' vol. ii. plates
77, et seqq.

2 Plates 53 and 54 in the volume upon
the ' Cave Temples ' are from this collec-
tion, which is now in the editor's posses-
sion, and might be published. For some
account of the caves, see ' Cave Temples,'
pp. 348-360, and plates ; also ' Archseo-
logical Survey of Western India,' vol. iv.
pp. 70, 71, and plates 42, 43.

1 'Cave Temples,' plate 55, fig. 2,
and p. 357 ; Schlagentweit, ' Buddh-

ismus in Thibet,' plate 3 ; Griinwedel,
' Mythologie des Buddhism in Tibet u.
Mongolei,' S. 65.

4 The Kholvi group is situated more
than 60 miles north of Ujjain, that of
Dhamnar about 22 further north, and
deeper into the Central Indian jungles.

5 Plans of these caves, with descrip-
tions and some architectural details, will
be found in Gen. Cunningham's ' Archceo-
logical Reports,' vol. ii. pp. 270-288,
plates 77-84. But till those of Kholvi are
photographed we shall not be able to
speak positively regarding them ; the
General's drawings are on too small a
scale for that purpose.

CHAP. VI.

ELURA.

2OI

ELCRA.
At Elvira there are numerous viharas at the extreme south

of the group and attached to the Vi^wakarma, or the great

202

BUDDHIST ARCHITECTURE.

BOOK I.

chaitya above described (p. 160). Like it, however, they are
all modern, but on that very account interesting, as showing
more clearly than elsewhere the steps by which Buddhist cave-
architecture faded, through devotion to the Mahayana idolatry,
into something very like that of the Hindus. Every step of
the process can be clearly traced here, though the precise
date at which the change took place cannot yet be fixed with
certainty. The caves at the extremity of the series, as will be
seen from the Woodcut No. 1 14 are very much ruined.

The great vihara, which is also evidently contemporary
with the chaitya, is known as the Maharwara (No. 5), seen
near the left in Woodcut No. 114, and, as will appear from
the plan (Woodcut No. 115), it differs considerably from any
of those illustrated above. Its dimensions are considerable,

being 1 10 ft. in depth
by 70 ft. across the
central recesses, its
great defect being
the lowness of its
roof. Its form, too,
is exceptional. It
looks more like a
flat - roofed chaitya,
with its three aisles,
than an ordinary
vihara ; and such it
possibly was intended
to be, and, if so, it
is curious to observe
that at Bedsa (Wood-
cut No. 63, p. 138) we
had one of the earliest
complete viharas,
looking like a chaitya
in plan ; and here we
have one of the latest,
showing the same
confusion of ideas :
a thing very common
in architectural his-
tory, where a new
style or a new ar-
with copying some

"5-

Plan of Maharwara Cave, Elura,
Scale 50 ft, to i in.

rangement generally hampers itself

incongruous form, which it casts off during its vigorous man-
hood, but to which it returns in its decrepitude — a sure sign
that it is passing away. But the form of this cave is, perhaps,

CHAP. VI. ELURA. 203

otherwise to be explained by the probability that, like the
so-called Darbar cave at Kanheri, this was a refectory, which
may account for its arrangements.

Close to the VLrwakarma (No. 10), is a small and very pretty
vihara (No. 8), in which the sanctuary stands free, with a passage
all round it, as in some of the Buddhist caves at Aurangabad and
in Saiva caves further on ; and the appearance of the Mahayana
warders on each side of the door would lead one rather to
expect an image of Siva inside than the Buddha which actually
occupies it. The details, however, of its architecture are the
same as in the great cave.

Communicating with this one is a small square vihara
(No. 7), the roof of which has been supported by four pillars
of the same detail as in the Dukhya-garh, which is the cave
next the chaitya on the north ; but though surrounded by cells
it has no sanctuary or images.

Higher up the hill than these are two others (Nos. 6 and 9),
containing numerous cells, and one with a very handsome hall,
the outer half of which has unfortunately fallen in ; enough,
however, remains to show not only its plan, but all the details,
which very much resemble those of the last group of viharas at
Ajanta.

In the sanctuaries of both of these caves are figures of
Buddhas sitting with their feet down. On each side of the
image in the principal one are nine figures of Buddhas, or rather
Bodhisattwas, seated cross-legged, and below them three and
three figures, some cross-legged, and others standing, probably
devotees, and — one of them a female — the Tara of later
Buddhism. Neither of these caves have been entirely finished.

There is still another group of these small viharas (Nos. 2,
3, 4), further to the south, at the right in Woodcut 1 14, called
the Dherwara or ' low caste's ' quarter.1 The first is square, with
twelve pillars on the same plan as those at Ajanta, though the
pillars are of the cushion form of Elephanta and the Maharwara,
but the capitals are much better formed than in the last example,
and more ornamented ; the lateral galleries here contain figures
of Buddha, all like the one in the sanctuary, sitting with their
feet down, and there are only two cells on each side of the
sanctuary. The next cave is similar in plan, though the detail
is more like that of the Vuwakarma. There are eleven cells,
and in the sanctuary Buddha sitting with the feet down ; it
never has been finished, and is now much ruined. The last
is a small plain vihara with cells, but with two pillars in front
of the shrine and cells, and much ruined.

1 ' Cave Temples of India,' plates 57 and 58. Possibly ' Dherwara' is a corruption
of Therawara or 'ascetics' quarter.'

204 BUDDHIST ARCHITECTURE. BOOK I.

The whole of the caves in this group resemble one another
so much in detail and execution that it is difficult to make out
any succession among them, and it is probable that they were
all excavated within the same century as the Vmvakarma.

The two temples, north of the Vi^wakarma, are particularly
interesting to the antiquarian, as pointing out the successive
steps by which the Buddhistical caves merged into the forms of
the Brahmanical.

The first is No. 1 1, the Don Tal or Dukhya-garh, a Buddhist
vihara of which the lower storey was long completely silted up
— hence its name of the ' two storeyed ' ; but in 1877 the ground
floor was excavated, consisting of a verandah 90 ft. in length,
with a shrine and the commencement of two cells. Most of its
details are so similar to those above described that it may be
assumed to be, most probably, of the same age. It is strictly
Buddhist in all its details, and shows no more tendency
towards Brahmanism than what was pointed out in speaking
of the Vijwakarma. All its three storeys have been left
unfinished.

The next, or Tin Tal (No. 12), is very similar to the last in
arrangement, but on a greatly enlarged scale, and its numerous
sculptures are all Buddhist, though deviating from the usual
forms by a large representation of the female divinities of the
Mahayana pantheon. Of its class, this cave is one of the most
important and interesting in India ; nowhere else do we find a
three-storeyed cave temple — adapted for worship rather than as
a monastery — executed with the same consistency of design and
the like magnificence, so that there is a grandeur and propriety
in its conception that it would be difficult to surpass in cave
architecture. Its sculptures are of extreme interest, and the
delineation or photographing of the whole would be of the
greatest value to the antiquary as illustrative of Buddhist
iconography.1

It is not easy, in the present state of our knowledge, to
determine whether the Elura Buddhist group is later or earlier
than those of Dhamnar and Kholvi. It is certainly finer than
either, and conforms more closely with the traditions of the
style in its palmiest days ; but that may be owing to local
circumstances, of which we have no precise knowledge. The
manner, however, in which it fades into the Hindu group is in
itself sufficient to prove how late it is. If we take A.D. 600 as
the medium date for the Vmvakarma and its surroundings, and
A.D. 750 as a time when Buddhism began to wane in Western
India, we shall probably not err to any great extent ; but we

1 'Cave Temples,' pp. 381-384, and plates 64, 65 ; ' Archaeological Survey of
Western India,' vol. v. pp. 16-22, and plates 14 fig. 2 ; 18 fig. 3 ; 19 ; and 20.

CHAP. VI. AURANGABAD, KUDA, AND OTHER CAVES. 205

must wait for some inscriptions or more precise data before
attempting to speak with precision on the subject.

A great deal more requires to be done before this great
cartoon can be filled up with anything like completeness ; but
in the meanwhile it is satisfactory to know that in these
" rock-cut temples," eked out by the few structural examples
that exist, we have a complete history of the arts and liturgies
of the Buddhists for the thousand years that ranged from
B.C. 250 to A.D. 750; and that, when any one with zeal and
intelligence enough for the purpose will devote himself to the
task, he will be able to give us a more vivid and authentic
account of this remarkable form of worship than can be gathered
from the books known to us.

AURANGABAD, KUDA, AND OTHER CAVES.

Besides the caves at Junnar, already noticed, there is a small
but important series near Aurangabad, forming three small
groups in the scarp of a range of hills to the north of the city,
and consisting of twelve or thirteen excavations. The third or
most easterly group consists of three unfinished caves without
sculpture ; but, except a chaitya cave in the first group — nearly
half of which has fallen away — most of the others are very rich
in sculpture, and the pillars are elaborately carved in the style
of the later Ajanta viharas. Two in the first group, and two
larger in the second, are planned on a purely Hindu arrange-
ment, there being a passage for circumambulation quite round
the shrine, with cells off this. The attendant figures in the
shrines, the dwarpals at the entrances, and numerous female
figures sculptured in these caves, indicate that they belonged to
a Mahayana or ritualistic sect of Buddhists. No inscription has
been found to help us in determining their date, but their
whole style indicates that they can hardly be placed earlier
than the 7th century of our era, and perhaps towards the end
of it. Since, however, they have been described and illustrated,
with numerous examples of their richly carved pillars and
remarkable sculptures, in the third volume of the 'Archaeo-
logical Survey of Western India Reports,' reference may be
made to that volume for further details.

The Kuda caves in the Konkan, south from Bombay, form
a group of twenty-two excavations, mostly plain and of small
dimensions ; but though they are rich in inscriptions, these
afford us no key to the date of the caves further than that
the alphabet of the inscriptions is closely allied to that used
in Karle, Nasik, and Kanheri inscriptions of Andhrabhritya

206

BUDDHIST ARCHITECTURE.

BOOK I.

times.1 The architectural style is plain, and the sculptures
comparatively few ; and as they have been described else-
where they need not occupy us here. And for the twenty-eight
excavations at Mahad, about sixty at Karhad, and other smaller
groups in the Konkan and Dekhan, as they present no special
features, we must also refer to the detailed accounts in the
same works.

No important Buddhist remains have yet been discovered
in the south of the peninsula, and the rapid manner in which
Hiuen Tsiang passes through these countries, and the slight
mention he makes of Buddhist establishments render it some-
what uncertain what important establishments belonging to
that sect then existed in Dravida-de^a. Yet we gather from
him that Buddhists as well as Jains must, at one time, have
been very numerous there, though the former had probably
lost much of their influence by the 7th century. Their viharas
and temples, being usually of brick, would become the spoil

of neighbouring towns and
villages for building materials
wherever the Buddhists ceased
to frequent them, and all traces
of them have long since dis-
appeared.

Negapattam, on the coast,
170 miles south from Madras,
was the great port of Tanjor
and the Kavert delta, and was
noted as a seat of Buddhist
worship. We learn that a
Buddhist temple here was
endowed by Rajendra Chola I.
in 1006 A.D., and that it had
been built by one "Chula-
manavaram King of Kidaram
or Kataha" — possibly in south
Burma or Siam. And in a
later grant Kulottunga Chola
I., in 1090, made gifts to at
»"""•"• least two Buddhist temples

Ancient Buddhist Tower at Negapattam. i 1-1 , T>

(From a sketch by Sir Walter Elliot.) nere> Whilst a .Burmese in-

scription of the i$th century

mentions a visit to Negapattam by some Buddhist priests from
Pegu.

116.

1 ' Inscriptions from the Cave Temples
of Western India,' etc. pp. 3-22 ; ' Indian
Antiquary,' vol. vii. pp. 253-257 ; ' Cave
Temples,' pp. 204-209, and plate 5, fig. i,

and plate 7, fig. I ; 'Archaeological Survey
of Western India,' vol. iv. pp. 12-18,
and plate 8, and the inscriptions at pp.
84-88,

CHAP. VI. AURANGABAD, KUDA, AND OTHER CAVES. 207

One remarkable fragment survived till 1867, about a mile
north-west of Negapattam,1 in a ruined brick tower of three
storeys about 70 feet high, locally known as Puduveli-gopura,
and to Europeans as the " China pagoda." The interior was
open to the top, but showed marks of a floor about 20 ft. from
the ground. The brickwork was described as good and closely
fitted together without cement, and the storeys were marked off
by outside cornices of stepped brickwork, with an opening for a
door or window in the middle of each side. Its general appear-
ance in 1846 is presented by the accompanying woodcut (No.
1 1 6). This structure had probably formed part of one of the
temples mentioned in the nth century. With the consent
of the Madras Government, it was pulled down by the Jesuit
priests who had been expelled from the French territory of
Pondicherry in 1845, and in its demolition several images of
Buddha were found — the pedestal of one of them bearing an
old Tamil inscription.2

1 Ante, p. 33.

* In 1859 the Jesuit missionaries asked
permission to pull it down and use the
materials for their college, and the
district engineer, reporting upon it as
not deserving the name of an ancient
monument, recommended that an esti-
mate of Rs. 400, sanctioned for its con-
servation, should be cancelled, and the
tower demolished. Sir W. Elliot opposed

this, and the building would have been
preserved, but the Jesuit priests threw
obstructions in the way, and nothing was
done. In 1867 they presented a fresh
petition for permission to demolish it,
which was granted. — 'Indian Antiquary,'
vol. vii. pp. 224 et seqq. vol xii. p. 311,
and vol. xxii. p. 45. The cut is taken
from Yule's 'Marco Polo' (3rd ed). vol. ii.
p. 326.

NOTE. (Ante, p. 175). — Among the sculptures mentioned in the 'Journal of the
Royal Asiatic Society, 1907,' p. 997,
as discovered in the excavations
made in recent years at Sarnath,
besides a fine capital and part of
the shaft of an inscribed Aroka Lat,
was an interesting flat capital which,
though differing from the usual
classic forms, bears a distinct re-
semblance to the capitals of the
pilasters of the temple of Apollo
Didymseos at Miletos. Conf. Durm,
'Die Baustile des Hdbuches. der
Architectur,' Bd. i. S. 189; Texier
and Pullan, ' Principal Ruins of Asia
Minor,' plates 6-8. It is of the
same style as the larger example
previously discovered by Dr. Waddell
at Patna — the ancient Pataliputra.

The abacus of the latter is 49 in. II7. Capital found at Patna.

long and 33 J in. in height, and is
represented in the accompanying
cut, No. 1 17. The Sarnath one is only 13 in. high and, when entire, was about 25 in.

ao8 BUDDHIST ARCHITECTURE. BOOK I.

across the top, having its frieze sculptured with a horseman at the gallop, parts of a
large plant being shown as beyond the horse (represented in 'Journal of the Royal
Asiatic Society' 1907, plate 3, fig. 4). In the Patna example a honeysuckle or similar
plant occupies the area, and the whole form is more elegant and classical in feeling.
(Waddell, ' Report on Excavations at Pataliputra,' p. 40 and plate 2.) Both capitals
belong to the same order and must be of about the same age; but they differ so
essentially from anything we know to be of the age of Aroka, and are so refined and
classical in taste that, viewed in connection with the remains found at Jamalgarhi and
elsewhere, they seem, more probably, to belong to the period about the commencement
of our era, when Hellenic influence in architecture was strongest. — fn^ra, p. 215.

118. Capital in Side Chapel of Cave xix., at
Ajanta.

CHAP. VII. GANDHARA MONASTERIES. 209

CHAPTER VII.

GANDHARA MONASTERIES.

CONTENTS.

Monasteries at Jamalgarhi,— Takht-i-Bahai and Shah-Dheri,— Greek

influence.

FEW of the later discoveries in India have been more fruitful
of important results for the elucidation of the archaeology
of India than those obtained from the excavations of ruined
monasteries in the neighbourhood of Peshawar. They supply
us with the materials for settling not only the question of the
amount of influence classical art exercised on that of India,
but also for solving many problems of Buddhist archaeology
and art.

As mentioned above, it is from their coins, and from them
only, that the names of most of the kings of Baktria and their
successors have been recovered ; but we have not yet found
a vestige of a building that can be said to have been erected
by them or in their age, nor one piece of sculpture that, so far
as we now know, could have been executed before their down-
fall, about B.C. 1 30. This, however, may be owing to the fact
that Baktria proper has long been inhabited by fanatic Moslims,
who destroy any representations of the human form they meet
with, and no excavations for hidden examples have yet been
undertaken in their country ; while it is still uncertain how far
the influence of the true Baktrians extended eastward, and
whether, in fact, they ever really possessed the valley of
Peshawar, where so many of the sculptures have been found.
No one, in fact, suspected their existence in our own territory
till Lieutenants Lumsden and Stokes, in 1852, partially explored
the half-buried monastery at Jamalgarhi, which had been
discovered by General Cunningham in 1848. It is situated
about 36 miles north-east from Peshawar, and from it these
officers excavated a considerable number of sculptures, which
afterwards came into the possession of the Hon. E. Clive Bayley.
He published a short account of them in the ' Journal of the
Bengal Asiatic Society,' in 1853, and brought the collection

VOL. I. O

210

BUDDHIST ARCHITECTURE.

BOOK I.

itself to this country. Unfortunately, they were utterly
destroyed in the disastrous fire that occurred in December 1866
at the Crystal Palace, where they were being exhibited, and
this before they had been photographed, or any serious attempt
made to compare them with other sculptures.

Since that time other collections have been dug out of
another monastery eight miles further westward, at Takht-t-
Bahai, and by Dr. Bellew at a third locality, 10 miles south-
ward, called Shahr-i-Bahlol, some of which have found their
way to this country. In 1874 Dr. Leitner brought home an
extensive collection, principally from Takht-i-Bahai, which have
now gone to Berlin.1 Again, since the extension of British rule
over the North -West Frontier Province during the last few
years, numerous fresh sites have been discovered and excavated.2
But since they were first discovered, numerous sites have been
rifled, at least once ; " mostly without definite plan and with
motives not altogether disinterested. The history of these
depredations would be long and lamentable — from the exploit
of the Colonel who, as Cunningham tells us, carried off the
statues from Jamalgarhion twelve camels, to those ' irresponsible
diggings,' the ravages of which in the recently opened district
of Swat, Sir H. Deane so justly deplores."3 Of the earlier
official excavations, the worst thing is that they were so
unsystematically carried on that it is impossible to ascertain
where hundreds of the sculptures now in the Lahor Museum
came from, and in almost no instance can the position of any
one piece of sculpture be fixed with anything like certainty.4

1 Quite recently the splendid collection
of Mr. M. L. Dames, has also gone to
Berlin, because the British Museum
would not, or could not, purchase it.

2 As an example of how such remains
were too often treated, we learn that in
1896 the contractors of the Military
Works Department, to obtain readily
stones for a culvert near Chakdarra,
destroyed a little vihara of great archaeo-
logical interest. — Foucher, ' L'Art Greco-
Bouddhique du Gandhara,' tome i.
p. 108.

3 Foucher, loc. cit. tome i. p. 14 ;
Cunningham, ' Archaeological Survey
Report,' vol. v. p. 46; 'Journal of the
Royal Asiatic Society,' 1896, p. 664.
M. Foucher adds: "Hardly anywhere
were the excavators at the trouble to
unearth the buildings to the foundations,
in order to determine their plans or
restore the scheme of their decoration ;
their only care was to lay hands on the
sculptures ; and they took no trouble to
put aside or protect pieces that were

thought too heavy or too fragmentary for
removal. In many places, headless
trunks and mutilated reliefs strew the
clearings and testify to the ignorance and
brutality with which the excavations were
conducted, if we may use the word — for it
is somewhat ironical to employ that term
— since they were mostly left, without
European supervision, to the direction
of some native subaltern, or even to the
discretion of the coolies of the nearest
village."

4 The mode in which the excavations
were conducted by Government was to
send out a party of sappers in the cold
weather to dig, but the officer in charge
of the party was the subaltern who
happened to be in command of the com-
pany at the time. A new officer was
consequently appointed every year, and
no one was ever selected because he had
any experience in such matters or any
taste for such pursuits ; and the result
was, as might be expected, painfully
disappointing.

CHAP. VII.

GANDHARA MONASTERIES.

The sculptures discovered have been partly collected in the
Indian Museums — those of Lahor and Calcutta having between
them 1700 or more specimens of this class of art, and small
collections were sent to Madras, Rangoon, and Bombay in 1884 ;
there is a large and fine collection at Berlin, and over a hundred
pieces at the Louvre, whilst the small collection in the British
Museum is due almost entirely to private donors.1

These remarkable sculptures have attracted more attention on
the continent than in India or England, and the encouragement
given by continental governments has conduced in a marked
degree to their study and the solution of some of the problems
they present.2

The essential elements of a Buddhist monastery were the
stupa and sangharama or quarters for the monks ; the vihara
proper, or shrines for the images, might be arranged to form a
court round the stupa, or they might surround a separate court,
between the stupa and sangharama — and, as in the Mahayana
schools the images were very numerous, the pantheon must
often have overflowed the capacity of the stupa courts.

The following plans (Woodcuts Nos. 119, 120), of the two
principal monasteries which have been excavated in the vicinity
of Peshawar, will explain their arrangements in so far as they
have been made out. As will be seen at a glance, they are
very similar to each other, or at least consist of the same parts.
First a circular or square court, AA, surrounded by cells,
evidently intended to contain images, though none were found
in situ. In the centre of each stands a circular or square
platform, being the basement of a stupa, approached by steps.3
The circular one at Jamalgarhi was 22 ft in diameter and adorned
with cross-legged, conventional, seated figures of Buddha, the
smaller one, at Takht-i-Bahai, was 15 ft. square and ornamented
by two rows of pilasters one over the other. Beyond this is an
oblong court, BB, called the "pantheon," from the number of
images, small models of topes, and other votive offerings of all
sorts, that were found in it. It, like the last court, is surrounded
by niches for images, and was the " vihara " properly so called.

1 Foucher, ' L'Art Greco- Bouddhique
du Gandhara,' tome i. pp. 23-30.

2 In 1893 Prof. A. Griinwedel, in a
handbook of the Gandhara sculptures at
Berlin, discussed the origin of ' Buddhist
Art in India,' elucidating the subject
from the bas - reliefs in the Royal
Museum there. An enlarged edition was
issued in 1900, and an English transla-
tion revised and greatly extended was
published (by Quaritch) in 1901. Dr.

A. Toucher's Mission to the North-West
Frontier, etc., has already been referred
to, ante, p. 89.

3 All the stupas of the Panjab and
Gandhara had steps up to the level of the
basement, and usually on the side facing
the monastery ; thus, at Jamalgarhi they
were a little to the east of south, whilst
at Takht-i-Bahai, they were on the north
side. Some had steps on two, and others
on all four sides.

212

BUDDHIST ARCHITECTURE.

BOOK I.

Beyond this again was the sangharama or residence, CC, with the
usual residential cells. At Takht-i-Bahai there is, at the north-

N

119, Plan of Monastery at Jamalgarhi. Scale 50 ft, to i in.

S

Plan of Monastery at Takht-i-Bahai, Scale 50 ft, to i in,

CHAP. VII.

GANDHARA MONASTERIES.

213

west corner, a court, D, 51 ft. square, surrounded by a high wall
with only one door leading into it. A corresponding court and
of similar size exists at Jamalgarh! ; but it lies about 30 yards to
the east, so that it could not be included in the woodcut. This
has been identified by M. Foucher with the Service Hall, so often
referred to in Buddhist literature, where all the Bhikshus or
" members of the order " met privately on the nights of new and
full moon to read their rules and go through their confessional
forms, and where they met for all their more solemn purposes —
as ordination, excommunication, and the like ; and it was often
used also as a refectory. This was known as the Upasthanajala
or Meeting Hall. If this was the purpose of these buildings,
which seems very probable, they must have been roofed in wood.1

When we attempt to compare these plans with those of
rock-cut examples in India, we at once perceive the difficulty
of comparing structural with rock-cut examples. The monastery
or residential parts are the only ones readily recognised. The
pantheon does not apparently exist at Ajanta, nor is anything
analogous to it attached to other series of caves. A group
of small rock-cut memorial dagabas exists outside the caves at
Bhaja, and a much more extensive one of structural topes
formed the cemetery at Kanheri, and similar groups may have
existed elsewhere : but these are nowise analogous to the above.
Numbers of small models of topes and votive offerings are found
in the neighbourhood of all Buddhist establishments, and were
originally no doubt deposited in some such place as this. The
circular or square base of the stupa marks the place which the
chaitya occupies in all the rock-cut chaitya halls.

One of the most remarkable ornamental features that adorn
this monastery is a series of bas-reliefs that adorn the front
of the steps of the stairs leading from the so-called pantheon
or vihara to the circular court at Jamalgarhi. They are sixteen
in number, and each is carved with a bas - relief containing
twenty, thirty, or forty figures according to the subject.2
Among these the Vishvantara and Sama jatakas can easily
be recognised,3 and so may others when carefully examined.

1 Foucher, ' L'Art Greco-Bouddhique,'
tomei. pp. 162-163. It had been suggested
that this roofless hall might have been
a cemetery (Cunningham ' Archaeological
Reports,' vol. v. p. 32) ; and it was
pointed out that Turner in his ' Embassy
to Tibet' (p. 317), describes a similar
enclosure at Teshu-lumbu in which the
bodies of the deceased monks were ex-
posed to be devoured by the birds ; and
what happened there in 1800 might
possibly have been practised at Peshawar
at a much earlier age ; but that this was

not the purpose of the two enclosures
referred to is quite obvious.

2 These were removed by Gen.
Cunningham, and several are now in
the British Museum^' Journal of Indian
Art and Industry," vol. viii. p. 40, and
plates 23, 24 ; ' Ancient Monuments, etc.
of India,' plate 151 ; Cunningham, 'Arch-
aeological Survey Report,' vol. v. p. 199.

3 ' Tree and Serpent Worship,' plates
24 (fig. 3) and 36 (fig. l) ; and 'Journal
of the Royal Asiatic Society,' 1893,
p. 3'3-

214

BUDDHIST ARCHITECTURE.

BOOK I.

Besides these there are representations of the chase, processions,
dancing, and domestic scenes of various kinds. In fact such
a series of sixteen bas-reliefs, one over another, is hardly known
to exist anywhere else, but is here only an appropriate part
of an exuberance of sculptural ornamentation hardly to be
matched, as existing in so small a space, in any other building
of its class.

121. Corinthian Capital from Jamalgarhi, (From a Photograph.)

Corinthian Capital from Jamalgarhi. (From a Photograph.)1

The architecture of this monastery seems to have been of
singular richness. General Cunningham brought away a dozen
of capitals of the Corinthian order, and others exist in the
Lahor Museum. As will be seen from the last two illustrations
(Nos. 121, 122), they are unmistakably classical, but of a form
to which it is not at first sight easy to assign a date. They
are more Greek than Roman in the character of their foliage,

1 The modillion cornice, though placed on the lower capital in the photograph,
belongs in reality to another part of the building.

CHAP. VII.

GANDHARA MONASTERIES.

215

but more like Roman than Greek in the form of their volutes
and general design. Perhaps it would be correct to say they are
Indian copies or adaptations of classical capitals of the style
of the Christian Era.

Not one of these was found in situ, nor, apparently, one quite
entire, so that their use or position is not at first sight apparent.
Some of them were square, and it is consequently not difficult
to see they may have formed the caps of the antae on each
side of the cells, and are so represented in General Cunningham's
plate (15). If this is so, the circular ones must have been
placed on short circular pillars, one on each side, forming a
porch to the cells. One at least seems to have stood free —
like a stambha — and, as the General represents it (on plate 48),
may have carried a group of elephants on its head.

All these capitals were apparently originally richly gilt,
and most of them, as well as some of the best of the sculptures,
show traces of gilding,1 and, as others show traces of colour,
the effect of the whole must have been gorgeous in the extreme.
From the analogy of what we find in the caves at Ajanta and
Bagh, as well as elsewhere, there can be little doubt that fresco-
painting was also employed : but no gilding, as far as I know,
has been found in India, nor indeed, with one or two exceptions,
any analogue to the Corinthian capital.2 The capitals found in
India are either such as grew out of the necessities of their own
wooden construction, or were copied from bell-shaped forms we
are familiar with at Persepolis, where alone in Central Asia they
seem to have been carried out in stone ; 3 and they may have
been so employed down to the time of Alexander, if not later.
Certain it is, at all events, that this was the earliest form we
know of employed in lithic architecture in India, and the
one that retained its footing there certainly till after the
Christian Era, and also among the Gandhara sculptures to
a still later date.

In the decorative sculptures of these monasteries, archi-
tectural elements are largely employed in the representation
of buildings in which scenes are pourtrayed, and in pillars
separating the panels. These present forms of Perso- Indian
pillars employed side by side, sometimes on the same slab,
with columns having classical capitals and bases. The capitals
of the old Perso-Indian type have new forms given to them —
the animal figures being changed, whilst the pillars themselves
are placed on the backs of crouching figures with wings. It is
the same absurd composition as is found in Assyrian and even

1 ' Archaeological Reports,' vol. v.
pp. 49 and 196.

2 Ante, p. 207, note.

3 'The Palaces of Nineveh and Per-
sepolis Restored.' By the Author. Part
II. sect, i., et passim.

2l6

BUDDHIST ARCHITECTURE.

BOOK 1.

Lombard architecture, where pillars were placed on animals
and monsters ; and a similar practice was also long prevalent in
Dravidian architecture.1 Structurally the architecture of the age,
we may suppose, would share in the mixed character of these
sculptured representations. But the evidence may not be quite

decisive ; a stupa, such as
the best preserved at 'Alt
Masjid,for example, affords
but little aid in recovering
the style of temples or
other structures. What
we see represented in the
sculptures, together with
such structural fragments
as remain to us must be
our chief guides.

It is not difficult to
restore, approximately, the
front of the cells in these
monasteries, from the
numerous representations
of them found among the
ruins, where they are used
as conventional frames for
sculptures. It probably
was owing to the fact that
their fronts may have been
adorned with paintings
representing scenes from
the life of Buddha, or
emblems of various sorts,
that these miniature repre-
sentations of them were
used to convey the same
design in sculpture. These

123. Conventional Elevation of the Fa?ade of a gable - end shaped panels
Cell from Jamalgarhi. were fixed on four sides of

the domes of the smaller

stupas at least, and whilst they may present the general features
of the facades of the more highly decorated cells, it is not to be
supposed that any of them were so richly sculptured (Woodcut
No. I23).2

The form of the wooden framework which filled the upper

. 1 Fergusson, 'Ancient and Mediaeval

Architecture,' 3rd ed., pp. 188, 593, 594.

2 Conf. ' Buddhist Art in India '

(Eng. transl.), p. 156, fig. 107 : Foucher's
' L'Art Greco- Bouddhique du Gandhara,'
pp. 183-185, figs. 70-72.

CHAP. VII. GANDHARA MONASTERIES. 217

part of all the great windows of the chaitya halls, from the
earliest known examples, is also used for the same purpose
in these Gandhara monasteries. Few things among these
sculptures are more common than these semicircular frames,
filled with sculpture of the most varied design. They are in fact
the counterparts of what would have been carried out in painted
glass had they possessed such a material.

It is to be feared that it is hardly likely we shall now recover
one of these cells or chapels in so perfect a state as to feel sure
of its form and ornamentation. It would, however, be an
immense gain to our knowledge of the subject if one were found,
for it is hardly safe to depend on restorations made from
conventional representations.

Meanwhile there is one monument in India which — mutatis
mutandis — reproduces them with considerable exactness. The
small detached rath at Mamallapuram is both in plan and
dimensions, as well as in design, an almost exact reproduction
of these Jamalgarht cells. Its lower front is entirely open,
flanked by two detached pillars. Above this are two roofs,
with a narrow waist between them — somewhat differently
arranged it must be confessed, but still extremely similar.
In the Jamalgarht representations of these cells everything
is simplified to admit of the display of sculpture. At Mamalla-
puram all the architectural features are retained, but they are
still marvellously alike, so much so, that there seems no doubt
this little rath (Woodcut No. 185, page 329), with its circular
termination, is as exact a copy of what a Buddhist chaitya
hall was at the time it was carved, as that the great rath
(Woodcut No. 89, p. 172) is a correct reproduction of a Buddhist
vihara at the same period.

If this is so, these Gandhara sculptures and these raths
represent the chaitya hall of the Buddhists in a much more
complicated and elaborate form than we find it in the simple
but majestic examples at Karle, Nasik, or Ajanta. The
Jamalgarhi cells are not at all so modern as the rath at
Mamallapuram, but they are certainly approaching to it in form.1

General Cunningham dug out a small vihara at Shah-Dhert,
the ancient Taxila, which seems more ancient than these
Peshawar monasteries. As will be seen from the plan

1 One curious peculiarity of these
Gandhara sculptures is that they gener-
ally retain the sloping jamb on each
side of their openings. In India and in
a structural building this peculiarity
would certainly fix their age as anterior
to the Christian Era. In Gandhara it

is found chiefly in decorative sculpture,
but it seems also to have been occasion-
ally employed structurally, as in the
small vihara near Chakdarra fort in
Swat, destroyed in 1896 by the Military
Works Department. — Ante, p. 2IO,
note 2.

2l8

BUDDHIST ARCHITECTURE.

BOOK I.

(Woodcut No. 124), it is not only small in dimensions, but
simple in its arrangements — as simple, indeed, as any of those at

Katak or in the western
Ghats. Like some of
them it has a raised
bench, not, however,
divided into beds as
there, but more like a
continuous seat.1 It no
doubt, however, was
used for both purposes.
Its most remarkable
peculiarity, however, is
its Ionic order. As will
be seen, the bases of the

Plan of Ionic Monastery, Shah-Dheri. (From a pillars are of the USUal

Scale 5o ft. formj ancj as correct as
any that could be found
in Greece or Rome, from before the Christian Era to the age of
Constantine, and, though the capital is not fully
made out, there can be little doubt what was
intended (Woodcut No. 125). Twelve coins of
Azes were found close by, from which it may
be inferred the building was not of earlier date
than his age, or the 1st century B.C.,2 and there
is nothing in the architecture to militate against
this idea. It seems the oldest thing yet found
in this province.

Plan by General Cunningham.)
to i in.

The extraordinary classical character and
the beauty of the sculptures found in these
125, Gandhara monasteries is of such surpassing

ionic pniar, Shah- interest for the history of Indian art, that it
(From a Drawing by ls °f the utmost importance their age should
General Cunningham.) be determined, if it is possible to do so. At
present, sufficient materials do not exist in this
country to enable the general public to form a correct opinion
on any argument that may be brought forward on the subject ;
nor will they be in a position to do so till the Government can
be induced to bring some of them home. They are quite thrown
away where they now are ; here, they would hardly be surpassed
in interest by any recent discoveries of the same class. Quite

1 Compare the plan of the Ramgarh
cave, ' Indian Antiquary,' vol. ii. p. 245 ;
or Cunningham's ' Archaeological Re-
ports,' vol. xiii. p. 39 and plate to.

2 Assuming that his age has been
correctly ascertained : Conf. ' Buddhist
Art in India,' p. 78.

CHAP. VII. GANDHARA MONASTERIES. 219

recently, however, the solution of most of the questions relating
to these sculptures has been taken out of our hands by the
French mission to India to study the materials on the spot ;
and M. Foucher has arranged and illustrated them with such
acumen that his work becomes a standard one on the subject.1

Among Indian antiquaries different views have been held
as to the age of these sculptures, General Cunningham's opinion
was that the Baktrian Greeks carried with them into Asia the
principles of Grecian sculpture and the forms of Grecian
architecture, and either during their supremacy or after their
expulsion from Baktria established a school of classical art in
the Peshawar valley. This view further assumed that, when
Buddhism was established there under Kanishka and his
successors, it bloomed into that rich and varied development
we find exhibited in these Gandhara monasteries. He admitted,
however, that, as all the sculptures are Buddhist, the earliest
must be limited to the age of Kanishka, which he assumed
to be about B.C. 4O,2 and that they extend to A.D. 100, or
thereabouts.

Another theory equally admitted the presence of the
classical element, derived from the previous existence of the
Baktrian Greeks, but spread the development of the classical
feeling through Buddhist art over the whole period during
which it existed in the valley, or from the ist to the 7th or
8th century of our era, and ascribed its peculiar forms as
much, if not more, to constant communication with the West,
from the age of Augustus to that of Justinian, rather than
to the original seed planted there by the Baktrians.3

Neither view satisfactorily met the conditions, and, in 1890,*
Monsieur Emile Senart reviewed the question afresh and argued
that the priestly type of Buddha with the nimbus — a veritable
mark of the Grseco-Buddhist school — first appearing on the coins
of Kanishka, supplies one limit. And next, the regular appearance
of this same type among the Amaravati sculptures, testifies that,
when they were carved, the art of the north-west of India had
a fixed type, and had extended its influence to the south-east
of the peninsula ; and since the Andhra inscriptions engraved
on them cannot be assigned to a later date than the 2nd century

The English reader will find an | Mr. Fergusson, placing its epoch in A.D.

account of these sculptures generally
and of their origin in ' Buddhist Art
in India,' with 154 illustrations (London,

1901).

'Archaeological Reports,' vol. v. ,

78 ; and the other, ably supported by Dr.
j. F, Fleet and already referred to, throw-
ing it back to B.C. 57, ante, p. 29.

3 'Journal R. Institute Brit. Architects,'
3rd ser. vol. i., 1894, pp. 93ff.

Introduction, p. vi., and Appendix pp. 4 'Journal Asiatique' VHIe serie, tome

I93"I94- The date of Kanishka has long i xv. pp .139-163. See also the remarks of

been a matter of controversy, the principal Count Goblet D'Alviella, ' Ce que 1'Inde

views respecting his era, being that of doit a la Grece,' pp. 58, 630°.

220 BUDDHIST ARCHITECTURE. BOOK I.

of our era, it follows that " the zenith of the art and period of
its greatest expansion falls before the second half of the 2nd
century." Such an argument must have much weight in
deciding this question.1

About the beginning of our era Greek art had become
a matter of commerce and export, and Graeculi travelled in
all directions with their wares and models, ready to employ
their skill in the service of Gaul, Skythian, or Indian to
provide images for their pantheons by imitations from their
own patterns. They could also represent for their employers
the different classical orders of architecture, and would teach
their pupils how to carve them ; but, with or without models,
the copy would be modified to suit the Indian taste ; and so,
for the acanthus of the Greek capital, were introduced the
palms with which the Indian workmen were familiar ; and the
figures of Nike — we see in the Corinthian capitals of antae
in the temple of Augustus erected about A.D. 10 at Ancyra,
or in those of Priene, were reproduced in Gandhara as little
figures of Buddha.2 It is an imitation of Greek forms with
divergencies — not a copy — but the suggestion must have come
from those travelling Greek artists — probably lonians — who
were the agents by whom the Gandhara sculptures were
inspired, and Greek statuary was the model from which the
Mahayana pantheon was evolved.3

Further, it is at least approximately correct to state that no
statue of Buddha, in any of his conventional attitudes, has been
found in India executed earlier than about the Christian Era.
Those on the facade at Karle and in the western caves are
avowedly insertions of the 2nd or 3rd centuries or later. There
are none found at Bodh - Gaya, Bharaut, or Sanchi ; nor do I
know of any one in India that can be dated before the ist
century. In these Gandhara monasteries they are very frequent,
and of a type which in India would be assumed to be as late
as the 2nd or 3rd century ; some of them even later.

It is true Buddhist books tell us frequently of statues of
Buddha having been made at much earlier dates.4 But Indian
books have this fatal defect, that they represent facts and beliefs
at the time they were written, or acquired the forms in which
we now find them, without much reference to facts at the time
at which they are supposed to have happened. The actual
remains and the period to which they belong are our surest

1 Foucher, ' L'Art Greco-Bouddhique
du Gandhara,' tome i. pp. 4off.

2 'Buddhist Art in India,' p. 153.
There is also a capital at Siah, in Syria,
on which a bust is introduced, which

— De Vogue, ' Syrie Centrale,' plate 3.
3 It may be accepted that Greek art
furnished India with the images that
served for the beliefs. — Goblet D'Alviella,
' Ce que 1'Inde doit a la Grece,' p. 152.

may be as early as the Christian Era. 4 'Buddhist Art in India,' pp. ijiff.

CHAP. VII. GANDHARA MONASTERIES 221

guides, and we have now sufficient evidence to interpret these
sculptures with reasonable confidence.

Besides the figures of Buddha, there are a great number of
figures which have all got nimbuses or glories at the back of
their heads. All have the tilaka on their foreheads, as Buddha
has, and none have any kingly attributes, but all wear the same
ornaments and amulets. These are recognised as representa-
tions of the Bodhisattwa or of Bodhisattwas. Until Gautama
assumed Buddhahood, he was the Bodhisattwa of that age, and
as such is represented with necklaces and ornaments. But the
Mahayana school introduced many others into their iconography
— mythical beings who are ultimately to be manifested as the
Buddhas of future ages.1

A more important point than the mere presence of these
conventional figures of Buddha or of saints in these monasteries,
is their excessive reduplication ; to consecrate one was evidently,
as among the Jains, a work of religious merit.

In India, no building or cave is known with a date anterior
to, say, A.D. 100, in which more than one such figure is repre-
sented. Even at Amaravati they do not occur on the great rail
which was erected at latest about the beginning of the Christian
Era (ante, p. 122) but appear first on the basement, which was
constructed in the 2nd century ; and they occur in such cases as
Nos. 19 and 26 at Ajanta, and are numerous in the later caves
at Kanhert, Elura, and Aurangabad, none of which seem to be
earlier than A.D. 200, and most of them much later.

In the Gandhara monasteries they exist literally in hundreds
— on the base of the stupas, on the walls, and in the cells. The
latter is, indeed, the most remarkable peculiarity of any.
Among the Jains, it is the practice to surround the courts of
their temples with cells which are small shrines ; and here we
find also numerous small cells surrounding the courts of the
stupas all consecrated as shrines for images of Buddha and
saints, the monastery being quite separate from the structures
for worship. And further, here are even separate courts con-
structed for secondary stupas and numerous additional image
chapels. This wealth of imagery, however, is accounted for by
the fact that the Mahayana or Greater Translation was much
more prevalent in the north of India than in the peninsula, and
was considerably in advance of the Hinayana school of Central
India in all complications of ritual observances.

The few inscriptions found on Gandhara sculptures or on the
same sites, are dated in an unnamed era, and range from 78 to
384. One is dated in the twenty-sixth year of King Guduphara

1 Among these the chief are Maitreya — who is expected to appear first—
Avalokitejvara or Padmapani, Manjom, Samantabhadra, Vajrapani, etc.

222 BUDDHIST ARCHITECTURE. BOOK I.

or Gondophernes of Takshajila or Taxila, and in the year 103 of
this era. Now early Christian tradition mentions this king in
connection with the mission of St. Thomas, which would fall in
the ist century ; and the only Indian era we know of about that
time is the Samvat commencing B.C. 57, which makes the twenty-
sixth year of Guduphara coincident with A.D. 46, and places his
accession in A.D. 2021. This is quite in agreement, not only
with the tradition, but with the place assigned to the coinage
of Guduphara ; and we can hardly suppose that the other
inscriptions should be dated in a different era.1 Among these
there are known three or four of the ist century A.D., one each
of the 2nd and 3rd century, and that of Hashtnagar is dated
in 384, that is A.D. 32/.2 The last is not later than might be
expected, for when Fah Hian passed through Gandhara at the
commencement of the 5th century, the monasteries were still
in a flourishing condition. It was only a century later that the
Buddhists were persecuted by the Hunas under Toramana and
his son Mihirakula, and by that time the art had probably
declined ; these dates, however, are sufficient to substantiate the
conclusion that the Gandhara sculptures belong to the early
centuries of our era, and that its most flourishing period may
be assigned to about B.C. 50 to A.D. 200.

From what has been said above regarding the sculptures of
Bharaut and Sanchi, it appears evident that the Indians had a
school of art of their own before they knew anything of the arts
of the Western world ; but that native art seems to have had
very little influence on the arts of Gandhara. The Western arts,
on the contrary, acting through that country, seem to have had
considerable influence on those of India at periods about and
subsequent to the Christian Era. It seems at least almost
impossible to escape the conviction that the arts of Amaravati
and the later caves, say of the Andhra period, betray marked
evidence of Western influence ; and it seems that it is only
through Gandhara that it can have reached them.

So strongly marked is all this that it may become a subject
of an interesting investigation to enquire whether the Greeks
were not the first who taught the Indians idolatry. There is no
trace of images in the Vedas or in the laws of Manu, or any of
the older books of the Hindus. As repeatedly mentioned, there
is as little trace of any image of Buddha or Buddhist figures
being set up for worship much before the Christian Era. But
the earliest, the finest, and the most essentially classical figures
of Buddha are to be found in Gandhara, and, so far as we

1 Dr. Vogel ('Archaeological Survey j the end of Ajoka's reign, and Guduphara

Annual,' 1903-1904, pp. 259(7".), proposes
the Seleukidan era of B.C. 312, for these
dates, which would place Mogas about

not long after ! Conf. ' Journal of the
Royal Asiatic Society,' 1906, pp. 706-71 1.
2 ' Buddhist Art in India,' p. 84.

CHAP. VII.

GANDHARA MONASTERIES

at present know, of an earlier date there than any found in
India Proper.

Further, if there are traces of Christian doctrine in the
' Bhagavat Gita/ and of classical learning in other poetic works
of the Hindus, we now know at least where they may have
come from. In short, when we realise how strongly Hellenic
influence prevailed in Gandhara in the first few centuries after
Christ, and think how many thousands, it may be millions,
crossed the Indus, going eastward during that period and
through that country, we ought not to be surprised at any
amount of Western thought or art we may find in India.

In the meanwhile the question that bears most directly on
the subject now in hand is the enquiry, how far the undoubted
classical influence shown in these Gandhara sculptures is due to
the seed sown by the Baktrian Greeks during the existence of
their kingdom there, and how much to the direct influence of
Hellenic intercourse between the times of Augustus and
Aurelian ? Both, most probably, had a part in producing this
remarkable result ; but we have abundant evidence that the
latter was very much more important than the former cause,
and that about the commencement of the Christian Era the
civilisation of the West exercised an influence on the arts
and religion of the inhabitants of this part of India far greater
than was formerly suspected.

The question of the subjects of the sculptures is beyond the
scope of this work, and for this and their origin the reader must
be referred to the excellent work of Mons. A. Foucher, and to
the translation of Grunwedel's ' Buddhist Art in India.'

126. Footprints of Buddha. (From a bas-relief at Amaravatl.)

224

BUDDHIST ARCHITECTURE.

BOOK. I.

CHAPTER VIII.

CEYLON.

CONTENTS.
Introductory — Anuradhapura — Polonnaruwa

INTRODUCTORY.

IF the materials existed for writing it in anything like a complete
and satisfactory manner, there are few chapters in this history
that ought to be so interesting or instructive as that which
treats of the architecture of Ceylon. It alone, of all known
countries, contains a complete series of Buddhist monuments
extending from the time of Asoka to the present day, and in the
' Dipawansa ' and ' Mahawansa ' it possesses a history so detailed
and generally so credible, that the dates and purposes of the
earlier buildings can be ascertained with tolerable precision.
We know, indeed, that the early chronology is based on the
legend that the kingdom was founded at the date of Buddha's
Nirvana, which was placed in B.C. 543, and 236 years before
the Council held in the eighteenth year of Ajoka. But from
Indian data we must place the Council about B.C. 246, and
this reduces the dates dependent on the Nirvana by fully
sixty years. Lassen accepted the native chronology from the
accession of Dutthagamani, B.C. 161 ; but there are indications
in Indian history that the correction must be continued till
at least the 6th century A.D., after which the error perhaps
diminishes till it finally vanishes in the I2th century at the
time of the accession of Parakramabahu 1. in 1153 A.D.1 With
this rectification we may be satisfied for our purposes.

1 Lassen, ' Indische Alterthumskunde,'
Ed. 2, Bd. It, Ss. ioof., 266, 28;f.,
I225f. In the 27th chapter of the
' Mahawansa,' Dutthagamani's accession
is placed 146 years after Devanampiya
Tissa, who began to reign about B.C.
246. Max Mtiller (' History of Ancient

Sankrit Literature,' pp. 268-269 an<^
2986".) ascribed the Nirvana to B.C. 477 or
478; Dr. J. F. Fleet ('Journal Royal
Asiatic Society, 1906, 'pp. 984^) discussed
the evidence afresh, and ascribed it to B.C.
482. I am indebted to the latter for the
substance of the above statement.

CHAP. VIII.

CEYLON.

225

Besides the intrinsic interest of Sinhalese architecture, if it
were possible to compare this unbroken series with its ascer-
tained dates with the fragmentary groups on the continent of
India, its parallelisms might throw much light on many questions
that are obscure and uncertain, and the whole acquire a con-
sistency that is now only too evidently wanting.1

The survey of the Ceylon monuments owed its first inception
to Sir Hercules Robinson, then Governor, and in 1871 a series
of photographs of the principal remains at Anuradhapura and
Polonnaruwa was taken by the late Mr. Lawton, under the
personal direction of Mr J. G. Smither, Government Architect,
and supplemented a little later by a second series, by Captain
Hogg, R.E. These threw some light on the matter ; but
photographs alone — without plans or dimensions or descriptions
— are most deceptive guides, and, for the time, they added little
to our scientific knowledge. In 1873, however, under directions
from the Governor, Sir William H. Gregory, a survey was made
by Mr. Smither of what was then known at Anuradhapura,
and detailed plans and other architectural drawings of the more
important ruins were eventually completed in accordance with
recommendations by the late Mr. James Fergusson.2 In 1894,
however, Mr. Smither's most valuable work on Anuradhapura
was published ; the plates of drawings in it are excellent,
and the collotype photographs add materially to its interest and
value. If we had delineations of the other remains in Ceylon
excavated and surveyed since 1890, prepared and described
with like skill and accuracy, they would be of the very highest
value for the history of Sinhalese architecture.

Meanwhile, much progress has been made, for in 1884 the
Governor, Sir Arthur H. Gordon, now Lord Stanmore, intrusted
Mr. S. M. Burrows of the Civil Service, under the supervision

1 Sir Emerson Tennent's book, pub-
lished in 1 859, was one of the best works
on the subject. He had, however, no
special qualifications for the task, beyond
what were to be expected from any well-
educated gentleman of talent, and his
description of the buildings is only meant
for popular reading.

The two papers by Captain Chapman,
in the third volume of the 'Transac-
tions' (1832), and thirteenth volume of
the 'Journal of the Asiatic Society,'
were for long the best account of the
ruins of Anuradhapura, and beyond these
a few occasional notices were nearly all
the printed matter we had to depend
upon. Of late several ' Guide Books '
have appeared : Burrows's ' Buried Cities

VOL. I.

of Ceylon' (4th ed. 1906) ; H. W. Cave's
' Ruined Cities of Ceylon' (8vo. ed. 1900);
and J. Still's ' Ancient Capitals of Ceylon '
(1907).

2 Nothing was generally known in
England of this survey till 1888, when
a paper by Mr. John Capper on the
Anuradhapura dagabas appeared in the
'Journal of the Royal Asiatic Society,'
N.S., vol. xx., pp. 165-180. His son,
Mr. G. Capper, had been employed
in the survey under Mr. Smither,
F. R. I. B. A. , whose work — ' Architec-
tural Remains : Anuradhapura, Ceylon,
comprising the dagabas and certain
other ancient ruined structures with fifty-
seven plates' (Atlas fol. ) — contains the
results of the surveys made in 1873-77.

226

BUDDHIST ARCHITECTURE.

BOOK I.

of the late Mr. R. W. levers, with the work of survey and
excavation at Anuradhapura, including the partial conservation
of the crumbling remains, and clearing them of vegetation.
This he carried on for a year and a half.1 In 1890 a general
survey was organised, and Mr. H. C. P. Bell of the same
service was appointed Archaeological Commissioner. He has
supplied Government from time to time with progress reports
on his work, especially at Anuradhapura ; 2 but no systematic
account with architectural drawings has yet been published
to which the student can apply for connected and detailed
information.

One of the most striking peculiarities of Ceylonese art, as
compared with that of the continent, was the almost total
absence of sculpture known to us previous to the excavations
made within the last thirty-five years. Now, however, there have
been brought to light, besides carved capitals, string courses,
friezes, and the like — a number of statues of Buddha, his
disciples and other personages. The Tamil invaders, who often
ravaged the richest provinces of the island, were Brahmanical
Hindus, and had no respect for Buddhist idols ; whilst the
reported wealth of the shrines was a strong temptation to their
destruction in search of treasure. The Sinhalese, moreover,
were chiefly adherents of the Hinayana school, and had no
pantheon to compare with that of the northern schools, and
their principal figures would be those of the Buddha with his
attendants. Further, as is the case in Burma, where is an
unlimited amount of painting and carving, but little sculpture
properly so called, something similar may have occurred in
Ceylon. So far as we can now see, all the great topes were
covered with chunam, which may have been painted to any
extent, and all the viharas, as in Burma, were in wood or
brick, and consequently unfitted for permanent sculpture. But
there are evidences to show that most of the religious struc-
tures were ornamented with figures in chunam, in more or
less relief; and brick cores are met with on which representa-
tions of men and animals were moulded. Besides this, such
information as we have would lead us to suppose that painting
was a more favoured art with the islanders than sculpture.

1 Between 1875 and 1880 the Ceylon
Government had employed first Dr. P.
Goldschmidt, and after his death, Dr.
Edward Miiller, to copy the inscriptions ;
and in 1883 a thin volume of texts and
translations, with an accompanying series
of plates, was issued by the latter scholar.

* These reports are printed by the
Ceylon Government as 'Sessional Papers,'

and the earlier ones are accompanied by
rough lithographs of plans and sketches
in pen and ink. Mr. Bell, however, issues
his reports for 1900 onwards, with "half
tone " block illustrations.

It is to be hoped the Ceylon Govern-
ment, after having incurred the expense
of the survey, will not fail to make the
results available by adequate publication.

CHAP. VIII. CEYLON. 227

When Fah Hian, for instance, visited the island in A.D. 412-413,
he describes an accompaniment to the procession of the Tooth
relic as follows : " The king next causes to be placed on both
sides of the road representations of the 500 bodily forms which
the Bodhisattwa assumed during his successive births" — the
jatakas in fact. "These figures," he adds, "are all beautifully
painted in divers colours, and have a very life-like appear-
ance." 1 It was not that they could not sculpture in stone, for,
as we shall presently see, some of their carvings are of great
delicacy and cleverness of execution, but they seem to have
preferred colour to the more permanent forms of representa-
tion. Early figures of the Buddha are comparatively few :
possibly they were destroyed by the Tamil invaders ; still the
excavations of the last thirty years have brought to light
quite a considerable number of various ages. On the embank-
ments of many tanks there are slabs carved with five or seven
headed serpents, which may be of any age, and at the foot of
every important flight of steps there are two dwarpals or door-
keepers with this strange appendage, and attached to each of
the chapels of the Abhayagiri dagaba are figures of a great
Naga. These may be regarded as an evidence of the early
prevalence of the worship of serpents in the island.

Another peculiarity of the Ceylonese monuments is their
situation in the two capitals of the island, for, it will have
been observed, none of the remains of Buddhist architecture
described in the previous chapters are found in the great
capital cities of the Empire. They are detached monuments,
spared by accident in some distant corner of the land, or
rock - cut examples found in remote and secluded valleys.
The Buddhist Palibothra has entirely perished — so has vSravasti
and Vaual! ; and it is with difficulty we can identify Kapila-
vastu, Kusinara, and other famous cities, whose magnificent
monasteries and stupas are described by the Chinese travellers
in the 5th or 7th century of our era. In a great measure
this may be owing to their having been built of brick and
wood ; and, in that climate, vegetation is singularly destructive
of the first, and insects and decay of the second. But much
is also due to the country having been densely peopled ever
since the disappearance of the Buddhists. It may also be
remarked that the people inhabiting the plains of Bengal
since the extinction of Buddhism were either followers of the
Brahmanical or Muhammadan religions — both inimical to them,
or, at least, having no respect for their remains.

1 Beal, 'Buddhist Pilgrims,' p. 157; or 'Buddhist Records,' vol. i., introd. pp.
Ixxv., Ixxvi.

228

BUDDHIST ARCHITECTURE.

BOOK I.

In Ceylon the case is different. Though the great capitals
were early deserted, the mass of the people are still Buddhists, as
they have been for the last 2000 years, and there, consequently,
cities are still found adorned with monuments, which, though
in ruins, convey a sufficient impression of what those of India
must have been in the days of her glory.

Anuradhapura seems to have become the capital of Ceylon
about 370 years before Christ, or about a century after the
death of Buddha, and the fabled introduction of his religion
into the island. It was not, however, till about B.C. 240, that
it became a sacred city, and one of the principal capitals
of Buddhism in the East, which it continued to be till the
9th century. Then, owing to the repeated and destructive
invasions of the Malabars, an alternative capital was formed
at Polonnaruwa, which gradually supplanted Anuradhapura,
and became the sole capital till the I3th century. That
city reached its period of greatest prosperity and extension,
apparently in the reign of Parakrama Bahu I., 1153-86, and
then sank during a long and disastrous period into decay.
The seat of Government, after 1236, was moved hither and
thither, till the country fell into the hands of the Portuguese
and Dutch, and finally succumbed to our power.

ANURADHAPURA.

The city of Anuradhapura until within quite recent years
stood almost deserted in the midst of dense and sparsely
inhabited jungle. Its public buildings must have suffered
severely from the circumstances under which it perished,
exposed for centuries to the attacks of foreign enemies. Besides
this, the rank vegetation of Ceylon had been at work for
1000 years, stripping off all traces of plaster ornaments, and
splitting the masonry in many places. Now, however, it is
a prosperous town of about 4000 inhabitants, the capital of
the North-Central Province, and on the railway from Colombo
to Jaffna.

The very desolation, however, of its situation has preserved
the ancient monuments from other and greater dangers. No
bigoted Moslim has pulled them down to build mosques and
monuments of his own faith ; no indolent Hindu has allowed
their materials to be used for private purposes or appropriated
as private plunder ; and only to a limited extent have English
officials rendered them available for mending station roads
and bridges.1 We may be sure, however, that these ruins

1 As in India, the ruins in Ceylon have
suffered at the hands of the Public Works :

thus we learn that " countless pillars and
steps have been broken up to go into

CHAP. VIII.

ANURADHAPURA.

229

deserve the greatest attention from the student of Buddhist
architecture, and that a vast fund of information may be
drawn from them when they have been sufficiently explored
and fully delineated and described.

The peculiar fortune of Anuradhapura is that it continued
the capital of Ceylon for about ten centuries ; and, alone
of all Buddhist cities, it retains something like a complete
series of the remains of its greatness during that period. We
possess, moreover, in the ' Mahawansa ' and other Ceylonese
chronicles, a tolerably authentic account of the building of
these monuments,- and of the purposes to which they were
dedicated. Among the vestiges of its former grandeur still
to be found, are the ruins of half-a-dozen or more large
dome-shaped stupas or dagabas, and many smaller ones of
numerous monasteries and of a terraced enclosure erected
to contain the sacred Bo-tree, besides numerous other ruins and
antiquities. Among these is the great mound usually called Elala
Sohona, or the tomb of the usurper Elala ; but this traditional
name is incorrect, for recent excavation has shown that it
covers the remains of a large stupa about 180 ft. in diameter —
possibly the Dakshina stupa referred to in the chronicles.

Two of the dagabas are of the largest size known : of
one, the Abhayagiri,1 the dome, continued down to the ground,
is exactly hemispherical, and has a diameter of about 328 ft,
being thus more than 1000 ft. in circumference, and with the base
and spire must have made up a total elevation of about 260 ft.,
which is not far short of the traditional height of 120 cubits
assigned to it in the 'Mahawansa.'2 It is ascribed to King
Walagam-bahu or Vattagamani-Abhaya, who reconquered his
kingdom late in the first century B.C. from foreign usurpers
who had deposed him and occupied his throne for about
fifteen years ; and to commemorate the event he built a
vihara on the site of a Jaina temple. Nothing is said about
his erecting the dagaba or chaitya, though there must have

culverts on a road not traversed by a cart
once in six months," and the ruins at
Puliyan-kulama were "sadly destroyed
for ashlar to build three or four large
culverts on a branch road." Besides
later damages caused by reckless blast-
ing elsewhere, minor ruins also about
Anuradhapura disappeared on lands sold
to natives prior to 1890. The restorations
most to be feared are those by the priests,
who "are erecting at Ruwanveli dagaba
a series of shrines in a modern style
absolutely frightful." — General L. de
Beylie, ' L' Architecture Hindoue en
Extreme-Orient,' p. 364.

1 Until the accuracy, or otherwise of
the current identifications are fully in-
vestigated, we can only follow the tradi-
tional account which, in the case of the
more notable dagabas can hardly be in
error. The capital and pinnacle of the
Abhayagiri were restored by prison labour
in 1890.

2 The cubit of Ceylon is nearly 2 ft.
3 in. ; it has sometimes been taken as
2 ft. 3f in. The present total height
from the platform to the top of the
ruined spire is 232 ft. — Smither, 'Archi-
tectural Remains : Anuradhapura,' p.
47-

230

BUDDHIST ARCHITECTURE.

BOOK I.

been one in connection with the vihara. It seems to be
referred to about a century later ; : and by Gajabahu I., in
the 2nd century A.D. we are told "it was constructed of
a greater elevation, and he caused arches to be built at the
four gates." Such casing of smaller dagabas to enlarge them
was not infrequent. The stupa stands on a stone-paved
platform 590 ft. square, raised about p| ft. above the ground
level, and ascended on each side by a flight of steps 27 ft.
wide. The excavations here and at the other large dagabas
have shown that, as with the Nepal and Indian chaityas,
at each cardinal point there were richly carved, oblong
projections from the circle of the lower pasadas or terraced
basement, which were, doubtless, the chapels or thrones for
the Dhyani Buddhas. We have a similar arrangement also
at Sanchi and in the Kalinga stupas.2

The second tope is the Jetawanarama, begun by King
Mahasena in the 4th century, and finished by his successor,
Kittis-Siri Meghavanna. In form and dimensions it is almost
identical with the last described, though somewhat more perfect
in outline, owing probably to its being more modern than its
rival.3 Its chapels seem to have been quite ruined.

Next to these, but far more important from its sacredness,
is the Ruwanveli dagaba, erected by King Dutthagamani,
between the years B.C. 102 and 78, over a very imposing collec-
tion of relics, of which a full account is given in the 28th to
3 ist chapters of the ' Mahawansa..' Its dimensions are very
similar to those of the last two described ; but it has been so
much defaced, that except the remains of the circular plinths
round its base, it has, like the rest, become only a huge shape-
less mound of solid brickwork. The excavations, however,
have made it plain that the dome had a diameter of 252 ft.
8 in., that, like all the others, it stood upon a basement of three
plinths called pasadas, or procession terraces, together 15 ft.
high, and rose from a stilted drum to a greater height than a
hemisphere. The ' Mahawansa ' says it was 120 cubits high, or
about 270 ft., but the present mound stands only about 179 ft.
above the paved platform, which is 5 ft. 7 in. above the ground

1 Mahawansa, ch. 35.

2 'Journal of the Royal Asiatic Society,'
1902, p. 32 ; ' Buddhist Art in India,'
p. 195. The dsana, or seat for Vairochana.
is in the relic chamber. At the Mirisa-
veti dagaba at Anuradhapura, and at
Rankot and Kiri dagabas at Polonnaruwa,
Mr. Bell has found several detached cells
or chapels — probably of late date. That
the'MahSyana ritual had found acceptance
in Ceylon by the gth century or earlier,

is evidenced by Mr. Bell's discovery at
the Vijayarama monastery of copper
plaques bearing invocations to Akasa-
garbha, Vairochana, TarS, etc. -
' Sessional Papers,' 1896, pp. 460,
464-467.

3 Its dome is 310 ft. in diameter and its
height to the top of the square capital,
187 ft. 6 in., and to the top of the ruined
spire 245 ft. — Smither's 'AnurSdhapura,'
plate 47.

CHAP. VIII.

ANURADHAPURA.

231

level, and measures 475 ft. from north to south by 473 ft.
across, the retaining wall being carved with elephants. The
four thrones or chapels are found facing the cardinal points
—that on the south being the least ruined.

The same king had previously erected another smaller
dagaba, about 133 ft. in diameter at the rise of the dome.
It is known as the Mirisavetiya dagaba, and like the last
described it is very much ruined. Like the other dagabas it
had three low pasadas or terraces round the base from 5 to
6 ft. in breadth, together projecting 16 ft. 7 in. from the bell
of the dagaba, and rising to 13 ft. 5 in. from the pavement.1

127. View of the north side of the west chapel, Ruwanveli Dagaba. (From a Photograph.)

An excavation on the west side, however, revealed a handsome
chapel, similar to those at the other dagabas, but differing in
detail, and in a much more perfect condition. The whole is
elaborately carved in horizontal bands with elephants and other
animals, rosettes, etc., and at the ends are richly-carved stelae
surmounted by lions. Behind the chapels at the Ruwanveli

1 For full description and drawings, see
Smither's ' Anuradhapura,' pp. 19-22,
and plates 14-21. Excavation on the
east side, in search of a corresponding
structure proved fruitless. In 1890 the

dagaba was "being elaborately restored
by prison labour at the expense of a
Siamese prince.'' — 'Sessional Papers,'
1890, p. 43. This restoraiion has been
left unfinished for want of funds,

232

BUDDHIST ARCHITECTURE.

BOOK I.

dagaba are flights of steps by which access was had from the
first to the second pasada or terrace. The first was reached
from the pavement by a stair on the east side of the south
chapel ; whilst behind this chapel there is only one stair from
the second to the third terrace. A somewhat similar arrange-

128. Part Elevation (restored) of front of the south chapel, Ruwanveli Dagaba. Scale i-85th.

ment of stairs also existed at the Abhayagiri and Mirisavetiya
dagabas ; and it is very probable also at the Jetawanarama
dagaba, though the published results of the surveys do not show.
The facade of the chapels consists first of a plain base, above
which is a row of kneeling elephants with paterae between them,1
very like those used in the metopes of the Roman Doric order ;
above this are three plain faces divided by ornamental string-
courses ; then a bracket cornice with paterae again, and above
this two more plain faces and string-courses. Over this there
was probably a frieze of animals and a band simulating a
Buddhist rail, with a blocking course over it, as at the Mirisa-
vetiya dagaba.

At each end of this projecting arrangement were two stelae
— the inner covered by foliaged and other patterns, the outer
in one instance, at the Ruwanveli, by a seven-headed serpent,
as will be observed in the Woodcut No. 127 ; at the Abhayagiri,
there are serpent figures at all the chapels — each on a separate
stone — and here the inner stele is adorned with a pattern so
nearly identical with that on the pillars of the western gateway
at Sanchi,2 that we may recognise them as belonging to about

1 At the Mirisaveti these are lotus
flowers.

2 ' Tree and Serpent Worship,' plate 19.
In some respects it resembles the Wood-
cuts Nos. 39 and 40. Similar stelse were

also found at Amaravati. — 'Archceolog.
Survey of Southern India : Amaravati
and Jaggayyapeta Stupas,' plate 33,
fig. 2 ; plate 38, fig. 7 ; plate 40, fig. 3 ;
plate 44, fig. i ; and plate 54, fig. T.

CHAP. VIII.

ANURADHAPURA.

233

the same age. On the other stele in this tope (Woodcut No. 129)
we recognise the shield, the Swastika, the Triratna, and other
Buddhist emblems with which we are already familiar.1

129. Stelae at the east end of the north chapel, Abhayagiri Dagaba. ( From a Photograph . )

All this is architecturally so unlike anything we find of
the same age on the continent of India, while its sculptured
details are so nearly identical, that, when we come to know more
about it, these differences and similarities may lead to most
important inferences ; but we must at present wait for the
requisite information to enable us to see the bearing of these
peculiarities.

Besides these four large buildings there are two smaller ones,
known as the Thuparama and Lankarama, very similar to one
another in size and arrangement. The first named is represented
in Woodcut No. 130. The dagaba itself, though small, was
originally of a singularly elegant bell-shaped outline.2 As it

1 For photographic illustrations of the
stelse, at each of the chapels of this dagaba.
— Smither's 'Anuradhapura,' plates 41-43.

2 Since the drawing was made from

which this cut is taken, it was thoroughly
repaired in 1842, and made as unlike
what it was as can well be conceived.
— Smither's 'Anuradhapura,' plates 2-8.

234

BUDDHIST ARCHITECTURE.

BOOK I.

appears since it was restored in 1842, is shown on Plate No. II.
Its diameter and height are now nearly the same — at the base
about 55 ft; and it stands on a platform raised about II ft. from

130. Thuparama Dagaba. (From an unpublished Lithograph by the late James Prinsep. )

the ground, on which are arranged four rows of tall pillars
of strikingly slender proportions, which form by far the most
important architectural ornament of the building. The inner
circle stands about 3 ft. from the dagaba, and the next two
about 10 ft. from each other. The shafts are monoliths 22 ft.
10 in. in height in the inner row, and diminish successively
by about 18 in. in the next two rows, in each of which the
lower part, to a third of the height, is left square, each side being
about i ft. These sustain octagonal capitals of singularly grace-
ful outlines, 2 ft. I in. in height, and 2 ft. 2 in. across the top.
They are carved with figures and foliage, and under the capitals
the pillars are ornamented with fringes 14 in. deep, depending
from kirttimukh faces carved in low relief on the angles. The
sculpture on the capitals of the first and second circles are
similar, namely squatting or dwarf human figures ; in the third
row the ornaments differ. The pillars in the fourth or outer
circle are monoliths, 14 ft. in height including the capitals ; they
are entirely octagonal, and their shafts are 10 in. diameter.
The capitals are very similar to those of the inner circles, but
differ in dimensions and ornamentation. They are 21 in. in
height and i8£ in. across the abacus, and are sculptured with
sixteen capering dwarfs (Woodcut No. 131). They have
octagonal seatings on the tops of 2\ in. high and 10 in. diameter,

CHAP. VIII.

ANURADHAPURA.

235

over which is a rounded boss 4 in. in diameter. Some of

the capitals in the two inner rows have raised pads, and various

forms of seatings that might have been supports

for images or symbols ; all in the third circle

have square pads on them. Originally there

have been 128 of these pillars belonging to the

three inner circles,2 and 48 more in the outer or

fourth row ; the latter are more slender than the

others, and stand in a circle 14 ft. beyond the

third.

This relic-shrine was originally erected by
the celebrated King Devanampiya Tissa, about
246 years B.C., to contain the right collar bone of
Buddha, which — say the Buddhist chroniclers —
descending from the skies, placed itself on the
crown of the monarch. As contemporary with
Ajoka it belongs to the most interesting period I3I. capital from outer
of Buddhist history, and is older, or, at least, circle at Thuparama
as old, as anything now existing on the continent Dasaba- Scale Jv^-1
of India ; and there is every reason to suppose
it existed till 1842, as nearly as may be, in the form in which
it was originally designed, having escaped alteration, and, what
is more unusual in a Buddhist relic-shrine, having escaped
augmentation. When the celebrated Tooth Relic was brought
hither from India some time during the 4th century, it was
deposited in a small building erected for the purpose to the
south-east of the circular platform. This was known as the
Dalada Maligawa or Tooth relic Temple.

The Lankarama (Woodcut No. 132) is extremely similar to
the last — though there is no distinct historical mention of its
erection in the Sinhalese chronicles. Its being encircled by
pillars, like the Thuparama, might suggest that it belonged to
about the same age,3 and this seems supported by Mr. Smither's
drawings, which point in that direction. The building, however,
has more than once undergone restorations that may have nearly
obliterated its more ancient features. Parakrama Bahu I. (1153-
ii 86) repaired many of the old monuments — most probably

1 From Smither's, ' Anuradhapura ',
plate 7, fig. 3.

2 These pillars were arranged thus :
52 in the inner circle, 36 in the second,
and 40 in the third ; those in the inner
circle are only 2 ft. 5 in. apart, except in
front of the chapels, where they are about
9 ft. 4 in. apart in each circle. None of
the other pillars in one circle are directly
opposite pillars in the next.

* Captain Chapman said it was built by

Abhaya Tissa, A.D. 231; Major Forbes
assigned it to Mahasena between 277 and
304 (Sinhalese dates) ; and Mr Smither
suggests that it might be the vihSra,
unnamed by Tumour, mentioned as built
by MahSsiva, dr. B.C. 190; this last,
however, was the Nagarangana vihara.
— 'Journal of the Royal Asiatic Society,'
vol. xx. pp. I75f. ; Forbes 'Eleven Years
in Ceylon ' ; Wijesinha, ' MahSwansa,'
part i., p. 81.

236

BUDDHIST ARCHITECTURE.

BOOK I.

this among them, and in the i8th century it underwent a special
restoration.

The base is 9 ft. high, and the dome upon this is hemi-

132. Lankarama Dagaba (1870). (From a Photograph.)

spherical and 38 ft. in diameter. The lower 4 ft. of the base has
apparently been extended by an addition of
varying breadth all round, which includes
the innermost row of very graceful pillars.
The second circle is 7 ft. 8 in. in advance of
the first, and the third — like the fourth at
the Thuparama — consists of more slender
shafts only 12 ft. 5 in. high, and stands
i6| ft. outside the second circle. These
pillars are all monolithic — the shaft and
capital being in one piece. Those of the
second circle stand 16 ft. 11 in. above the
raised pavement on which the dagaba stands,
the level of the heads of the inner circle
being 5 in. higher. In both, the shafts are

133. Capital of Lankarama 1 3 ™- square to about a third of their

pillars in the inner circles.1 height, above which they are octagonal.

Scale ^th. The capjtais (Woodcut No. 133) are 26| in.

high and 2 ft. across at 6 in. below the top, from which they

1 From Smither's ' Anuradhapura,' plate 13, fig. i.

CHAP. VIII. ANURADHAPURA. 237

curve inwards to 12 in. broad at the top, which is flat. All
these capitals are carved with seated lions. In the third and
outermost circle the shafts are 10 in. in diameter and octagonal
throughout ; their capitals are 25 £ in. high and at greatest width
19 in., diminishing to 9^ in. at the top, ornamented, like those
of the Thuparama, with sixteen dwarf figures instead of lions.
There were 20 pillars in the inner circle, 28 in the second, and
40 in the third — those in one circle having no relation in
position to those in the next.

As will be observed, the two last-mentioned dagabas present
us with a peculiarity not found on any example we have else-
where met with outside Ceylon, inasmuch as they are sur-
rounded by three circles of slender monolithic pillars of very
elegant design. The purpose of those pillars is somewhat per-
plexing : it has been suggested that they may once have carried
a roof,1 but they are so slender, and arranged without relation of
those in one circle to those in the next, that any roof — of however
light materials — could hardly have been placed on them ; and the
examination of the capitals does not favour such a hypothesis.
They rather correspond to the rails of Indian stupas, marking
out procession paths or pradakshinas, whilst they were probably
also employed as stambhas or lats. The tallest of them, in the
two inner circles at the Thuparama, had tenons of different
sizes and forms on the capitals — hardly two of them being alike
— which indicate that they may have supported various religious
symbols and images, such as Dharmachakras, Triratnas, etc.
The pillars of the outer circles at the Thuparama and Lankarama
had also pads or fastenings on their capitals as if to retain metal
or other symbols — perhaps of a different type from those on
the two inner rows.2

There is still another — the Kujjatissarama, better known as
the Selachaitiya dagaba — between the Ruwanveli and Abhayagiri
stupas, but so ruined that its architectural features were undis-
tinguishable until excavated in 1895. It was a mere mound of
ruined brickwork, rising about 1 5 ft. above the platform. The
base has been about 37 ft. 5 in. in diameter, and it stands on a
paved platform 46 ft. 9 in. square, rising 7 ft. 6 in. above the
ground level, and enclosed by a stone parapet, with entrances
on the east and south sides. It may perhaps belong to the
reign of Lajji Tissa — about 55 B.C.3 The spot at all events is
said to have been hallowed by the presence of Ka^yapa, the
Buddha preceding Sakyamuni.

1 General L. de Beylie, ' L'Archi- and plate 8.

lecture de Hindoue en Extreme-Orient,' 3 Mr. Bell identifies it with the 5ila

p. 361. I Thupa built by that monarch. — 'Sessional

2 Smither's ' Anuradhapura,' pp. 5-7, i Paper,' xl. 1904.

238 BUDDHIST ARCHITECTURE. BOOK I.

Besides these, there are on the hill of Mihintale, eight miles
to the east of the city, two important relic-shrines ; one of the
first class — the Mahaseya, erected on its summit to cover a hair
that grew on the forehead of Buddha over his left eyebrow.
The other — the Ambasthala dagaba — on a shoulder of the hill
immediately below this — is of the same class as the Thiiparama ;
it stands on the traditional spot where King Devanampiya-Tissa
first met the Thera Mahinda, and is said to have been erected
by that king. The small central building stands on a base
29 ft. in diameter and about 3 ft. high ; and the dome, where it
rises from this, is 23 ft. in diameter and about 20 ft. high, rather
oval in curvature and surmounted by a square capital support-
ing a stunted spire — the total height being about 30 ft. from
the circular pavement on which it stands. It is surrounded
by two concentric rows of pillars, which, as appears to have
been usual when this mode of decoration was employed, rose
to half the height of the central mound. The inner circle of
twenty pillars stands 5 ft. from the basement, and the outer,
of thirty-two shafts, is 12 ft. farther out. They are 12 ft. high
with octagonal capitals 2 ft in height. The platform is reached
on the west side by a granite stair.1

There are, in addition to these, a great number of dagabas
of various sorts scattered over the area once covered by the old
city, but whether any of them are particularly interesting, either
from their architecture or their history, has not been ascertained,
nor will it be till the whole site has been systematically and
carefully surveyed.

There is another ruin at Anuradhapura, which, if a little
more perfect, would be even more interesting than those topes.
It goes by the name of Loha Mahapaya, or Great Brazen
Monastery. We have a full account in the ' Mahawansa ' of its
erection by the pious King Dutthagamani (dr. B.C. ioo),2 accord-
ing to a plan procured from heaven for the purpose — as well
as a history of its subsequent destruction and rebuildings.

When first erected it is said to have been ioo cubits or
230 ft. square, and as high as it was broad ; the height was
divided into nine storeys, each containing ioo cells for priests,
besides halls and other indispensable apartments. Nearly
200 years after its erection it required considerable repairs,
but the first great disaster occurred in the reign of Mahasena
(4th century), who is said to have destroyed it utterly.3 It was
re-erected by his son, but with only five storeys instead of nine ;
and it. never after this regained its pristine magnificence, but

1 Smither, ' Anuradhapura,' p. II.

'2 ' Mahawansa,' Tumour's translation, p. 163, ch. 27.

8 Loc. cit. p. 235, ch. 37.

CHAP. VIII. ANURADHAPURA. 239

gradually fell into decay even before the seat of government
was removed to Polonnaruwa. Since that time it has been com-
pletely deserted, and all that now remains are the 1600 pillars
which once supported it. These generally consist of unhewn
blocks of granite about 12 ft. high; some of the central ones
are sculptured, and many have been split into two, apparently
at the time of the great rebuilding after its destruction by
Mahasena ; as it is, they stand about 6 ft. apart from centre to
centre in a compact phalanx, forty on each face, and covering
a space of 250 ft. or 260 ft. each way. Upon the pillars must
have been placed a strong wooden framing from which the
remaining eight storeys rose, as in the modern Burmese
monasteries.

There is only one difficulty in understanding the arrange-
ment of the superstructure of this building, and that is the
assertion of the ' Mahawansa ' that it consisted of nine storeys
— afterwards of five — each containing 100 apartments. For
myself I have no hesitation in rejecting this statement as
impossible, not only from the difficulty of constructing and
roofing such a building, but because its form is so utterly
opposed to all the traditions of Eastern art. If we turn back
to Fah Hian or Hiuen Tsiang's description of the great
Dakhani monastery (page 171) or to the great rath at Mamalla-
puram (Woodcut No. 89), or, indeed, to any of the 1001 temples
of Southern India, all of which simulate three, five, or nine-
storied residences, we get a distinct idea of what such a build-
ing may have been if erected in the Indian style. It would,
too, be convenient and appropriate to the climate, each storey
having its terrace for walking or sleeping in the open air, and
the whole easily constructed and kept in order. All this will
be clearer in the sequel, but in the meanwhile it hardly appears
doubtful that the Loha Mahapaya was originally of nine, and
subsequently of five storeys, each less in dimension than the
one below it. The top one was surmounted as at Mamallapuram
by a dome, but in this instance composed of bronze — whence
its name ; and, gilt and ornamented as it no doubt was, it must
have been one of the most splendid buildings of the East. It
was as high as the dagabas, and, though not covering quite so
much ground, was equal, in cubical contents, to the largest of
our English cathedrals, and the body of the building was higher
than any of them, omitting of course the spires, which are mere
ornaments.

Besides these there are scattered about the ruins of Anura-
dhapura many groups of pillars and basements that evidently
belonged to viharas, monasteries and halls for various

240

BUDDHIST ARCHITECTURE.

BOOK I.

purposes. They were all raised on platforms or stylobates,
and approached by one or more flights of steps, of a highly

134. Vihara with Steps west of Ruwanveli Dagaba. (From a Photograph.)

135.

Moonstone at Foot of Steps leading to the Platform of the Bo-tree, Anuradhapura.
(From a Photograph.)

ornamental character. One of these, leading to a group of
pillars attached to the Ruwanveli dagaba, will convey some
idea of their general character (Woodcut No. 134). At the
foot of a flight of steps is a richly carved semicircular stone

CHAP. VIII. ANURADHAPURA. 241

threshold, popularly known as a " moonstone " l (Woodcut
No. 135). Many of these are found at Anuradhapura, and
as many probably at Polonnaruwa. Some are large and some
smaller than others, but they are all broadly similar in design.
They are not peculiar, however, to Ceylon : in temples, especially
in the south of India and in the cave temples, they are usually
found in front of the entrance to the shrine and often at the
outer doorways, and are known as lotus-slabs — the general
pattern resembling the lotus flower. Inside an outer ornamental
ring, in Anuradhapura examples, is a procession of animals,
divided from the next compartment by a richly elaborate scroll ;
within that again a row of birds bearing lotus buds, and then
a lotus flower with a disc edged with leaves. The animals
are always elephants, horses, lions, and bulls : the birds — hansas
or sacred geese — chakwas.2 These, it will be recollected, are
the animals which Fah Hian and Hiuen Tsiang describe as
ornamenting the five storeys of the great Dakhani monastery,
and which, as we shall afterwards see, were also arranged at
Halebid in the I3th century in precisely the same manner.
For 1 500 years they, and they only, seem to have been selected
for architectural purposes, but why this was so we are yet
unable to explain.

The risers of these stairs, though not adorned with storeyed
bas-reliefs, like those of the Jamalgarhi monastery in Gandhara,
are all richly ornamented, being divided, at Anuradhapura, into
two panels by figures of dwarfs, and framed by foliaged borders,
while the jambs or flanking stones are also adorned by either
figures of animals or bas-reliefs.

These steps lead to platforms on which stood various
structures, as witnessed by the monoliths still standing on
some of them ; and, so far as information is available, the
buildings were ecclesiastical, surrounded by brick walls, and
the roofs supported on them and the pillars. In the case of the
so-called Mahasena's pavilion, and many other ruins of that
type, there was a central and four subsidiary structures in the
corners of the enclosure, which together constituted the vihara
— the larger and central building being probably a temple

1 In Sinhalese — "Sanclakada pahana,"
in Sanskrit — " padmasilam." — Tawney's
' Prabandha-chintamani,' p. 57. They
are also called " ardhachandras " — "of
half moon form." At the entrance
to a vihara north of the Lankarama
dagaba, a fine flight of steps was ex-
cavated about twenty years ago, the
large threshold stone to which presents
the lotus only. — Cave's ' Ruined Cities

of Ceylon' (8vo. ed.), p. 106, and plate
35. One from the Dalada Maligawa at
the Thuparama is represented in a photo-
collotype in Smither's 'Anuradhapura,'
plate 57, fig. 3 ; and another drawn to
a small scale from the mis-named Maha-
sena's pavilion, on plate 59 ; and in
Cave's ' Ruined Cities,' plate 32,

2 The Polonnaruwa examples are more
crowded with ornament.

VOL. I. Q

242

BUDDHIST ARCHITECTURE.

BOOK I.

containing an image of the Buddha.1 In others, such as that
styled the Vijayarama vihara, there was a small dagaba as
well as a temple and other structures ; and connected with
each such vihara would be the indispensable Service-hall or
Uposathagharam. The "preaching halls" which Fah Hian
mentions at the head of the four principal streets, where the
religious members of the community of all classes assembled
on stated days to listen to the preaching of the doctrine or
" bana," may have been connected with certain of these viharas.2

Besides these there is at Anuradhapura a temple called
Isurumuniya, partly cut in the rock, partly structural. Till
within the last forty years the pillars of its porch still carried
the wooden beams of a roof, but whether it was the original
one or a subsequent addition is by no means clear. From the
mortises in the face of the rock I would be inclined to believe
that it was at least in the original form, but the building has
been so knocked about and altered in modern times, that it is
impossible to speak with certainty regarding it. So far as can
be judged from photographs, I would be inclined to ascribe
the original excavation to the 6th or 7th century.3 The
architecture of the steps and the Naga dwarpals are all of the
old pattern, but coarser and showing unmistakable signs of
decadence.

The excavations directed by Mr. Bell, among other important
discoveries, have brought to light a regular " Buddhist railing "
surrounding a rectangular site, near the Abhayagiri Dagaba.
The pillars of this rail were only 3 ft. 10 in. in height by 8 in.
square — quite diminutive as compared with Indian examples
— and standing I ft apart with three cross - bars, surmounted
by a coping 8 in. high ; but it stood on a moulded basement
about 3 ft. 9 in. in height, thus raising the whole to about 8 ft.
3 in. high. Except a little carving on the jambs of the entrance,
the whole is perfectly plain.4

To us these are the most interesting of the remains of the
ancient city, but to a Buddhist the greatest and most sacred of
the vestiges of the past is the celebrated Bo-tree. This was long
reverenced and worshipped even amidst the desolation in which it
stood, and has been worshipped on this spot for more than 2000
years ; and thus, if not the oldest, is certainly among the most
ancient of the idols that still command the adoration of mankind.

1 Smither's ' Anuradhapura,' pp. 59-60
and plates 58, 59 ; and Spence Hardy,
' Eastern Monachism,' pp. 200-201.

2 Fah Hian, chap. 38.

3 " The doorway is fine, and the temple
is unique in many respects." — Cave's

' Ruined Cities of Ceylon,' p. 63.

4 ' Second Archaeological Survey Re-
port,' pp. 3, 4. A small fragment of
Buddhist railing had also been found
here in 1873. — Smither's 'Anuradha-
pura,' p. 50 and plate 44.

CHAP. VIII.

ANURADHAPURA.

243

When A.roka, according to tradition, sent his son Mahinda
and his daughter Sangamitta, to introduce Buddhism into
Ceylon, one of the most precious things which they brought
was a branch of the celebrated tree which still grows at Bodh-
Gaya1 (Woodcut No. 19). The branch, so says the legend,
spontaneously severed itself from the parent stem, and planted
itself in a golden vase prepared for its reception. According
to the prophecy, it was to be " always green, never growing
nor decaying," and certainly present appearances would go
far to confirm such an assertion, for, notwithstanding its age, it
is small, and does not seem to increase. Its being evergreen
is only a characteristic of its species, the Ficus religiosa ; our
acquaintance with it, however, must extend over a longer series
of years than it yet does, before we can speak with certainty
as to its stationary qualities. Its branches, however, are already
propped up to preserve them.

136. View of the Sacred Bo-tree, Anuradhapura. (From Sir E. Tennent's ' Ceylon.')

It grows from the top of a small pyramid, which rises in
three terraces, each about 12 ft. in height, in the centre of a
large square enclosure popularly known as " Udamaluwa," but
by the priests called Maha-Vihara. But though the place is
large, sacred, and adorned with stairs of some pretension, none
of the architectural features which at present surround it are
such as to require notice in a work like the present.

1 Singularly enough, the natives of
Bihar ascribe the planting of their Bo-
tree to Dutthagamani, the pious king of

Ceylon. — Buchanan Hamilton's 'Stat-
istics of Bihar,' Montgomery Martin's
edition, vol. i. p. 76.

244

BUDDHIST ARCHITECTURE.

BOOK I.

POLONNARUWA.1

Although very much more modern in date, and consequently
less pure in style, the ruins at Polonnaruwa are scarcely less
interesting than those of the northern capital to which it suc-
ceeded. They form a link between the ancient and modern
styles at a time when the Buddhists had ceased to exist, or at
least to build, on the continent of India, and, when properly
illustrated, will enable us to speak with confidence of much
that we find beyond the Ganges. Much of what we know of
these ruins is due to the publications of Sir Emerson Tennent,2
which, though most valuable contributions, were far from
exhausting the subject. According to this authority, the
principal ruins extend in a line nearly north and south for
about a mile and a half from the palace to the Gal Vihara,
and comprise two dagabas, besides a number of smaller edifices.
The greater part seems to have been erected during the reign
of Parakrama Bahu I., 1153-86, though, as the city became the
capital of the kingdom in the 9th century, it is probable that
investigation may yet reveal some of earlier date ; while, as it
was not finally deserted till 1293, some of them may also be
more modern.

If not the oldest, certainly the most interesting group at
Polonnaruwa is that of the rock-cut sculptures known as the
Gal Vihara. They are not rock - cut temples in the sense in
which the term is understood in India, being neither residences
nor chaitya halls. On the left, on the face of the rock, is a
figure of Buddha, seated in the usual cross-legged conventional
attitude, 15 ft. in height, and backed by a throne of exceeding
richness : perhaps the most elaborate specimen of its class
known to exist anywhere. Next to this is a cell, with two
pillars in front, on the back wall of which is another seated
figure of Buddha, but certainly of a more modern aspect than
that last described. Beyond this is a figure standing in the open
air, now supposed to represent Ananda — the cousin of Buddha ; 3

1 In inscriptions the city is called
Pulastipura and Kalingapura, and its
modern name is Topawaewa or Topawa.
As, however, that here given is the only
one by which it is known in English
literature, it is retained.

1 ' Christianity in Ceylon,' Murray,
1850; 'An Account of the Island of
Ceylon,' 2 vols., Longmans, 1859. Mr.
Lawton's and Captain Hogg's photo-
graphs added considerably to the pre-
cision but not to the extent of our
knowledge. Mr. Smither's volume on
Anuradhapura is by far the most im-

portant addition yet made to the archi-
tecture of Ceylon. Mr. Bell's ' Progress
Reports ' are a mine of information, but
require to be digested and arranged with
fuller and better architectural illustrations.
As yet the survey of Polonnaruwa is not
published, but Mr. Bell has very kindly
supplied me with a proof of his ' Annual
Report ' for 1903, which, with three
preceding, is devoted chiefly to the
remains at that place.

3 This is a modern local designation :
except perhaps in China, Ananda scarcely
appears in Buddhist iconography.

CHAP. VIII.

POLONNARUWA.

245

and still further to the right another of Buddha, lying down
in the conventional attitude of his attaining Nirvana. This
figure is 46 ft. long, while the standing one is only 23 ft. high.1
These Nirvana figures are rare in India, but there is one in
the most modern cave at Ajanta, No. 26, 23 ft. 3 in. long (Wood-
cut No. 29 on page 101), and others in the latest caves at Nasik
and Salsette. None of these, however, so far as I know, ever
attained in India such dimensions as this.

»Not far south of the Gal Vihara stands one of the principal
religious groups of the city, consisting first of the Jetawanarama
Temple, built of brick, 170 ft. long by 70 ft. wide, with walls
12 ft. in thickness covered with chunam, and still about 70 ft.
high. It was divided into two halls, the inner or shrine being
wider than the outer or eastern one, and containing an erect statue
of Buddha built of brick, 58 ft. in height, much injured. The
entrance is flanked by two polygonal turrets, on the bases of
which were dwarpalas or yakshas in high relief, and the highly
carved stone steps at the entrance were each 20 ft. long. On
the north side of it is the Kiri Dagaba — about 70 ft. in diameter
and nearly 100 ft. in height — with two smaller topes, standing on
raised platforms ; the whole space, measuring 577 ft. by 500 ft,
was apparently at one time entirely filled with objects of
religious adoration. The whole certainly belongs to the age
of Parakrama Bahu I. It was, however, built of brick, and
plastered, which gives it an appearance of inferiority even
beyond what is due to the inferior style of that age.

Next in importance to this is the Rankot Dagaba, about
500 yards south of the Jetawanarama, 186 ft. in diameter, and
of about the same in height. This, though only half that of
some of those in the older capital, is still larger than any
known to exist on the continent of India. It is ascribed to
Kirti Nissanka Malla, a Kalinga prince, at the end of the
1 2th century, and is in fair preservation. Its base is surrounded,
like those in Burma, by eight small brick shrines — two at each
of the cardinal points — having conical roofs, and between
each pair is an dsana, or seat for a Dhyani Buddha.

At some five furlongs south from this stands the Sat Mahal
Prasada (Woodcut No. 1 37), which is one of the most interest-
ing buildings of the place, as it is one of the most perfect
representations existing of the seven-storeyed temples of Assyria.

1 There are two colossal statues of
Buddha, one at Sseseruwa, in the North-
Western Province, 39 ft. 3 in. high, the
other at a place called Aukana, to the
east of the Kalawsewa tank, in the North
Central Province, 39 ft. high. They are

extremely similar to one another, and —
except in dimensions and position of the
arms — to that at the Gal Vihara. —
'Sessional papers, 'xl., 1904, pp.6and 12.
A descriptive inventory of the monu-
ments of Ceylon is a great desideratum.

BUDDHIST ARCHITECTURE.

BOOK I.

It is also interesting as affording a hint as to the appearance
of the five or nine-storeyed monasteries mentioned in a previous

137. Sat Mahal Prasada and Galpota, irom the south. (From a Photograph.)

page (239). This one, however, never was a residence, nor
does it simulate one, like the raths at Mamallapuram or other
buildings in the Dravidian style, which will be described in a
subsequent chapter. Its base is 28 ft. 6 in. square, each storey
diminishes in size and height — the uppermost being ruined — but
the total height is still 53 ft. Statues of stucco, in high relief,
ornamented each storey ; and there is a flight of steps, but it
reaches only to the top of the first storey.1 The style of this
peculiar tower suggests a comparison with those structures
known in Cambodia as " Prasats," from which it seems to be
copied ; and about the time when this one was erected by
. Nissanka Malla at the end of the I2th century, Ceylon was
in pretty close intercourse with Cambodia.2

In front of it lies a splendid stone table 26 ft. 10 in. long,
4 ft. 7 in. broad, and from 16 to 26 in. thick. It is known
as the Galpota or stone book, and bears a long inscription

1 Cave's ' Ruined Cities of Ceylon,'
p. 154; Mr. Bell's 'Report for 1903.'

2 Compare illustrations of Prasats in
Lajonqui&re, ' Inventaire Descriptif des

Monuments du Cambodge,' tome i. pp.
xx., xxii., 199, 2OI, 218; Aymonier,
1 Le Cambodge,' tome ii. p. 427, etc. ;
' Mahawansa,' ch. 76, vv. 21, 22.

CHAP. VIII.

POLONNARUWA.

247

recording the acts and virtues of King Kirti Nissanka Malla
(1187-1196). The carving on its border represents a double
row of hansas or sacred geese — always a favourite subject of
the Buddhist sculptors.1 At each end of this stone is engraved
a representation of Sri with her two elephants with their water-
pots (Woodcut No. 3).

Beside the Sat Mahal Prasada is the Wata-Da-ge, a
circular building, which, so far as is at present known, is all
but unique.2 It is a circular enclosure, open to the sky, 58 ft.
2 in. in diameter, and surrounded by a wall 14 ft. high, once
decorated with paintings (Plate No. III.). Inside were found the
remains of a small brick dagaba and broken figures of Buddhas
at the cardinal points, as also the broken shafts of pillars of
two circles that had surrounded the dagaba. Round the outer
circumference of the wall is a narrow passage, enclosed by a
highly ornamental screen about 3 ft. high, adorned with a
range of thirty-two slender pillar shafts, 6 ft. in height with
highly carved capitals, like those of the outermost circle at
the Thuparama dagaba (Woodcut No. 130). Below this is a
richly carved stylobate, about 4 ft. 6 in. high, standing on a
circular platform 120 ft. in diameter, about 20 ft. broad and
4 ft. 6 in. above the ground level. The principal entrance is
from the north side, but at the other cardinal points also are
flights of steps leading up to the enclosure, more elaborate than
any others that have yet been discovered in Ceylon. They
all have highly carved thresholds or moonstones to start from.
Their risers are each adorned with twelve figures of dwarfs,
and their side pieces, or jambs, are of exceptional richness, and
each has a pair of Naga-headed dwarpals at the sides of its steps.3
Altogether this is one of the most interesting buildings in Ceylon,
as well as one of the richest in sculptural decorations.4

Close to the Wata-Da-ge, on its south-west side, is the
Thuparama temple, a large, oblong brick structure, built by
Parakrama Bahu I., the walls of which are full 5 ft. in thickness
(Plate No. IV.). The principal entrance is on the east side and
a smaller one on the north, and it has four narrow windows
divided by round mullions. The temple consists of a vestibule
and inner hall, vaulted in the Hindu method by corbelling
inwards the successive layers of brick. Over this is a low, square
tower. Round the base of the building runs a low dado of
lions somewhat boldly worked in stucco.

1 They occur also on Ajoka's pillars in
the earliest known sculptures in India
(Woodcut No. 6).

2 Mr. Bell mentions the ruins of a very
similar "Circular Relic-shrine" at Mediri-
giriya, 20 miles north of Polonnaruwa. —

'Annual Report, 1903.'

3 This is probably the Dalada Mali-
gawa, erected by Parakrama Bahu in the
second half of the I2th century. —
' Mahawansa,' ch. 73 and 78.

4 Smither's ' Anuradhapura,' p. 12.

248

BUDDHIST ARCHITECTURE.

BOOK I.

Besides these, there are at Polonnaruwa several of those
groups of pillars, without roofs or walls, which we tried to
describe in speaking of Anuradhapura. One, called the
Audience Hall,1 seems to be very similar to those of the
northern capital ; a vihara, the so - called Heta - Da - ge, close
to the Sat Mahal Prasada, is more extensive, and has been
profusely ornamented ; 2 but no mere description is of much
use, and till we see the plans and more details it is needless
speculating on what they may or may not have been.

Polonnaruwa likewise possesses another point of interest
of considerable importance, though hardly germane to our
present subject. Among its ruins are several buildings in
the Dravidian style of architecture : one of these, miscalled the
Dalada Maligawa, is really a Saiva temple, erected probably
by Nissanka Malla, about A.D. 1190. It is built of granite
and, except the roof and outer mandapa, is in fair preservation.3
Another building, though called the Vishnu Dewala, was also
dedicated to the worship of Siva, as is testified by the presence
of the bull alongside of it, and also apparently on its roof
(Plate No. V.). It is the lowest and flattest of those buildings I
have yet met with, and whilst in general style and carving
resembling the preceding, it is more like a direct literal copy
from a constructive vihara than even the raths at Mamalla-
puram (Woodcut No. 185, p. 329). This may arise either
from its being a copy of an actual vihara existing at the time
it was built, or to its being very old. Those at Mamallapuram,
even if older than this one, may have gone through certain
stages towards their present conventional forms before they
were cut in the rock.4

It is unfortunate for the history of architecture in Ceylon
that the oldest and finest of her rock-cut temples — as those,
for instance, at Dambulla — are only natural caverns, slightly
improved by art; and those mentioned above, as the Isuru-
muniya at Anuradhapura, and Gal Vihara rock temple at
Polonnaruwa, with a recumbent figure of Buddha entering
Nirvana hewn on the rock, besides being comparatively
modern, have little architecture about them, and that little by

1 ' Indian Antiquary," vol. ii. p. 247 ;
Mr. Bell's 'Annual Report, 1900,' p. 9.

2 The proof of Mr. Bell's ' Annual
Report, 1903,' contains a description of
this vihara, but no plan or section is
there given.

1 Nissanka Malla was of a Kalinga
family, and would naturally incline to
the Hindu style of architecture. There

are remains also of a group of Hindil
temples, chiefly of brick, but too much
ruined to be of architectural importance.
4 The Editor is indebted to Lord Stan-
more, G.C.M.G., Mr. James G. Smither,
F.R.I.B.A., and Mr. H. C. P. Bell,
Archaeological Commissioner in Ceylon,
for much valued assistance in the revision
of this chapter.

CHAP. VIII. CONCLUSION. 249

no means of a good class. Generally speaking, what archi-
tecture these Sinhalese caves do possess is developed on applied
facades of masonry, never of the same age as the caves them-
selves, and generally more remarkable for grotesqueness than
beauty. Besides, the form of these caves being accidental, they
want that interest which attaches so strongly to those of India,
as illustrating the religious forms and ceremonies of the early
Buddhists. Indeed, their only point of interest seems to consist
in their being still used for the celebration of the same rites to
which they were originally dedicated 2000 years ago.

There are some interesting ancient bridges, formed on
upright stone pillars, over which stone lintels are placed, and
on these other stone beams, about 5| feet in length from
one lintel to another, to form the road ; but these have no
architectural features worth attention.1

CONCLUSION.

Although the above sketch cannot pretend to be anything
like a complete and exhaustive treatise on the subject, it
may probably be accepted, as far as it goes, as a fairly
correct and intelligible description of Buddhist architecture
in India. We certainly know the beginning of the style,
and as certainly its end. The succession of the buildings
hardly admits of doubt, and their dates are generally ascer-
tained within narrow limits of error. Much has been done,
during the last fifty years, to delineate the numerous examples
of Buddhist architecture known in India, and all that is most
essential to complete the history is now available for the
purpose. It is hardly probable that anything will be now
discovered in India which will materially alter the views put
forward in the preceding pages. Another discovery like that
at Bharaut may reward the industry of explorers ; but even
that, though it has given breadth and precision to our enquiries,
and added so much to our stores of knowledge, has altered
little that was known before. It is difficult, however, to form
an opinion on the chances of any such discoveries being now
made.

But even such a sketch as that contained in the preceding
pages is sufficient to prove that it is almost impossible to
overrate the importance of architecture and its associated
arts in elucidating and giving precision to our knowledge
of Buddhist history and mythology, from the time when it

1 'Third Archaeological Report,' p. 7.

250

BUDDHIST ARCHITECTURE.

BOOK. I.

became the religion of the state till it perished in so far as
India was concerned. In the rails at Bodh-Gaya and Bharaut,
we have a complete picture of Buddhism as it existed during
the great Mauryan dynasty (B.C. 320 to B.C. 180). At Sanchi
and the western caves we have as complete a representation
of the form it took from the 2nd century before our era to the
3rd after it. At Amaravati, and from the Gandhara monasteries,
we learn what modifications had been introduced between our
era and the 3rd century ; and from the Ajanta and later caves
we trace its history downward through its period of decay till
it faded away altogether.

During the first half of this thousand years we have no
contemporary records except those written in stone, and during
the latter we have no books we can depend upon ; but the
architecture, with its sculptures and paintings, remain, and
bear the indelible impress of the thoughts, the feelings, and
the aspirations of those who executed them, and supply us
with a vast amount of exact knowledge on the subject which
is not attainable by any other means now known to us.

ar

138. Capital of a Pilaster, Pitalkhora.

Map of

INDIA

showing the Distribution of

Indo-Aryan, Chalukyan
& Dravidian Styles.

Miles

London,: JvJm. Murray

BOOK II.
ARCHITECTURE IN THE HIMALAYAS.

CHAPTER I.

KASHMIR.

CONTENTS.

Temples — Martand — Avantipur— Buniar — Pandrethan — Malot.

ALTHOUGH neither so beautiful in itself, nor so interesting
either from an artistic or historical point of view as many others,
the architecture of the valley of Kashmir has attracted more
attention in modern times than that of almost any other style
in India, and a greater number of special treatises have been
written regarding it than are devoted to all the other styles put
together. This arises partly from the beauty of the valley in
which the Kashmiri temples are situated. The beauty of its
scenery has at all times attracted tourists to its verdant snow-
encircled plains, and the perfection of its climate has induced
them to linger there, and devote their leisure to the investiga-
tion of its treasures, natural and artistic. In this respect their
fate is widely different from that of temples situated on the hot
and dusty plains of India, where every official is too busy to
devote himself to such a task, and travellers too hurried to
linger for a leisurely and loving survey of their beauties.

Apart, however, from this adventitious advantage, the
temples of Kashmir do form a group well worthy of attention.
When one or two spurious examples are got rid of, they form
a complete and homogeneous group, extending through about
five centuries (A.D. 600 to A.D. 1 100), singularly uniform in their
development and very local, being unlike any other style known

ARCHITECTURE IN THE HIMALAYAS BOOK II.

in India. They have besides this a certain classical element,
which can hardly be mistaken, and is sufficient in itself to
attract the attention of Europeans who are interested in detect-
ing their own familiar forms in this remote valley in the
Himalayas.

The earliest of the modern investigators of the subject were
Messrs. Moorcroft and Trebeck, who visited the valley in
iSip-iSas.1 They were both acute and intelligent observers, but
having no special knowledge of the subject, their observations
on the architecture of the valley do not add much to our know-
ledge of its history.

They were followed by G. T. Vigne in 1833, who being an
artist drew the buildings with wonderful correctness, so as to
bring out the peculiarities of the style, and also to approximate
their history with very tolerable exactness.2 About the same
time Baron Hugel gave his impressions on the subject to the
public, but in a manner much less critical than his predecessors.3

In 1848, Captain (afterwards General Sir) A. Cunningham
published in the ' Journal of the Bengal Asiatic Society,' an
essay on what he called the " Aryan Order of Architecture," but
which was wholly devoted to that of Kashmir.* It was
illustrated by seventeen folding plates, containing map, plans,
elevations, and views, and in fact all that was required for
settling the history of the style, and, but for one or two
unfortunate mistakes, would have left little to be done by his
successors in this field of enquiry.

In 1866, the Rev. W. C. Cowie published in the same
journal an essay on the same subject, as a supplement to
General Cunningham's paper, describing several temples he
had not visited, and adding considerably to our knowledge
of those he had described. This paper was also extensively
illustrated.6

In consequence of all this wealth of literature, very little
remained to be done, when in 1868 Lieutenant Cole, R.E.,
obtained an appointment as superintendent in the Archaeological
Survey of India, and proceeded to Kashmir with a staff quite
sufficient to settle all the remaining outstanding questions.6
Unfortunately, however, Lieutenant Cole had no previous know-

1 ' Travels in the Himalayan Provinces
and in Ladakh and Kashmir,' London,
Murray, 1841.

2 ' Travels in Kashmir, Ladak, Iskardo,
etc., 'two vols. 8vo., London, Colburn,
1842 ; 2nd ed. 1844.

3 ' Travels in Kashmir and the Punjab.'
Translated by Major Jervis, London,
1845.

4 Loc. cit. vol. xvii. part ii. pp. 241-

327-

s< Journal of the Bengal Asiatic Society,
vol. xxxv. pp. 91-123.

6 ' Illustrations of the Ancient Build-
ings in Kashmir, etc., prepared, under
the authority of the Secretary of State
for India in Council,' by Lieut. H. II.
Cole, R.E., quarto, London, 1869.

CHAP. I.

TOMB OF ZAINU-L-'ABIDIN.

253

ledge of Indian antiquities in general, and had not qualified
himself by any special study for the investigation he was
deputed to undertake. All, therefore, he could do was to adopt
blindly General Cunningham's dates, and in this there would
have been no great harm, but, when he came across a temple
which had escaped his predecessor's attention, he arbitrarily
interpolated it into the General's series with a date of his own.
As all these dates are given as if perfectly ascertained, without
any of the reasoning on which they are based, they would, if
accepted, lead to the most erroneous conclusions. Putting
these, however, aside, Lieutenant Cole's plans and architectural
details were a valuable contribution to our knowledge of the
subject, and with his photographs and those now available by
others, enable those who have not had an opportunity of visiting
the valley to form an opinion of their own, and with all these
lights there seems little difficulty in ascertaining all the really
important facts connected with this style.

The first and most misleading mistake that has been made
with reference to Kashmiri architecture, was the assumption
by General Cunningham that the enclosure to Zainu-l-'Abidin's
tomb in 5rinagar originally belonged to an ancient Kashmiri
temple. Lieutenant Cole boldly printed on his plates " probable
date A.D. 400 to 500," a mistake as nearly as may be of 1000
years, as it is hardly doubtful that it was erected for or by the
prince whose name it bears, and who in A.D. 1417 succeeded his
brother — their father being Sikandar, who bore the ill-omened
nickname of Bhutshikan, the idol-breaker.1 As will be seen from
the woodcut (No. 139), it consists of a series of small pointed
arches in rectangular frames, such as are very frequently found

in Muhammadan art, and, though it occupies ,~; , ,.„.,,., HII IMI

the site of an early temple, and parts may
be much older, the peculiarities of the
gateway and other parts are just such as

are found in all contemporary Moslim art I

in India. All the mosques and tombs for |f;
instance at Ahmadabad, A.D. 1400-1 S72, are

f •,..,, i /• ,11- T39- TombofZamu-l-'Abidin.

made up of details borrowed from the archi- Elevation of Arches.

tecture of the Hindus and Jains, and the (

bases of their minarets and their internal pillars can only be

distinguished from those of the heathen by their position, and

by the substitution of foliage for human figures in the niches or

places where the Hindus would have introduced images of their

gods.

1 He boasted of having demolished all
the temples in Kashmir. The tomb of
his queen is constructed on a base, and

with materials of Hindu shrines. —
'Calcutta Review,' vol. liv. (1872),
p. 27.

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

In this instance, however, there is no incongruity, no
borrowed features ; every stone was carved for the place where
it is found. There are niches, it is true, on each side of the
gateway, like those found at Martand and other pagan temples ;
but like those at Ahmadabad they are without images, and the
arch in brick which surmounts this gateway is a radiating arch,
which appears certainly to be integral, but, if so, could not
possibly be erected by a Hindu.1 With the knowledge we now
possess, it is not likely that any one can mistake the fact, that
this enclosure was erected in its present form, by the prince
whose name it bears, to surround his tomb, in the Muhammadan
cemetery of the city in which it is found.

Assuming this for the present, it gives us a hint as to the
age of the other anomalous building in Kashmir — the temple
that crowns the hill, called the Takht-i-Sulaiman, near the
capital. Inside the octagonal enclosure that surrounds the
platform on which the temple stands is a
range of arches (Woodcut No. 140), similar
to those of the tomb of Zainu-l-'Abidin
(Woodcut No. 139), not so distinctly pointed,
nor so Saracenic in detail, but still very
nearly resembling them, only a little more
debased in style. At the bottom of the steps
is a round-headed doorway, not, it is true,
surmounted by a true arch, but by a curved lintel of one stone,
such as are universal in the Hindu imitations of Muhammadan
architecture in the i/th and i8th centuries. The same is the
case in the small temples alongside, which are evidently of the
same age.2 The temple too, itself, is far from having an
ancient look. The one most like it, that I am acquainted with,
is that erected by Chait Singh of Benares (17701781) at
Ramnagar, at the end of the i8th century. I know of no
straight-lined pyramid of a much older date than that, and no
temple with a polygonal plan, combined with a circular cell, as
is the case here, that is of ancient date.3 The cell itself with
the Linga is undoubtedly quite modern ; and the four pillars in
the cell, with the Persian inscriptions upon them, are avowedly
of the 1 7th century. It is suggested, moreover, that they belong
to a repair ; my conviction, however, is, from a review of the
whole evidence, that the temple, as it now stands, was

140. Takht-i-Sulaimin.

Elevation of Arches.

(From a Drawing by

Lieut. Cole.)

1 I cannot make out the span of this
arch. According to the rods laid across
the photograph (No. 4) it appears to be
15 ft. ; according to the scale on the
plan, only half that amount.

2 Lieut. Cole's plates, 1-4.

3 The polygonal basement, however, is
constructed of remarkably massive blocks
and without mortar, and must thus be
relegated to an earlier period. — Stein's
' Rajatarangini, ' vol. ii. p. 290.

CHAP. I.

KASHMIRI TEMPLES

255

commenced by some nameless Hindus, in honour of Siva.,
during the tolerant reign of Jahangir, and that the building
was stopped at the date engraved on the staircase, A.H. 1069
(A.D. 1659), the first year of the reign of the bigot Aurangzeb.
It was then unfinished, and has consequently remained a ruin
ever since, which may give it an ancient look, but not such as to
justify any one putting it 1879 years before what seems to be
its true date.

If we may thus get rid of these two anomalous and
exceptional examples, the history of all the remaining temples
in the valley is more than usually homogeneous and easily
intelligible. The date of the principal example — the temple at
Martand — is hardly doubtful (A.D. 750) ; and of the others,
some may be slightly older, but none can be carried further
back than the reign of Ranaditya, in the 6th century, if the
temple founded by him at Simharotsika still exist.1 Nor can
any one be brought down below, say 1000, which is the latest
date we can possibly assign to that of Payer.2 Between these
dates, with a very little local knowledge, the whole might easily
be arranged. Such a classification is, however, by no means
necessary at present. The style during these six centuries is so
uniform that it may be taken as one, for the purposes of a
general history.

TEMPLES.

Before proceeding to speak of the temples themselves, it
may add to the clearness of what follows if we first explain
what the peculiarities of the style are. This we are able to
do from a small model in stone of a Kashmiri temple
(Woodcut No. 141), which was drawn by General Cunning-
ham ; such miniature temples being common throughout India,
and copies of their larger prototypes.

The temple in this instance is surmounted by four roofs
(in the built examples, so far as they are known, there are
only two or three), which are obviously copied from the
usual wooden roofs common to most buildings in Kashmir,
where the upper pyramid covers the central part of the
building, and the lower a verandah, separated from the centre
either by walls or merely by a range of pillars.3 In the

1 Stein's 'Rajatarangini,' bk. iii. v.
462, and note ; also note on vv. 453-454.

2 Vigne regarded this temple as more
modern than any of the others, whilst
Cunningham ascribed it to the end of the
5th century. Vigne called the village
Payech, which has been followed by

subsequent writers; the real name is
Payer — it is in the pargana of Savur. —
Loc. cit. vol. ii. p. 473.

3 See drawing of mosque by Vigne,
vol. i. p. 269 ; and also ' Journal of the
Asiatic Society of Bengal," vol. xvii.
(1848) pt. ii. p. 253, containing General

256

ARCHITECTURE IN THE HIMALAYAS BOOK II.

wooden examples the interval between the two roofs seems to
have been left open for light and air ; in the
stone buildings it is closed with ornaments.
Besides this, however, all these roofs are
relieved by dormer windows, of a pattern
very similar to those found in mediaeval
buildings in Europe ; and the same steep,
sloping lines are used also to cover door-
ways and porches, these being virtually a
section of the main roof itself, and evidently
a copy of the same wooden construction.

The pillars which support the porticoes,
and the one on which the model stands, are
by far the most striking peculiarity of this
style, their shafts being so distinctly like
those of the Grecian Doric, and unlike any-
thing of the class found in other parts of
India. Generally they are from three to four
diameters in height, diminishing slightly
towards the capital, and adorned with sixteen
flutes, rather shallower than those of the
Grecian order. Both the bases and capitals
are, it is true, far more complicated than
would have been tolerated in Greece, but at
Paestum and in Rome we find, with the Doric
order, a complexity of mouldings by no

* -!'.«•* •' means unlike that found here. These

i. Model of Temple peculiarities are still more evident in the
in Kashmir. annexed representations of two pillars, one
found in Srinagar (Woodcut No. 142),
which is a far more highly ornamented example than the last,
but equally classical in its details, and, if anything, more
unlike any known examples of true Hindu architecture.
The other (Woodcut No. 143) is from Shadipur, and is
perhaps more modern ; the diameter of the pillar is 13^ in.,
and the upper fillet of the abacus is 2oJ in. square.
Nowhere in Kashmir do we find any trace of the bracket
capital of the Hindus, nor of the changes from square to

OROUNO LEVEL

Cunningham's paper on the subject, from
which this woodcut is taken. These
miniature models of temples occur here
and there throughout Kashmir : on the
Pir-Panjal road between .Supiyan and
Ramuh, is one ; another is at Kohil near
Payer ; a third is built into the embank-
ment of the Nali Mar canal ; two in the
.Snnagar lake that are often submerged ;

and one photographed by Major Cole
(' Illustrations of Ancient Buildings,' No.
44), near the Jami Masjid. In these
there is an interior cell scarcely a foot
square ; but near the village of Pattan are
two such models which are not hollowed
out, the place of the doorway being
represented by a small carved panel. —
' Calcutta Review,' vol. liv. p. 26.

CHAP. I.

KASHMIRI TEMPLES.

257

142. Pillar at Srinagar. (From a Draw-
ing by W. Carpenter, Esq.)

octagon, or to the polygon of sixteen sides, and so on. Now
that we are familiar with the
extent of classical influence that
prevailed in Gandhara (ante, p.
217) down to about the 5th
century, we have no difficulty
in understanding whence these
quasi-Grecian forms were de-
rived, nor why they should be
found so prevalent in this valley.
It adds, however, very consider-
ably to our interest in the
subject to find that the civilisa-
tion of the West left so strong
an impress on the arts of this
part of India that its influence
can be detected in all the
Kashmiri buildings down to
the time when the local style
perished under Muhammadan
influence in the I4th century.
Although, therefore, there can
be no mistake about the forms
of the columns in the archi-
of

tecture of Kashmir being
derived from the classical
styles of the West, and as
little doubt as to the countries
through which it was introduced
into the valley, it must not be
overlooked that the classical
influence is fainter and more
remote from its source in
Kashmir than in Gandhara.
Nothing resembling the Cor-
inthian capitals of the Jamal-
garhi monastery are found in the
valley. The classical features in Kashmir are in degree more
like those of the M^nikyala tope and the very latest examples in
the Peshawar valley. The one style, in fact, seems to com-
mence where the other ends, and to carry on the tradition
for centuries after it had been lost in the country from which
it was introduced.

The fact, however, of a quasi-Doric,~order being currently
used in the valley from the 8th to the 1 2th century, renders it
probable that if remains of greater antiquity had been preserved,
VOL. I. R

143. Capital from Shadipur. (From a
Drawing by F. H. Andrews, Esq.)

258

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

we should have found that it was introduced at a much earlier
period, or about coeval with the appearance of the Corinthian
order of the Gandhara monasteries. As both were evidently
derived from the same source, it seems most unlikely that there
should be any break in the continuity of the tradition.

No example of the Doric order has yet been found in
Gandhara, but, as both Ionic and Corinthian capitals have
been found there, it seems more than probable that the
Doric existed there also ; but as our knowledge is still some-
what limited, we ought not to be surprised at any deficiencies
in our series that may from time to time become apparent.

There is still one other peculiarity of this style to account
for. This is the trefoiled arch, which is everywhere prevalent,
but which is not to be accounted for by any constructive
requirement. Now at Takht-i-Bahai and at 'Alt Masjid we
meet with trefoiled arches, over niches for sculptures,1 and in

144. Restoration of Vihara Cells on the west side of the Court of
Takht-J-Bahai Stupa.a Scale about 12 ft. to i in.

the gablets from Gandhara stupas, such as that represented
in Woodcut No. 123 (p. 216), the frames are of this form.
And round the stupa court at Takht-f-Bahai most of the
small shrines or cells were roofed by a sort of double dome
— a smaller one set upon a larger and flatter one (Woodcut
No. 144) ; and if we conceive a vertical section made of one
of these cells, it will be perceived that the outline would be
just that of Woodcut No. 123, or such a trefoil as is every -

1 Foucher, ' L'Art Greco-Bouddhique
du Gandhara,' tome i. p, 199, fig. 80,
and pp. 19, 201, figs. 2 and 81.

2 From Foucher, ut sup., p. 126 ;

modified from Cunningham, ' Archaeo-
logical Reports,' vol. v. plate 9. — Conf.
'Ancient Monuments of India,' part i.
plate 69.

CHAP. I.

MARTAND.

259

where prevalent in Kashmir. Or, if we refer to Woodcut
No. 60 or to 72,1 the section of the cave at Ajanta, which it
represents, affords a similar outline ; and, as in Kashmir and
everywhere else in India, architectural decoration is made up
of small models of large buildings applied as decorative features
wherever required, it is by no means improbable that the
trefoiled facade may have been adopted in Kashmir as currently
as the simple horse-shoe form was throughout the Buddhist
buildings of India Proper. All these features, however, mark
a local style differing from anything else in India.

MARTAND.

By far the finest and most typical example of the Kashmiri
style is the temple of Martand, situated about 5 miles east
of Islamabad, the old capital
of the valley. It is the archi-
tectural lion of Kashmir, and
all tourists think it necessary to
go into raptures about its beauty
and magnificence, comparing
it to Palmyra or Thebes, or
other wonderful groups of ruins
of the old world. Great part,
however, of the admiration it
excites is due to its situation.
It stands well on an elevated
plateau, from which a most ex-
tensive view is obtained, over
a great part of the valley. No
tree or house interferes with
its solitary grandeur, and its
ruins — shaken down apparently
by an earthquake — lie scattered
as they fell, and, unobscured

by vegetation, they are the
most impressive remains of early

145. Temple of Martand. (From a Draw-

Kashmir architecture; nor are
they vulgarised by any modern accretions. Add to this the
mystery that hangs over their origin, and a Western impress
on its details — unusual in the East — but which calls back the

1 See also Woodcut No. 80. On the
Toran attached to the rail at Bharaut
are elevations of chaitya halls, shown

in section, which represent this trefoil
form with great exactness. — Cunningham,
' Stflpa of Bharhut,' plates 6 and 9.

26o ARCHITECTURE IN THE HIMALAYAS. BOOK II.

memory of familiar forms and suggests memories that throw
a veil of poetry over its history more than sufficient to excite
admiration in the most prosaic spectators. When, however, we
come to reduce its dimensions to scale (Woodcut No. 145),
and to examine its pretensions to rank among the great
examples of architectural art, the rhapsodies of which it has
been the theme seem a little out of place.

The temple itself (Woodcut No. 146) is a very small build-

view of Temple at Martand, from the East. (From a Photograph.)

ing, being only 60 ft. in length by 36 ft. in width. The width of
the west facade, however, is eked out by two wings or adjuncts,
which make it about 60 ft, thus making its length and breadth
about equal. General Cunningham also estimated its height,
when complete, at 60 ft. — making the three extreme dimensions
equal ; but this is only conjectural. £T"3I

The roof of the temple has so entirely disappeared that
Baron Hugel doubted if it ever possessed one. General
Cunningham, on the other hand, had no doubts on the subject,1

'Journal of the Asiatic Society of Bengal,' vol. xvii., Sept. 1848. p. 267.

CHAP. I.

MART AND.

261

and restored it in stone on his plate No. 14. The absence,
however, of any fragments on the floor of the temple that
could have belonged to the roof, militates seriously against
this view; and, looking at the tenuity of the walls and the
large voids they include, I doubt extremely if they ever could
have supported a stone roof of the usual design. When, too,
the plan is carefully examined, it will be seen that none of

147. Central Cell of Court at Martand. (From a Drawing by Gen. A. Cunningham.)
Scale, 10 ft. to i in.

the masses are square ; and it is very difficult to see how the
roof of the porch could, if in stone, be fitted to that over the
cella. Taking all these things into consideration, my impres-
sion is, that its roof — it certainly had one — was in wood ; and
knowing how extensively the Buddhists used wooden roofs for
their chaitya halls, I see no improbability of this being the
case here at the time this temple was erected.

The courtyard that surrounds and encloses this temple is,
in its state of ruin, a more remarkable object than the temple
itself. Its internal dimensions are 220 ft. by 142 ft.,1 which
are respectable, though not excessive ; they are not much

1 Cunningham in the ' Journal of the Asiatic Society of Bengal,' vol. xvii.
(1848), pt. ii. p. 269.

262 ARCHITECTURE IN THE HIMALAYAS. BOOK II.

more than those of the temple of Neminath at Girnar (Wood-
cut No. 280), which are 165 ft. and 105 ft., though that is by
no means a large Jaina temple. On each face is a central cell,
larger and higher than the colonnade in which it is placed
(Woodcut No. 147), but even then only 30 ft. in height to
the summit of the roof, supposing it to be completed, and
the pillars on each side of it are only 15 ft. high, which are
not dimensions to go wild about, though their strongly-impressed
Grecian aspect is certainly curious and interesting.

General Cunningham broached " a suspicion that the whole
of the interior of the quadrangle was originally filled with water
to a level within one foot of the bases of the columns, and that
access to the temple was gained by a raised pathway of slabs,
supported on solid blocks at short intervals, which connected
the gateway flight of steps with that leading to the temple.
The same kind of pathway must have stretched right across
the quadrangle from one side doorway to the other. Similar
pathways still exist in the Shalimar gardens, as passages across
the different reservoirs and canals. On the outside of the
quadrangle, and close by the northern side of the gateway,
there is a drain by which the surplus water found its exit,
thus keeping the surface always at the same level. The temples
at Pandrethan Ledari, and in the Baramula Pass, are still
standing in the midst of water. A constant supply of fresh
water was kept up by a canal or watercourse from the River
Lambadari, which was conducted alongside of the mountain
for the service of the neighbouring village of Simharotsika," etc.
" The only object," the General goes on to remark, " of erecting
temples in the midst of water must have been to place them
more immediately under the protection of the Nagas, or human-
bodied and snake-tailed gods, who were zealously worshipped
for ages throughout Kashmir."1 But for this hypothesis there
is no sufficient basis, for there are no springs on the arid plateau
where the temple stands, and the old irrigation canal from the
Lidar could not have served the purpose. Moreover, the temple
was undoubtedly dedicated to Surya-Narayan or Vishnu-Surya ;
and the polycephalous snake-hoods over some of the abraded
figures on the walls are only indicative of Surya or Vishnu.2

The ' Rajatarangini ' distinctly states that the " wonderful
temple of Martanda with its massive walls of stone within a

1 Loc. fit. p. 273. Temples were very by the rise of their surroundings,

frequently placed beside springs (Nagas), a As an example we may refer to the

which were enclosed in separate walled j figure of Vishnu in Cave III. at Badami. —

basins. — Stein, in 'Vienna Oriental j ' Archaeological Survey of Western India,'

Journal,' vol. v. p. 347. The Pandrethan I vol. i. plates 25 and 30 ; or Le Bon,

and Baramula temples have been flooded i ' Les Monuments de 1'Inde,' p. 149.

CHAP. I.

MARTAND.

263

lofty enclosure" was built by King Lalitaditya-Muktapida,1

who ruled about A.D. 725 to 76o.2 Jonaraja, a Kashmir

chronicler of the first half

of the 1 5th century, tells

us that Sikandar Shah

Bhutshikan (1393 - 1416)

destroyed the image, and

probably he wrecked the

temple itself. The court,

however, had been used as

a fortification in Jaya-

simha's reign, A.D. 1128 to

"49-

Unfortunately, the stone

of which the temple is built
is of so friable a nature
that the sculptures are now
barely recognisable, but, so
far as can be made out
from such photographs as
exist, the principal figures
in the niches have snake-
hoods, which are recognised
adjuncts of certain forms
of Vishnu (Woodcut No.
148). Any one on the spot,
with a competent know-
of Hindu mythology, could
determine the character of 148.
these sculptures ; but no
one has yet visited it with
the preparation necessary to settle this and other uncertain
points regarding the architecture and mythology of the place.
A monograph, however, of this temple would be a work well
worthy of any pains that might be bestowed upon it by any
Indian archaeologist ; for, besides its historical and mythological
importance, many of its details are of great beauty, and they
have never been drawn with the care they so well merit (Wood-

Niche with Figure at Martind.
(From a Photograph.)

1 Gen. Cunningham, loc. cit. p. 263,
misinterpreting this passage (' Raja-
tarangini,' bk. iv. v. 192, and iii. v. 462),
ascribed the temple to Ranaditya, and the
enclosure only to Lalitaditya. Simharot-
sika, where Ranaditya built a temple to
Martanda in the 6th century, cannot now
be identified.

a Stein, ' Rajatarangint,' vol. i. introd.,

p. 88, and bk. iv. ver. 192. Lalita-
ditya was a great patron of Vaishnavism,
but he also countenanced Buddhism,
which flourished in Kashmir in the
8th century, and built a vihara at
Parihasapur with a colossal image of
Buddha, and another vihara and a
stupa at Hushkapur, bk. iv. verses 188,
200, 203.

264

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

cut No. 149). As

149. Soffit of Arch at Martand.

(From a Sketch by the late

Mr. Wilson, B.C.S.)

the typical example of a quasi-classical
style, a perfect knowledge of its
peculiarities would be a landmark
in the history of the style both
before and after its known date.

The site of the ancient city of
Parihasapura, where Lalitaditya
erected four Vaishnava temples
and a Buddhist vihara, was dis-
covered by Dr. Stein, in 1892, near
the village of Divar, with the ruins
of half-a-dozen temples, said to
have been destroyed by Sikandar
Shah — the remains of the spacious
courts that had surrounded them
being still traceable. In 1896,
many of the stones, till then in
situ, had been removed and broken
up by contractors for road-metal.1

AVANTIPUR.

Next in importance to Martand, among Kashmiri temples,
are those of Avantipur, now Vantipor, on the right bank of the
Jehlam, halfway between Srinagar and Islamabad, all erected
certainly within the limits of the reign of Avantivarman, the first
king of the Utpala dynasty, and who reigned from A.D. 855 to
A.D. 883. The stone with which they are erected is so friable,
and the temples themselves are so ruined, that there might be a
difficulty in ascertaining to what religion they were dedicated if
the ' Rajatarangini ' were not so distinct in describing this monarch
as conducting himself as a follower of Siva, whilst he had been
brought up as a worshipper of Vishnu,2 and naming these temples
as dedicated, one — built before his accession — to the latter, and
after that event, the temple of Avantmvara to Siva.

The two principal ruins stand in courtyards of nearly the
same size, about 200 ft. by 160 ft. or 170 ft. internally. One
has pillars all round, like Martand, and almost identical in
design and dimensions. The other is astylar, but the temple
itself was much more important than in the first example.3

1 Stein's ' Rajatarangini,' vol. ii. pp.
300-303.

Ibid., bk. v. verses 43-45.

3 Plans of these temples with details
are given by Cunningham, plates 1 7 and
1 8, and by Lieut. Cole with photographs,

plates 20 to 27, and 2 to 5 for details.
Mr. Cowie also adds considerably to
our information on the subject. The
dimensions quoted in the text are from
Lieut. Cole, and are in excess of those
given by General Cunningham. The

CHAP. I.

AVANTIPUR.

265

The central shrines of both have been reduced to heaps of
stones, and it is now impossible to determine which was the
Vaishnava and which the Saiva shrine. Of the smaller temple,
owing to part of the court having been long since silted up,
there are more remains than of the other, from which every
pillar has been removed, possibly by Shah Jahan and other
Mughal emperors, for their summer palaces and Shalimar
gardens near Srinagar. Portions of the gateways of both still
remain.1

The characteristic that seems most clearly to distinguish the
style of the temples at Martand from that of those at Avantipur
is the greater richness of detail which the
latter exhibit ; just such a tendency, in fact,
towards the more elaborate carvings of the
Hindu style as one might expect from their
difference in date. Several of these have
been given by the three authors to whose
works I have so often had occasion to allude,
and to which the reader is referred ; but the
annexed fragment (Woodcut No. 1 50) of one
of its columns is as elegant in itself, and
almost as interesting historically, as the
Doric of the examples quoted above, inas-
much as if it is compared with the pillars
of the tomb of Mycenae 2 it seems difficult
to escape the conviction that the two forms
were derived from some common source.
At all events, there is nothing between the Peloponnesos and
Kashmir, so far as we now know, that so nearly resembles it.

At -Sankarapura, now Patan, between Srinagar and Baramula,
Sankaravarman (A.D. 883-902) the son and successor of Avanti-
varman, with his Queen Sugandha, erected two -Saiva temples
which still exist, though the corridors that doubtless once
enclosed their courts have disappeared. Like most other
Kashmiri temples they consisted only of a shrine or vimana,
without mandapa, but had recessed porches forming small
chapels on three sides. 5ankaravarman is said to have brought
the materials for his buildings here from Parihasapura, about
7 miles off.3

150. Pillar at Avanti-
pur. (From a Draw-
ing by Mr Wilson,
B.C.S.)

latter gives the dimensions inside the
court of the one as 191 ft. by 171 ; and
of the other as 172 ft. by 146^. The
second is in the village, and he proposed
to identify it with the Avantiswamin
temple and the first, about half a mile
to the north-west, as the Avantuvara or
Saiva temple.

1 Cunningham, loc cit. pp. 276 et seqq.,
and plate 17; Bernier's 'Travels A.D.
1656-1668' (ed. 1891), p. 400.

2 ' History of Ancient and Mediseval
Architecture,' vol. i. Woodcut 125, p.
244.

3 Gen. Cunningham has given de-
scriptions and outline plans of these

266

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

BHANIYAR OR B£TNIAR.

At a place near the remote village of Buniar or Bhaniyar,
on the road between Ort and Naoshahra, there stands one
of the best-preserved temples in the valley. During long ages
of neglect, silt and mud had so accumulated as to half bury
the place. It was, however, excavated a good many years
ago by orders of the Maharaja, and hence its nearly perfect

151. View of Court of Temple at Buniar. (From a Photograph.)

state.1 Its dimensions are less than those of the temples last
described, the court being only 145 ft. by 120 ft, but, except
from natural decay of the stone, it is nearly perfect, and gives
a very fair idea of the style of these buildings. The trefoiled
arch, with its tall pediment, the detached column and its
architrave, are as distinctly shown here as in any other existing
example of a Kashmiri colonnade, and present all those quasi-
classical features which we know were inherited from the
neighbouring province of Gandhara. The central temple is
small, only 26 ft. square over all ; the cell is 13^ ft. square inside
with walls over 6 ft. thick, supported on a basement 4 ft. high,
and its roof is now covered with wooden shingles ; but whether
that was the original covering is not certain. Looking, however,
at the central side-cell of the colonnade (Woodcut No. 151),

temples. — Loc. cit. pp. 282, 283, and
plate 20 ; Cole, ' Illustrations of Ancient
Buildings,' plates 28-35 '• conf- ' R&jatar-
angini,' bk. v. verses 156-158, and 161.
rCole, 'Ancient Buildings,' p. 23,

and plates 37 and 38. Another ruined
temple, but far more decayed, is on the
road between Buniir and Uri. — Stein,
' Rajatarangini,' vol. ii. p. 404.

CHAP. I.

BUNIAR.

267

it seems doubtful whether General Cunningham was justified
in restoring the roof of the temple, or of the central cell at
Martand in stone. My impression rather is, as hinted above,
that the temple-roof was in wood ; that of the side-cell in stone,
but flat.

At a place called Waniyat or Vangath — 32 miles from
.Srinagar, near the sacred Haramukh peaks — are two groups
of temples, together about seventeen in all, which were carefully
examined and described by the Rev. Mr. Cowie,1 and plans
and photographs are found in Lieutenant Cole's book.2 They
differ somewhat from those we have been describing, inasmuch
as they do not seem to have been enclosed in colonnaded courts,
and each group consists of one large and several smaller
temples, unsymmetrically arranged. The larger ones are 30 ft.
and 32 ft. square in plan over all ; the smaller 10 ft. or 12 ft.
They are of various ages, and the two principal temples are
most probably those of Bhute-ra in the east group, and Jyeshtha
in the other.3

There are no inscriptions, nor any historical indications
that would enable us to fix the date of the Waniyat temples
with certainty, and the stone has decayed to such an extent
that the details cannot be defined with the precision necessary
for comparison with other examples ; but whether this decay
arises from time or from the nature of the stone there are no
means of knowing.4 This Ttrtha at Haramukh was famous
from very early times, and we learn that Lalitaditya-Muktapida
built here a stone temple to Jyeshtha in the 8th century, and
made gifts to the Bhute.ya temples. The Jyeshtha shrine is
thus probably among the earliest. Early in the nth century
the temples were plundered, after which they were probably
restored and modernised by Uchahala (A.D. noi-nii), and
again plundered by hillmen before 1150. They would almost
certainly suffer also at the hands of Sikandar Shah at the end
of the 1 4th century.

Among the remaining examples, perhaps the one that
most clearly exhibits the characteristics of the style is that
at Pandrethan, about 3 miles from Srinagar (Woodcut No. 152).
It still is a well-preserved little temple, standing in the middle
of the village, and is in all probability the Vaishnava temple
built during the reign of King Parthva (A.D. 906-921) by his

1 'Journal of the Asiatic Society of
Bengal,' vol. xxxv. (1866), pp. 101 etseqq.

2 ' Illustrations of Ancient Buildings in
Kashmir,' pp. iiff., plates 6 to II.

3 Stein's ' Rajatarangini,' bk. v. vv. 55-
59 and note, and bk. i. v. 107, note.

4 Lieut. Cole, basing his inferences on
certain similarities he detected between
them and the temple on the Takht-i-
Sulaiman, which he believed was erected
B.C. 220, ascribed their erection to the
first century after Christ.

268

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

minister, and named Meruvardhana-swamin.1 It now stands

in the water of a
shallow pool that
occupies the former
courtyard. Origin-
ally, it seems to have
had a third storey or
division to its roof,
but that has fallen ;
the lower part of the
building, however, ex-
hibits all the charac-
teristic features of the
style in as much per-
fection as almost any
other known example.
It consists of a shrine
only, 1 1 ft. 7 in. square
inside, with doors on
all four sides, and
the inside roof covered
with sculpture.

One last example
must conclude our

152. Temple at Pandrethan. illustrations of Kash-

(From a Drawing by Gen. Cunningham.) ... . . — ,

mm architecture. The

temple at Payer, though one of the smallest, is among the
most elegant, and also one of the most entire examples of
the style (Woodcut No. 153). Its dimensions are only 8 ft.
square for the superstructure, with a door on each side, and
21 ft. high, including the basement; but with even these
dimensions it acquires a certain dignity from being erected
with only six stones — four for the walls and two for the
roof.2 It stands by itself on a knoll, without any court, or any
of the surroundings of the older temples, and is dedicated to
Siva. It would be interesting if its date could be ascertained,
as it carries with it that of the cave of Bhaumajo or Bumazu, and
of several other temples.3 Vigne, from its perfect preservation

1 ' Rajatarangini,' bk. v., v. 267;
Moorcroft, ' Travels,' vol. ii. p. 240 ;
Hiigel, ' Kaschmir,' Bd. i., S. 260;
Vigne, ' Travels,' vol. ii. p. 38 ; Cun-
ningham, ' Journal of the Asiatic Society
of Bengal,' vol. x vii. , pt. ii. pp. 283 et seqq. ,
and plate 21 ; Foucher, ' L'Art Greco-
Bouddhique,' pp. 143-146.

" Cunningham, he. cit. p. 256; Growse

says ' ten stones,' adding four for the
tympana over the doors.

3 'Journal of the Asiatic Society of
Bengal,' vol. xxxv. pp. IOO et seqq. ; and
Stein's 'Rajatarangini,' bk. vi., vv. 177,
178, note. It is probably the same as
the temple of Bhimakejava built (950-958
A.D.)byBhima .Sahi, King of Kabul or
Udabhanda.

CHAP. I.

TEMPLE AT PAYER.

269

regarded it as the most modern of the series, and Cunningham,
founding on a mistake,1 argued for its being a temple erected
by Narendraditya, before the end of the 5th century. It is

153-

Temple at Payer. (From a Photograph.)

evidently much later than Martand, and may be ascribed,
architecturally, with some confidence to the loth century.

In order to write a complete monography of the Kashmiri
style, we ought to be able to trace it to a much earlier period.
We know that it was influenced by that of Gandhara, from
which it was not distant, and from which it was subject to
invasion.

There were among the Gandhara viharas as we now know,
buildings with sloping stone roofs, and that all along the
Himalayas, these double and triple roofs, sloping to four sides —

' Rajatarangini,' bk. iii. ver. 383 ;
the temple of Narendraswamin referred
to was undoubtedly Vaishnava, and this
upsets his argument, for the Payer temple

is .Saiva ; svdmin is the invariable ter-
mination of the names of Vaishnava
shrines in Kashmir.

270

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

known as "for four waters" — have always been in common
use.1

We learn from tradition that in the time of A.yoka, B.C. 250,
missionaries were sent to convert the inhabitants of the valley
to the Buddhist faith, and that, at a later date, the Turushka
king Kanishka ruled over Kashmir, and was a patron of the
Buddhist religion ; and we know that in the 7th century Hiuen
Tsiang found Buddhism, if not the only religion, at least one of
the dominant cults of the people. The details he mentions,
and the fact of his lingering here for two whole years (A.D. 633
and A.D. 634) to study its forms and scriptures, proves how im-
portant this religion then was.2 More than a century later (A.D.
759), U-k'ong, another Chinese Buddhist, reached Kashmir, and
spent fully four years in literary study, and visiting the sacred sites
and monasteries, stating that he found more than three hundred
convents, and many stupas.3 But scarcely a vestige of a chaitya
or of a vihara has yet come to light ; and though there are
mounds which may contain stupas, it is most improbable
that they will contain any architectural forms that may be
of use for our purposes.4 All the ancient monasteries seem
to have been destroyed on the decline of the religion.

We now know sufficiently the forms and age of the Gandhara
monasteries (ante, pp. 211 et seqq.) to supply most of the
missing links connecting the Kashmiri style with that of the
outer world ; but till the temples in the Salt Range, and other
little- frequented parts of the Panjab are examined, we shall not
know all that we desire. Meanwhile the annexed woodcut
(No. 154), representing a temple at Malot, in the Jehlam district,
shows how nearly the Panjabi style resembled that of Kashmir.5
There are the same trefoil-headed openings ; the fluted pillars,
with quasi-classical bases and capitals ; and a general similarity
of style not to be mistaken. There is another temple very
similar, but smaller, at Katas, 1 2 miles north-west from Malot ;

1 Foucher, ' L'Art Greco-Bouddhique
du Gandhara,' pp. 116, 131-136.

2 Beal, ' Life of Hiuen Tsiang,' pp.
68-72 ; Julien, • Vie de Hiouen Thsang,'
pp. 91-96.

3 Dr. Stein has succeeded in identify-
ing the localities referred to by U-k'ong
in his ' Notes on Ou-k'ong's Account,
of Ka9mir' (Wien, 1896). See also
4 L'Itin6raire d'Ou-k'ong (751 - 790)
traduit et annote par MM. Sylvain Levi
et Ed. Chavannes ' in 'Journal Asiatique,'
ix. s6r. tome vi. (1895), pp. 341-384.

* Near Ushkiirthe ancient Hushkapura,
on the Jehlam, opposite to Barimula was

a Buddhist stupa— still intact till 1882—
when, under Gen. Cunningham's in-
structions, one of his assistants dug into
it and razed it to the ground. It had
been constructed along with a vihara, by
Lalitaditya early in the 8th century. —
Lawrence's ' Valley of Kashmir,' p. 163 ;
Stein's ' Rajatarangini,' bk. iv., note on
v. 1 88.

5 It appears from Hiuen Tsiang,
that in the 7th century the northern
Panjib was subject to Kashmir. — Beal,
' Buddhist Records,' vol. i. pp. 136, 143,
147, and 163 ; ' Life of Hiuen Tsiang,'
p. 192.

CHAP. I. KASHMIRI TEMPLES. 271

both are near Find Dadan KMn, and there are others in the
neighbourhood, which may form a connecting link between the
Kashmiri temples and other varieties of style in the Himalaya
region.

So many and so various are the points of interest connected
with the style of the ancient buildings in Kashmir, that they
deserve much fuller illustration than is compatible with the

Temple at Malot, in the Salt Range. (From a Photograph.)

scope of the present work. Though not magnificent, they are
very pleasing and appropriate examples of art, and they have
this advantage over most of the Indian styles, that Kashmir
possesses, in the ' Rajatarangini,' what may be said to be the
only Indian history in existence,1 and Dr. A. M. Stein's
admirable edition and translation of that work has done much
to fix the dates of many of the buildings, and to supply a basis
for a scientific and historic treatment of the whole.

1 See ante p. 8.

272

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

The earliest references to the religious beliefs of Kashmir
connect them with the worship of Nagas or serpent deities,
supposed to preside over springs, lakes, and rivers ; hence they
correspond closely with the classical Naiads or Potameids.
Numerous temples were erected to them at all the more famous
springs, and to these the earliest and more popular pilgrimages
were made, and continue to be even till the present time. Even
the Muhammadan Kashmirians pay superstitious and hardly
obscure reverence to them ; the " ziarats " or shrines of their Pirs
or saints are largely fixed at the old sacred spots, and sometimes
they seem to have been the native shrines appropriated by the
ruling caste.1 The Naga divinities were accepted by the
Buddhists and worked into the mythology of the Mahayana
school. Until the 6th century Buddhism was probably the
predominant religion of the country. Mihirakula, a White
Hun — whose coins indicate that he was a Saiva — acquired the
sovereignty about A.D. 530, and was a bitter persecutor of the
Buddhists, at the same time fostering the Brahmanical cult.
When Hiuen Tsiang visited the country in the /th century,
Buddhism seems to have considerably revived. The kings of
the Karkota and Utpala dynasties were tolerant, and, as we
learn, built Buddhist viharas as well as Hindu temples, and
U-k'ong, who reached Kashmir in A.D. 759 — probably in the reign
of Lalitaditya-Muktapida — speaks of the Buddhist establishments
as being numerous and very flourishing.2 By the I4th century,
however, the Hindu rulers had become weak and effete, and a
military adventurer from the south murdered Kota Rani, the
widow of the last sovereign, A.D. 1339, and usurped the legal
power as Shah Mir. The immigration of foreigners that followed
rapidly led the way, under the new Moslim dynasty, to the
general conversion of the people to the Musalman religion, and
by the end of the century this had become an accomplished fact.

As Muhammadanism rose in power the old temples were
either destroyed, as under the iconoclastic zeal of Sikandar
Shah, 1393 to 1416, or they were neglected and fell to ruin;
after that we have only the tomb of Zainu-l-'Abidm and the
temple on the Takht-i-Sulaiman that can be classed as examples
of the style, though the latter can hardly even claim a title to
that affiliation.

1 It seems not improbable that the
Ziarat of Pir Haji Muhammad Sahib at
.Snnagar, may represent the Rana-
swamin temple of Ranaditya, erected
in the 6th century ; the Bumazu temple
also is now regarded as the Ziarat of
Kaba Bamadin Sahib. — ' Rajatarangin!,'
bk. Hi., vv. 453-454, note, and bk.

vi., vv. 177, 178, note.

2 We read of the iconoclast Harsha
(1089-1101) sparing two colossal images
of Buddha : one in .Srinagar, and the
other at Parihasapur — probably that
established by Lalitaditya in the 8th
century. — ' Rajatarangini,' bk. vii.. vv.
1097, sq., and bk. iv., vv. 200.

CHAP. II. NEPAL

273

CHAPTER II.

NEPAL AND TIBET.

CONTENTS.
Stupas or Chaityas — Wooden Temples — Tibet — Temples in Kangra.

ANY one looking at the map, and the map only, would probably
be inclined to fancy that, from their similarity of situation and
surroundings, the arts and archaeology of Nepal must resemble
those of Kashmir. It would not, however, be easy to make a
greater mistake, for there are no two provinces of India which
are more diametrically opposed to one another in these respects
than these two Himalayan states. Partly this is due to local
peculiarities. The valley of Nepal proper — in which the three
old capitals, Patan, Bhatgaon, and Kathmandu, are situated —
is only about 15 miles north and south, by 20 east and west.
It is true, the bulk of the population of the Gurkha state live
in the valleys that surround this central point ; but they are
sparse and isolated communities, having very little communica-
tion with each other. Kashmir, on the other hand, is one of
the most beautiful and fertile valleys in the world, measur-
ing more than 100 miles in one direction and more than 70
in another, without any ridges or interruptions of importance,
and capable of maintaining a large population on one vast,
unbroken fertile plain.

Another point of difference is, that Kashmir never was a
thoroughfare. The population who now possess it entered it
from the south, and have retained possession of it — in all
historical times, at least — in sufficient numbers to keep back
any immigration from the north. In Nepal, on the contrary,
the bulk of the population are of Tibetan, or Mongol origin
from the north, left there apparently in their passage southward ;
and, so far as we can gather from such histories as exist, the
southern races who are found there entered the valley in the
beginning of the I4th century, and never in such numbers as
VOL. I. S

274

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

materially to modify the essentially Turanian character of the
people.

Nepal also differs from Kashmir from the fact that the
Muhammadans never had possession of their valley, and never,
consequently, influenced their arts or their religions. The
architectural history of the two valleys differs, consequently,
in the following particulars: — In Kashmir we have a Buddhist
period, developing by the 8th century into an original quasi-
classical style, that lasted till it, in its turn, was supplanted by
that of the Moslim in the i$th century. In Nepal we have no
succession of styles — no history in fact — for we hardly know
when any of the three religions was introduced ; but what we
find is the Vaishnava, Saiva, and Buddhist religions existing
side by side at the present day, and flourishing with a rank
luxuriance unknown on the plains of Bengal, where probably
their exuberance was checked by the example of the Moslims,
who, as just remarked, had no influence in the valley.

Owing to the principal monuments in Nepal — except the
older chaityas — being modern, and to the people being too
poor to indulge in such magnificence as is found on the plains,
the buildings of Nepal cannot compare, as architectural objects,
with those found in other parts of India. But, on the other
hand, the very fact of their being comparatively modern gives
them an interest of their own, and though it is an exaggera-
tion, it is a characteristic one, when it is said that in Nepal
there are more temples than houses, and more idols than men ;l
it is true to such an extent that there is an unlimited field for
enquiry, and even if not splendid, the buildings are marvellously
picturesque. Judging from photographs and such materials as
are available, I have no hesitation in asserting that there are
some streets and palaces in Kathmandu and Bhatgaon which
are more picturesque, and more striking as architectural com-
positions, than are to be found in any other cities in India.2
The style may be called barbarous, and the buildings have
the defect of being principally in wood ; but their height, their
variety of outline, their wealth of carving and richness of colour,
are such as are not to be found in Benares or any other city
of the plains.

The real point of interest in the architecture of Nepal to

1 The towns of Kathmandu, Patan and
Bhatgaon, which are within a short
distance of one another, are crowded
with sacred edifices — Buddhist, Saiva,
and Vaishnava. The number of these
shrines is estimated at 2000.

- We have now further illustrations

in ' Le Nepal, etude historique d'un
Royaume Hindou,' par Sylvain Levi,
3 tomes, Paris, 1905-1908; Dr. G. Le
Bon, 'Voyage au Nepal' in 'Tour du
Monde,' 1886, ler. sem. ; and ' Les
Monuments de 1'Inde,' pp. 223-246,
and figs. 364-398.

CHAP. II. NEPAL. 275

the true student of the art lies in its ethnographic meaning.
When fully mastered, it presents us with a complete microcosm
of India as it was in the /th century, when Hiuen Tsiang
visited it — when the Buddhist and Brahmanical religions
flourished side by side ; and when the distinctive features of
the various races were far more marked than they have
since become under the powerful solvent of the Muhammadan
domination.

From all these causes I believe that if the materials existed,
and it were possible to write an exhaustive history of the
architecture of the valley of Nepal, it would throw more light
on most of the problems that are now perplexing us than
that of any other province in India. It only, however, can be
done by some one on the spot, and perfectly familiar not only
with the Nepalese buildings, but with all the phases of the
question ; l but even then its value would be more ethno-
graphic than aesthetic. If this were an ethnographic history
of architecture, to which the aesthetic question were subordinate,
it would be indispensable that it should be attempted, however
incomplete the materials might be ; but the contrary being the
case, it must suffice here to point out the forms of the archi-
tecture, merely indicating the modes in which the various styles
are divided among the different races.

Like that of so many other countries of India, the mythic
history of Nepal commences even before the Kaliyug, and
among its pre-historic visitors are mentioned Vipajyi and the
other five Buddhas that preceded Sakyamuni, together with
Manjujn Bodhisattwa, Svayambhu — the self-existent, .Siva as
Pa^upati, Vishnu, and other gods of the Hindu Pantheon.
These do not concern us : tradition adds that A^oka visited
the valley and built five chaityas, one in the centre of Patan
and the others at the four cardinal points round it, which
are still pointed out. We come to historical fact in the
5th century A.D. when we meet with the earliest inscriptions.2
They belong to the later kings of the Lichchhavi dynasty,

1 Nepal is fortunate in having pos- found in Nepal, and the services he
sessed in the late Mr. Brian H. Hodgson rendered to this cause are incalculably
one of the most acute observers that i great. Nor did he neglect the archi-

ever graced the Bengal Civil Service.
At the time, however, when he was
Resident in the valley, none of the
questions mooted in this work can be
said to have been started ; and he was
mainly engrossed in exploring and
communicating to others the unsuspected
wealth of Buddhist learning which he

tecture, as the numerous drawings in
his collections bear witness.

2 The Nepal inscriptions were first
copied and translated by Pandit Bhag-
wanlal Indraji. — ' Indian Antiquary,'
vol. ix. pp. 163-194, and commented
on, vol. xiii. pp. 411-428.

276 ARCHITECTURE IN THE HIMALAYAS. BOOK II.

whose ancestors seem to have come from Vaijali and estab-
lished themselves in Nepal, and who seem to have been some
of them Vaishnavas and others Saivas. Their inscriptions
apparently range from late in the 5th century till into the
7th,1 when Amjuvarman founded a new dynasty, and possibly
employed a Tibetan era in his inscriptions.2

Buddhism had no doubt got a strong foothold among the
Newars at an early date — not improbably in the time of
Asoka. ; but about the end of the 5th century, or soon after,
we hear of the patriarch Vasubandhu in his old age going
on a mission to Nepal vyith 500 disciples, and founding
monasteries and making converts.

The Newars had entered the country from the north, and
were undoubtedly of Tibetan origin.3 Like most of the
Himalayan tribes they were snake - worshippers, and the
Buddhist missionaries who visited them accepted their legends
and made them part of their system. Hindu emigration into
the valley must have begun early, and the kings of the long
dynasty that ended about A.D. 600 all bear Hindu names,
whilst their inscriptions indicate that they worshipped the
Hindu gods. The Anmivarman or Thakuri dynasty were
Vauyas like Harshavardhana, and were succeeded by other
Rajput families. In 1097 Nanyadeva from Tirhut invaded
and subjugated the country, and again in 1324 Harisimha,
of the same race — fearing the invasion of the Muhammadans
under Ghyasu^-din Tughlak — moved up from Simraun in the
Tirai and, overcoming the petty chiefs, assumed the govern-
ment.4 But his dynasty does not appear to have ruled for
a long period, and the four chief towns — Bhatgaon, Banepa,
Patan, and Kathmandu had each their own princes till the
year 1768, when a weak sovereign having called in the assist-
ance of a neighbouring Gurkha Raja, he seized the kingdom,
and his successors still rule in Nepal. They apparently were
originally of the Magar tribe,5 but having mixed with the
immigrant Hindus, call themselves Rajputs, and have adopted
the Hindi! religion, though in a form very different from that
known in the plains, and differing in a manner we would

1 The dates range from 386 to 518 of an era of his own from A.D. 595, or
an undefined era. There are difficulties eleven years before Ilarsha's.
in supposing the 5aka era to be meant, 3 The traditional connection of the

and M. Sylvain Levi assumes a Lichch-
havi era beginning A.D. in. The
inscriptions are in classical Sanskrit, and
testify to the literary culture of the

country at that age. — ' Nepal,' tome ii. 1 p. 324f.

Newars with the Nayyars of Malabar
is only a myth of Brahmanical invention
in order to get over caste difficulties.
4 Sylvain Levi, ' Nepal,' tome ii.

pp. 112, 114

2 M. Sylvain Levi supposes he started the Kingdom of Nepal,' p. 22.

5 Buchanan Hamilton, 'Account of

CHAP. II.

NEPAL,

277

scarcely be inclined to expect. When the religion of the
destroyer was introduced into a country that professed the
mild religion of Buddha, it might naturally be supposed that
its most savage features would be toned down, so as to meet,
to some extent at least, the prejudices of the followers of
the religion it was superseding. So far from this being the
case in this instance, it is said that when first introduced the
gods were propitiated with human sacrifices, till warned in a
dream to desist and substitute animals.1 Besides this, the
images of Durga or Kali, though hideous and repulsive enough
in the plains, are ten times more so in Nepal, where Tantric
rites and sorcery prevail as in Tibet ; and, in fact, throughout
there is an exaggeration of all the most hideous features of
the religion, that would lead to the belief that it found a
singularly congenial soil in the valley, and blossomed with
unusual exuberance there. So far, too, as the architecture of
the Saiva temples in Nepal is concerned, it seems to indicate
that the worship came into the valley from the north, rather
than from the plains of Bengal. The architecture of the temples
of Vishnu, on the contrary, seems evidently to be an offshoot
of the art of the plains.

STCPAS OR CHAITYAS.

The Buddhist chaityas must be regarded as the oldest
monuments in Nepal. Four of them are ascribed to A^oka,
who is said to have visited the valley and built one in the centre
of Patan, and others at the four cardinal points round the city.2
They were not called stupas, since they contained no relics, but
are strictly chaityas or monuments intended to call forth pious
thoughts. The chaityas of the cardinal points still exist intact
in their great outlines ; and their general appearance, as M.
Sylvain Levi remarks, does not contradict the tradition : — a
hemispherical mound of earth, covered by a revetment of brick,
surrounded by a plinth also of brick which serves as a circular
path. Four chapels perhaps of later date are placed round the
dome at the four points of the compass and joined to it — each
containing the image of one of the four " cardinal " Buddhas.3
These chaityas still preserve the form of the earliest Buddhist
monuments. The plinth is the only feature of an architectural

1 Buchanan Hamilton, ' Account of
the Kingdom of Nepal,' pp. 35 and 211.

2 Oldfield's ' Sketches from Nepal/
vol. ii. p. 246 ; D. Wright's ' History
of Nepal,' p. 116; and Sylvain Levi,
' Nepal,' tome i. pp. 263 and 331.

3 See ante, p. 230 note. The Buddhas

are : Amitabha on the west ; Amogha-
siddha on the north ; Akshobhya on the
east ; and Ratnasambhava on the south.
The Araka chaityas and all the largest
temples have also a shrine for Vairochana
to the right of Akshobhya's : his proper
place is in the centre of the chaitya.

278

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

kind, but still rudimentary, which modified the rough outline
of the primitive stupa ; but the lofty spire of brick and masonry,
with its thirteen discs representing chhatras or umbrellas, is a
development of much later date, which has even been changed
into a solid cone or pyramid.1

After these, the two most important Buddhist monuments
in the valley of Nepal are those of Swayambhiinath and of

155. Temple of Swayambhunath, Nepal. (From a Drawing in the Hodgson Collection.)

Bodhnath;2 the former beautifully situated on a gentle eminence
about half a mile from Kathmandii, the latter at Bodhnath
about three and a half miles east from it; it is greatly reverenced
by the Tibetans under the name of the Ma-gu-ta chorten.

1 Sylvain Levi, ' Nepal,' tome ii.

pp. I, 2.

• A view of this chaitya forms the

'History of Nepal,' plate ix. p. 100;
Oldfield's < Sketches from Nepal,' vol. ii.
p. 260 ; and in Sylvain Levi's ' Nepal,'

frontispiece of Buchanan Hamilton's j tome i. p. 151, from a photograph,
volume; it also figures in Wright's I

CHAP. II.

CHAITYAS.

279

No very precise information is to be had about the date of
either, but, in their present form at least, they are not the oldest
in the valley. According to Brian H. Hodgson, there are
several low, flat, tumuli - like chaityas, with very moderate
chhattravalis or finials, which are older, and may be of any age ;
but, as will be seen from the previous woodcut (No. 155),
that at Swayambhunath is of an irregular clumsy form, and
chiefly remarkable for the exaggerated form of its tee or finial.1
This is, in fact, the most marked characteristic of the modern
Tibetan chaitya, which in China is carried frequently to such an
extent that the stiipa becomes evanescent, and the chhattrdvali
or spire changes into a nine or thirteen storeyed tower.
This chaitya stands on a narrow plinth projecting about 2 ft.
from the face of the dome; and the five shrines of the Dnyani
Buddhas, built partly into this plinth, were constructed by Raja
Pratapa Malla in the i/th century.2

The great Bodhnath chaitya is ascribed to King Manadeva
of the 6th century, as also to a Tibetan Lama, named Khasa, of
later date.3 It is raised on three successive platforms or terraces,
together about 45 ft. high, on which stands the great dome, 90 ft.
in diameter and rising another 45 ft.,and over this is the pyramidal
brick spire, reconstructed in 1825-1826. and of about the same
height. These chaityas are so subject to periodical repairs and
" restorations " that it is hard to say how much of them is original.

In Mr. Hodgson's collection there are nearly one hundred
drawings of chaityas in Nepal, all different, most of them small,
and generally highly ornamented ; but none of them grand,
and none exhibiting that elegance of form or beauty of detail
which characterises the buildings of the plains. From a low,
flat mound, one-tenth of its diameter in height, they rise to such
a tall building as this, which is a common form, bearing the
name of Kosthakar (Woodcut No. 1 56), in which the chaitya is
only the crowning ornament, and between these there is every
conceivable variety of shape and detail. Among others, there
is a four-faced lingam of Siva, with a corresponding emblem with
four Buddhas ; and altogether such a confusion of the two
religions as is scarcely conceivable.

By far the most characteristic and beautiful temples of the

1 For a detailed account of this chaitya
and its surrounding shrines, see Oldfield's
' Sketches from Nepal,' vol. ii. pp. 219-
246.

2 In an upper room of a small temple
on the west of it is preserved a perpetual
flame as a symbol of Adi-Buddha, which
is believed to have been derived from
heaven. It is tended by a family of

Tibetan lamas, in two cauldrons half-
filled with ghl, on which the lighted wicks
float. If by any mischance it should be
extinguished, it must be renewed from
another similar flame preserved at the
temple of Khasa Bodhnath.

3 The legend of its erection is given in
Waddell's ' Lamaism,' pp. 315-317; conf.
S. L£vi, ' Le Nepal,' tome i. p. 151.

280

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

Nepalese are those possessing many storeys divided with sloping

roofs. They are un-
like anything found
in Bengal, and all
their affinities seem
with those in Burma
or China. Usually,
they seem to be
dedicated either to
Vaishnava or to Saiva
worship, but in the
temple of Maha-
buddha at Patan,
Sakyamuni occupies
the basal floor, Ami-
tdbha the second
storey, a small stone
chaitya the third, a
Dharmadhatu Man-
dala or relic shrine
the fourth, and a
Vajra-dhatu Mandala
the fifth or apex of
the building, which
externally consists of
a small chudamani, or
jewel-headed chaitya.
This temple is
perhaps the most elaborately carved in the valley. It is about
75 feet high, and of unusual shape in Nepal. It was built about
the close of the i6th century, by a Buddhist Newar, named
Abhaya-raja, who in the reign of Amara Malla had gone on
pilgrimage to Bodh-Gaya, and brought back plans or a model
of the Mahabodhi temple, and, with his family, began to
construct this model of it.1

One of the most elegant of the sloping roofed class is the
Bhawani temple at Bhatgaon, represented in the woodcut
(No. 157). It was built in 1703 by Bhupatindra Malla to
enshrine a secret Tantric goddess, which to this day is not
allowed to be seen. It is five storeys in height, but stands
particularly well on a pyramid of five steps, which gives it a
greater dignity than many of its congeners.2 Another at Patan,

156. Nepalese Kosthakar. No scale.

1 Wright ' History of Nepal,' p. 204 ;
Oldfield, 'Sketches from Nepal,' vol. i. ,
plate at p. 272, and vol. ii. p. 269.

2 The stair up these five stages is
guarded by pairs of colossal figures ;

below are two athletes, above them
two elephants, then two lions, two tigers,
and at the top the goddesses or demons
— Singhini and VySghrini. The temple
itself is mostly of wood.

CHAP. II.

NEPAL.

281

Devi Bhawani Temple, Bhatgaon. (From a Photograph.

282

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

dedicated to Mahadeva, is seen in the centre of the wood-
cut (No. 158). It is only two storeys in height, but has the

Temples of Mahadeva and Krishna, Patan. (From a Photograph.)

CHAP. II.

NEPAL.

283

same characteristic form of roof, which is nearly universal in all
buildings, civil or ecclesiastical, which have any pretension
to architectural design. The temple on the left of the last
cut is dedicated to Krishna, and will be easily recognised by
any one familiar with the architecture of the plains from its
.yikhara or spire, with the curvilinear outline, and its clustering
pavilions, not arranged quite like the ordinary types, but still
so as to be unmistakably Bengali.

About 3 miles east from Kathmandu, on the right bank of
the Bagmati stream, is the sacred village of Pa^upati — the
Benares of the Nepalese worshippers of Siva. The place consists
almost entirely of temples and chapels of stone and wood, and
is sacred to Pa^upati or Siva as the god of beasts. A general
view of the village is given in the woodcut (No. 159). On
the right is prominent the double roof of the great temple of
Pa^upatinath — the most venerated Linga shrine of the Saivas
in Nepal. Its doors are overlaid with silver carved in the style
of those in the palace at Bhatgaon and at Patan. The truula
of the god may be seen to the right of the temple as well as on
its summit ; but the great Nandi or bull that rests in front of
the shrine is hid by the surrounding buildings. Close by it is
the place where widows are burnt as Satzs along with the bodies
of their dead husbands ; and the little chapels along the side of
the river are commemorative of the more notable. None of the
temples here are of any antiquity, most of them — if not all —
dating since the beginning of the I7th century.1

One other example must complete our illustration of
the architecture of Nepal. It is a doorway leading to the
darbar at Bhatgaon, and is a singularly characteristic specimen
of the style, but partaking much more of China than of India
in the style of its ornaments (Woodcut No. 160, p. 285). It is
indeed so like an archway in the Nankau Pass, near Pekin —
given further on — that I was at first inclined to ascribe them
to the same age. The Chinese example, however, is dated in
J345 2 > this one, according to Mr Hodgson, was erected as late
as 1725, yet their ornamentation is the same. In the centre is
Garuda, with a seven-headed snake-hood ; and on either hand
are Nagas, with seven-headed hoods also ; and the general
character of the foliaged ornaments is so similar that it is
difficult to believe in so great a lapse of time between them ;
but I cannot question Mr Hodgson's evidence. Since he was
in Nepal the building on the left-hand side of the cut has

1 Sylvain Levi's ' Le Nepal,' tome i.
PP- 357-366. The illustration (No. 159) is
from G. Le Bon's ' Monuments,' p. 245.

- ' Journal of the Royal Asiatic
Society ' (N.S.), vol. v. p. 18.

284 ARCHITECTURE IN THE HIMALAYAS. BOOK IT.

• "^ r ;

,'

m

m

CHAP. II.

NEPAL.

285

286 ARCHITECTURE IN THE HIMALAYAS. BOOK II.

been " improved." His drawings show it to have been one
of the most picturesque buildings in the valley. It certainly
is not so now.

In speaking of the architecture of Kanara (vol. ii., pp. 76, 77,
83), the similarity that existed between that of that remote
province and the style that is found in this Himalayan valley
will be remarked ; and scarcely any one can look at the illus-
trations referred to, especially Woodcuts Nos. 303 and 306, and
not perceive the similarity between them and the Nepalese
examples, though it might require a familiarity with all the
photographs to make it evident, without its being pointed out.

WOODEN TEMPLES.

In the Himalayan districts between Kashmir and Nepal, in
Kullu, Chamba, Kangra, and Kumaon, there are a vast number
of temples, regarding which it would be extremely interesting to
have more information than we now possess. They are all in
wood, generally Deodar pine, and, like most buildings in that
material, more fantastic in shape, but at the same time more
picturesque and more richly carved than buildings in more
permanent and more intractable materials. What we now
know of them, however, is mainly derived from photographs,
taken without any system, only as pictures, because the build-
ings were either picturesque in themselves or so situated as to
improve the landscape. No one yet has thought of measuring
them, nor of enquiring into their age or traditions ; and till this
is done it is impossible to treat of them in anything like a
satisfactory manner.

General Sir A. Cunningham in his Report for 1878-1879 made
some mention of the temples he visited at Barmawar, Chaitran,
and Chamba ; but beyond stating to what divinities they were
severally dedicated, and the inscriptions found, his remarks on
the architecture are of the briefest.1 In 1902 and in 1903, the
provincial archaeological surveyor visited the same places, but
in search of epigraphical materials, and the chief addition of an
architectural nature was a number of photographs, which, with-
out plans and descriptions made on the spot, do not help us.

In 1883 the late Mr. Wm. Simpson read to the Royal
Institute of British Architects a paper on ' Architecture in the
Himalayas,' in which he brought to notice that most of the
houses in the hill country between the Satlaj and the Ganges
valley are built of wood and stone ; timber being used in

1 'Archaeological Survey Reports,' vol. xiv. pp. 110-114.

CHAP. II. WOODEN TEMPLES. 287

alternate layers to bind the courses of stone together.1 Upon
this base of wood and stone stands the real dwelling which is
altogether of wood, and on beams it overhangs the more solid
structure beneath, and provides the dwelling with a verandah.
The gable line of the pointed roof is not straight, but has an
angle in it, making it steeper above, where protection was most
indispensable, whilst the slope diminished over the external
parts, that is in the eaves where leakage was least to be dreaded.
This gives a Chinese look to these Himalayan houses.

That this style of structure is not devoid of a certain
picturesqueness may be judged from a sketch of a Hindu temple
at Chergaon (Woodcut No. 161), on the Satlaj above Sarahan.

161. Hindu Temple at Chergaon in Chamba.

(From a Sketch by the late Mr. W. Simpson.)

It illustrates, too, the form to which the .rikhara is reduced in
such a building ; and it may be noted that the corners of the
projecting roofs are ornamented with quaint forms of gargoyles,
sometimes representing the form of a bird stuck on the snout of
a saurian or serpent.

Whenever this chapter of Indian architectural history comes
to be written, it will form a curious pendant to that of the
wooden architecture of Sweden and Norway, the similarities
between the two groups being both striking and instructive.
It cannot be expected that any ethnographical or political
connection can be traced between peoples so remote from one
another which could influence their architectural forms ; but

1 'Transactions of the Royal Institute of British Architects,' Session 1882-1883,
pp. 65ffg.

288

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

it is curious to observe how people come independently
to adopt the same forms and similar modes of decoration
when using the same materials for like purposes, and under
similar climatic influences. Although it may consequently
be impossible to trace any influence that the people of the
Himalayas could have exerted on the peoples of the north-
west of Europe, it is by no means clear that in these wooden
structures we may not find the germ of much that is now
perplexing us with regard to the earlier forms of Hindu stone
architecture. Like Buddhist architecture, there can hardly be a
doubt that much of it was derived from wooden originals, and it
is difficult to see any locality where wooden styles were likely
to be earlier adopted and longer practised than in those valleys
where the Deodar pine is abundant, and forms so excellent and
so lasting a building material.

An exploration of
these valleys, would,
no doubt, bring to
light many curious
monuments, which
would not only be
interesting in them-
selves, but might
throw considerable
light on many now
obscure points of our
enquiries. One
monument, for in-
stance, was dis-
covered by Major
Godwin Austen near
the foot of the Naga
hills in Asam, which
is unlike any other
known to exist any-
where else.1 The
temple — if temple it
may be called — con-
sists of a long cor-
ridor, about 250 ft. in
length and 2 1 ft. wide,
the roof of which was
supported by pillars richly carved, spaced 1 5 ft. to 2 1 ft. apart ;

1 The following particulars are taken from a paper by Major Austen in the
'Journal of the Asiatic Society of Bengal,' vol. xliii. pt. i., 1874, pp. 1-6.

162. Monoliths at Dimapur.

(From a Drawing by Major Godwin Austen.)

CHAP. II. WOODEN TEMPLES. 289

but its most remarkable features are two rows — one of sixteen,
the other of seventeen monoliths — standing in front of this. The
tallest is 15 ft., the smallest 8 ft. 5 in., the general range being
from 12 to 13 ft. in height, and 18 to 19 ft. in circumference.
No two are exactly alike, though all have a general similarity
of design to those represented in the preceding woodcut
(No. 162), which may be considered as typical of the style.
Another similar monolith was found a small distance off,
measuring 16 ft. 8 in. in height, and 23 ft. in circumference.

The natives were quite unable to give any account of these
curious monuments, nor is it easy to guess why they were placed
where they are. So far as I know, no similar monument exists
anywhere, for the pillars seem perfectly useless, though attached
to two rows of stones that may have borne a roof; otherwise
they look like those rows of rude stone monuments which we
are familiar with in this country and in Brittany, but which a
more artistic people may have adorned with rude carvings,
instead of leaving them quite plain, as our forefathers did. As
for their carving, the only things the least like them, so far as
I know, in India, are the pillars in the temple at Mudabidri
(Woodcut No. 305), and in other places in Kanara, but there
the pillars are actual supports of roofs ; these are round-headed,
and evidently never were intended for any utilitarian purpose.

Judging from the gateway and other remains of the town of
Dimapur, in which these pillars are found, they cannot be of
any great age. The gateway is of the Gaur type, with a
pointed arch, probably of the i6th or i/th century; and, if
Major Austen's observation is correct, that the sandstone of
which they are composed is of a friable and perishable nature,
they cannot be of any remote antiquity.

It would be very interesting if a few more similar monuments
could be found, and Asam is one of the most promising fields
in India for such discoveries. When Hiuen Tsiang visited it,
in the 7th century, it was known as the kingdom of Kamrup,
one of the three principal states of Northern India, and con-
tinued populous and important till the Pathan sovereigns of
Delhi attempted its conquest in the I5th century. Owing to
the physical difficulties of the country, they never were able
to succeed in this attempt ; but they blockaded the country
for many years, and, cut off from the rest of the world,
the savage hill tribes on either hand, aided by famine, so
depopulated the country that the jungle overpowered the feeble
remnant that survived, and one of the richest valleys in the
world became one of the most sparsely inhabited. When the
jungle has again been cleared, and rendered fit for human
population, there can be little doubt but that the remains of
VOL. I. T

290

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

many ancient cities will be found. Captain Dalton has
given an account of the ruins of Gauhati, which was almost
certainly the ancient capital of the province. " Its former
importance," he says, " is well attested by the immense extent
of its fortifications, and the profusion of carved stones which
every excavation of the modern town brings to light. The
remains of stone gateways and old stone bridges are found
both within and without the old city walls." l Captain Hannay
gives a view of one of these bridges. Like all the rest, it is
constructed without arches, on the horizontal principle,2 but it
may be as old as the time of the Chinese Pilgrims. Besides
these, other ruins have been found and described, in more or
less detail, in the pages of the 'Journal of the Asiatic Society
of Bengal.' When more fully known they will certainly be of
considerable historic and ethnographic value, though they hardly
can compare with the vast monuments of such provinces as
Orissa or Gujarat, and other parts of India Proper.

TIBET.

It would be extremely interesting if, before leaving this part
of the world, it were possible to compile anything like a satis-
factory account of the Buddhist style in Tibet, for it is there
that Buddhism exists at the present moment, in inexplicable com-
bination with Saivism and demon worship as the only religion,
and there only is it entirely and essentially a part of the system
of the people. We would gladly, therefore, compare the exist-
ing state of things in Tibet with our accounts of India in the
days of the supremacy of the same religion. The jealousy of the
Chinese, however, who are supreme over that nation of priests,
long prevented free access to the country, and it was only by the
expedition of 1903-1904, that Lhasa was reached and its mysteries
made known to the public with abundance of photographic
illustrations.3 But the reported architectural results are un-
important and present little that is novel. Relic worship, as an
essential element in Buddhism, is evidenced by the " chortens "
or stupas * everywhere met with, especially near the monasteries,
and the splendid tombs of the Grand Lamas at Tashi-lhunpo,

1 'Journal of the Asiatic Society of
Bengal,' vol. xxiv. pp. i et seqq.

2 Ibid. vol. xx. pp. 291 et seqq.

3 Capt. Turner, it is true, who was
sent to Tashi - Ihunpo by Warren
Hastings, published with his interesting
narrative a number of very faithful views
of what he saw, but they are not selected
from that class of monuments which is

the subject of our present enquiry.

4 Chorten, in Tibetan wChhoof-rten —
"relic receptacle," is equivalent to
dagaba, chaitya or stupa ; they are of the
usual characters of shrines enclosing
relics, cenotaphs, and symbols of the
religion, and receive formal worship
from the priests and people.

CHAP. II. TIBET. 291

Gandan, and in the Potala palace are special objects of worship.
But as no country in the world possesses a larger body of
priests in proportion to its population, and as these are vowed
to celibacy and live together, their monasteries are more
extensive than any we know of elsewhere — some containing
2000 or 3000 lamas, and, if we may trust the information
supplied to Mr. Rockhill, Debung Lamasery contains 9000,
Se-ra 7000, and Gandan 4OOO.1

The Tibetan monasteries are not built with any regularity, nor
grouped into combinations of any architectural pretension, but
consist of long streets of cells, mostly surrounding small court-
yards, three or four on each side. They are generally placed on
sites chosen with taste — either on the tops of hills with a wide
view round them, or in fertile valleys sheltered from the colder
winds. They occupy large areas in order to accommodate the
numerous population, and have the appearance of towns, consist-
ing, as they do, of a large aggregation of separate dwellings for
the monks, and surrounded by a high wall having four gates
towards the cardinal points. Outside are the houses and shops
of the tradesmen and shopkeepers. The houses are in the
usual style of the country — the walls often with more or less
batter — having the kitchen and storerooms on the ground floor
and the living rooms on the upper storey, which has a flat roof
forming a terrace. In the centre of the monastery is a large
square for assemblies, in the middle of which stand the temple,
library, meeting-hall of the authorities and mansion of the
superior or abbot, distinguished by a painted band or frieze of
reddish brown running round it under the eaves. The temples
are rectangular stone buildings, commonly constructed on a
general model, the walls often rough-cast in white with a broad
band of red or yellow colour under the eaves. The roofs are
formed of beaten clay or with tiles, on the middle of which is
raised a sort of pavilion with a Chinese roof decorated with little
gilt pyramids at the angles and apex. These temples have no
windows, the only daylight being admitted by the doors. Inside
they are divided longitudinally by two ranges of pillars into a
nave and side aisles, as in the chaitya caves in India. The
pillars and joists are painted yellow or bright red, and painted
silks are hung from the roof. At the inner end of the nave is the
altar or shrine with its three large images under the chhattra, and
lighted by lamps. The aisles are supplied with cushions for the
inferior lamas or monks, and their walls are covered either with
frescoed pictures or large paintings on silk of Buddhas, Jataka

1 'Journal of the Royal Astatic Society,' 1891, p. 278. Debung and Sera are both
in the vicinity of Lhasa, and rfGa^/dan — pronounced Gandan — is about 35 miles east
from Lhasa.

292

ARCHITECTURE IN THE HIMALAYAS. BOOK II.

scenes, divinities, etc. Attached to the temple and scattered
among the residences are numerous little chapels to the inferior
gods, goddesses, and demons.

The introduction of Buddhism into Tibet is ascribed to King
Sron-btsang-gam-po (629-650 A.D.), who married a Chinese and
a Nepalese princess — both of them Buddhists. He is said to
have built the monastery of Labrang in the centre of Lhasa
(A.D. 644), with perhaps the largest temple, as it is the most
ancient in the country. His Chinese queen had brought with
her a famous sandal-wood image of Sakyamuni and another of
Ananda, and for these was erected, about 650, the Ramoche
temple, a little to the north-east of Labrang.

The monastery of Potala1 (Plate VI.) outside the city of
Lhasa, where the Dalai Lama resides, seems to be of more
magnificence than all the rest — the centre being occupied by a
great block, dominating the others, which contains the temples,
audience halls and chaityas of the Dalai Lamas. It is known
as the Red Palace, and on its roof are the gilded pavilions of
Chinese style that render it so conspicuous in the landscape.
It was built by the first Dalai Lama, between 1642 and 1650,
on the ruins of the ancient fortress of Sron-btsang-gam-po, on a
hill in the west of Lhasa rising about 300 feet above the plain.
It is a great edifice of heavy though imposing aspect with its
gilded roofs and bells surmounting the chortens or chaityas that
enshrine the relics of the Dalai Lamas since the middle of the
1 7th century. Inside it is richly decorated, and, besides the
reception and state rooms and sanctuaries, it is said to contain
about 10,000 chambers for its myriad occupants. Around this
central palace are grouped a number of smaller ones, where
the inferior members of this great ecclesiastical order reside ;
but of all this it is difficult to form a distinct idea without
some better drawings than are at present available.

The Dalai Lama, who resides in this palace, is believed by
the Tibetans to be the living incarnation of the Bodhisattwa
Avalokitejvara, and, in consequence, is the principal object of
worship in Lhasa. There are, however, four or five subordinate
incarnations in different parts of Tibet and Mongolia, who,

1 Hue, from a mistaken etymology, has
' ' Buddha-la. " The later Buddhists speak
of three Potalas, as former residences of
Avalokiteivara — one at Tatta in the Indus
delta, but Hiuen Tsiang places it in the
extreme south of India, if not in Ceylon
(Beal, 'Buddhist Records,' vol. ii, p. 233),
and it is probably the same as Sumanakuta
or Adam's Peak ('Journal of the Royal
Asiatic Society,' N.S., vol. xv.p.339); the

second is P'u-t'o-shan among the Chusan
islands ; and the third, that at Lhasa, the
capital of the Dalai Lamas since 1643.
But there was another Chinese Potala, or
an imitation one, at Je-ho or Cheng-tu Fu,
about no miles north-east from Pekin.
— 'Journal of the Royal Geographical
Society,' vol. xxiii. (1904), p. 614; and
vol. xxix. (1907), pp. 1 80 and 185.

CHAP. II.

TIBET.

293

though inferior to this one, are still objects of worship in the
places where they reside, and by particular sects of Buddhists.1
It is this worship of a living, rather than of a dead deity, that
marks the difference of the forms of Buddhism in India and
Tibet. In the countries we have hitherto been describing no
actual incarnation of a Buddha is believed to have taken place
since the death of Sakyamuni though there have been many
saints and holy men ; in India, therefore, they have been
content to worship images of the departed, or relics which
recall his presence. In Tibet, where their divinity is still
present among them, continually transmigrating, but never
dying, of course such a form of worship is absurd ; no relic of
a still living god can logically exist, though this has probably
never been thought of, and the chaityas of the Great Lamas
are honoured, and worshipped in the palace or monastery
occupied by their successors.

The earliest monastery founded in Tibet is that of Sam-yas,
about 35 miles south-east from Lhasa, near the Sang-po river.
It was established by a famous teacher, Padma Sambhava, who
went from Bihar with other Buddhist teachers, about the middle
of the 8th century. He is said to have modelled it after the
great temple monastery of Otantapuri, near Nalanda, and it
became the metropolis of the Red-cap order.2 The monastery,
with its large temple and four separate colleges, is enclosed by
a circular wall about a mile and a half in circuit, and contains
a notable library and the State treasury. Another Indian
Pandit, named Atisha, came from the Vikrama^ila monastery
about 1038 and restored the Lamaism of his time, establishing
what afterwards became the Yellow -cap or Gelugpa order
of Lamas, which became the State church when its chiefs, the
Dalai Lamas, usurped the temporal power.

The monastery of Sakya, about 50 miles west-south-west
from Shigatse, was founded in 1071. Its Grand Lama was
acknowledged by Khubilai Khan in 1270 as head of the
church, and made tributary prince of Tibet. This position his
successors maintained for a century, and the sect played an
important role in the history of Tibet till the Gelugpas super-
seded it early in the I5th century. The establishment is said
to contain the largest single building in Tibet: it is seven

1 The heads of the Pan-chhen Rin-po-
ches of Tashi-lhunpo are regarded as
perpetual incarnations of Amitabha. The
Tar&natha Dalai Ldmas have their seat
at Urga in Mongolia, whither the late
Lama Pope fled.

2 Otantapuri and Vikramarila' monas-
teries were most probably among those

destroyed by Muhammad Bakhtiyar
Khalji about 1194. — 'Jour. Asiat. Soc.
Bengal,' vol. Ivi. pt. i. p. 19, and vol. iii.
(1907), p. 221 ; Griinwedel, ' Mythologie
des Buddhismus in Tibet,' etc., p. 55 ;
L. de Milloue, ' Bod-Youl ou Tibet,' pp.
28iff. ; Elliot, 'History of India,' vol. ii.
PP- 3°5f-

294 ARCHITECTURE IN THE HIMALAYAS. BOOK II.

storeys high, and has a spacious assembly hall. Its library is
famous for its collection of Sanskrit and Tibetan manuscripts —
probably all Buddhist.1

About the beginning of the i$th century Tsong-Khapa, a
Lama of Kumbum, re-organised the sect founded by Atisha
and re-named it the Gelugpa. His first monastery was that
of Gandan, founded in 1409, about 30 miles east from Lhasa, of
which he was abbot till his death in 1417. The chief object
of veneration here is the lofty mausoleum of the founder, built
of marble and ornamented with malachite and with a gilded
roof. It encloses a chorten or stupa said to be all of gold, in
which is deposited the embalmed remains of the sage. One of
Tsong-khapa's disciples founded Sera monastery about 2. miles
north of Lhasa in 1417, in which are about $500 monks.
Depung, 3 or 4 miles west from Lhasa, is also a monastery of
the Gelugpas, and contains fully 7000 inmates, mostly devoted
to exorcism and magic. It was founded in 1414, and is said to
be named or modelled after the early Indian monastery of
Dhanyakataka or Amaravati.2 Within the enclosure is a large
temple surrounded by four chapels, and a palace of the Lhasa
Lama. Out of scores of other such establishments, we may
mention Tashi-lhunpo, visited by Boyle and Turner in the end
of the 1 8th century. It is in western Tibet, near Shigatse,
about 140 miles west of Lhasa, was founded by Tsong-khapa
in 1445, and contains about 3500 monks. It is the seat of the
Pan-chhen Grand Lama, who is next to the Lhasa pontiff* in
dignity and influence. Here is the tomb erected by the Chinese
emperor Kiu-long for the Lama Erdeni who died on a visit to
Pekin in 1779. It is figured by Turner as is also the Go-ku-pea
— some nine storeys high for the display of religious pictures.3

Lastly, at Gyan-tse4 on the route followed by our troops
in 1904, is a large fortified monastery, itself forming a little
town on the southern slope of a hill to the north of the fort.
Its buildings, standing on the edge of the plain, rise in tiers,
like a large amphitheatre round the great temple at their base.
This temple, shown in Plate VII., is of interest from its
form.5 It is locally known as Gandho-la — a name usually
applied to the great temple at Bodh-Gaya, of which local
tradition names this as being a model. It is about 100 ft. high,

1 Waddell's ' Lamaism,' p. 274.

2 As Amaravati and its monastic estab-
lishment had been deserted and dis-
appeared a thousand years before this, the
connection between the two is imaginary,
as in the cases of Sam-yas and Gyan-tse.

3 Turner, ' Account of an Embassy to
the Court of Teshoo Lama ' ( 1800), plates

ii and 12; 'Mission of Geo. Bogle to
Tibet,' etc., pp. 96ffg.

4 Gyan-tse lies about 106 miles west-
south-west from Lhasa, in latitude 28° 53'
N., and longitude 89° 34' E.

5 This form reminds us of the Jaina
samosaranas at Girnar and Satrunjaya.

PLATE VII.

GOLDEN TEMPLE AT GYAN-TSE. (From a Photograph by Lieut. F. M. Bailey.)

[To face page 294, Vol. I.

CHAP. II.

SIKHIM.

295

with a circumference of 600 ft. at the base, and is built in five
stepped terraces with recessed angles, on the plan of the
vimanas of Indian temples. Above these is a circular drum of
one storey, and over it a smaller square one surmounted by a
spire of thirteen great rings of gilt copper crowned by a
chhattra canopy of the same material. In the different storeys
are numerous shrines to the different Buddhas, which are
reached by inside stairs, and the terrace roofs of the successive
storeys form a series of chaityanganas 1 for the circumambulation
of the different groups of cells.

From Sikhim, which is overrun by Lamas, and has borrowed
its architecture from Tibet, we may gain further acquaintance
with the characteristic features of the style. The view (Wood-
cut No. 163) of the doorway of the temple at Tashiding is

163. Doorway in the Temple at Tashiding. (From Dr. Hooker's ' Himalayan
Journals,' vol. i. p. 319.)

curious as showing a perseverance in the employment of sloping
jambs, which we do not meet with in the plains of India, but
is usual in Tibet.2 It will be recollected that this feature is
nearly universal in the Bihar and early western caves (Woodcuts
Nos. 55, 58, and 64), but there we lose it. It may have con-

1 This term is used among Buddhists for
the pradakshinapath or terrace. Waddell's
' Lhasa and its Mysteries,' pp. 217, 229-232.

2 E. Schlagintweit's ' Buddhism in
Tibet,' pp. 188%, and Milloue, ' Bod-
Youl ou Tibet,' pp. 279ffg.

296 ARCHITECTURE IN THE HIMALAYAS. BOOK II.

tinued to be employed during the Middle Ages, though the

examples have perished ; but it is curious to find it cropping
up here again after a lapse of 2000 years.1

1 It is found currently employed in
the decorative sculpture of the Gandhara
monasteries, but rarely as a constructive

feature. See Foucher, ' L'Art Greco-
Bouddhique du Gandhara,' tome i. pp.
107-108,

CHAP. II. TEMPLES AT KANGRA. 297

Another view in the porch of the temple at Pemiongchi
is also interesting, as showing the form of roof which we are
familiar with in the rock examples, and also as illustrating
the extent to which the bracket capital of India may be
carried under the influence of wooden architecture (Woodcut
No. 164). It hardly seems doubtful that the idea was origin-
ally derived from wooden construction, but was equally appro-
priate to masonic forms, and is used in masonry so judiciously
by Indian architects that we lose sight of its origin in most
instances altogether.

Interesting as these minor styles undoubtedly are from
their variety, and valuable though they may be for the hints
they afford us in understanding the history of the other styles,
they never can be so important as the greater architectural
groups that are found on the plains of India itself. A mono-
graph of the styles of Kashmir or Nepal, or of the intermediate
valleys, would be an invaluable addition to our knowledge ;
but hardly more is required in a general history than that
their places should be indicated, and their general character-
istics so defined as to render them recognisable. Even these
minor styles, however, will become more intelligible when
studied in connection with the Dravidian and northern styles,
which are those it is next proposed to define and describe.

TEMPLES AT KANGRA.

Though a little out of their place in the series, there are
two small temples in one of the Himalayan valleys which
it may be expedient to describe here before leaving this part
of the subject, as their peculiarities will assist us in under-
standing much that has just been said, or that will be presently
advanced. Besides this, they do not exactly fit into any other
series, but they can hardly be passed over, as they possess what
is so rare in Indian temples — an ascertained date.

The temples are situated in the village of Kiragrama or
Baijnath, about 25 miles east of Kangra, and two inscriptions
in the larger of the two principal shrines record that it was
built by two brothers, Manyuka and Ahuka, wealthy merchants,
under Lakshmanachandra lord of Kiragrama, and apparently
brother-in-law of Jayachchandra, king of Jalandhara or Trigarta.
The date is partly obliterated, but in all probability it was
equivalent to A.D. 1204. This temple was dedicated to Siva-
Vaidyanatha, and consists of a mandap or hall, 20 ft. square
inside, with four round pillars supporting the roof, and a shrine
(puri] for the Linga, 8 ft. square — separated from the hall by

298 ARCHITECTURE IN THE HIMALAYAS. BOOK II.

a small antechamber.1 Over all this temple measures 51 ft. by
31, and stands in a walled enclosure about 120 ft. long and
60 ft. wide at the east end, and 75 at the west. This contains
some rooms for the priests with small temples along the north
side. But in 1786 it underwent a thorough repair at the hands
of Raja Sansarachandra II., which has obliterated many of
its features ; and to this repair it probably owes the porch with
its four pillars in front ; but with the exception of the
balcony windows on each side, the walls were not, perhaps,
materially meddled with. The roof and spire, however, were
either rebuilt or so overlaid with plaster as to hide the original
work. The woodcut (No. 165) shows only the pillars of the
portico of the temple, with the Nandi kiosk in front and
a small temple of Jamadagni beyond. This latter, though
ruinous, is more interesting, because it has escaped the hand
of the spoiler. As will be seen from the woodcut, it has all
the features of a very old temple — great simplicity of out-
line, no repetitions of itself, and the whole surface of the
upper part covered with that peculiar horse-shoe diaper which
was so fashionable in those early days. It looks here as if
it must be copied from some brick or terra-cotta construction ;
otherwise its repetition over a whole surface seems unaccount-
able. The amalaka stringcourses are subdued and in good
taste, and the crowning ornament well proportioned.

There is little doubt that the jikhara of the larger temple
was similarly adorned, but all its details are so completely
obliterated by the coating of plaster it has received that it
has lost its interest. The pillars, however, of its porch retain
their forms up to their capitals, at least. The architraves, as
may be seen from the woodcut, belong to the repair in 1786.
The shafts of the pillars are plain cylinders, of very classical
proportions, and the bases also show that they are only slightly
removed from classical design. The square plinth, the two
toruses, the cavetto or hollow moulding between, are all classical,
but partially hidden by Hindu ornamentation, of great elegance,
but unlike anything found afterwards. The capitals are, how-
ever, the most interesting parts, though their details are con-
siderably obliterated by whitewash. They belong to what may
be styled the Hindu-Corinthian order, though the principles on
which they are designed is diametrically opposed to those of
the classical order of the same name. The object of both — as
is well known — is to convert a circular shaft into a square

1 The inscriptions are in the mandap, j lated in ' Epigraphia Indica,' vol. i. pp.

high up in the side walls, right and left | 97-118; vol. ii. p. 482; vol. v. App.

from the entrance, — a most unusual posi- ! p. 78, and ' Indian Antiquary,' vol. xx.

tion for such records. They are trans- ! p. 154.

CHAP. II.

TEMPLES AT KANGRA.

299

165. Temples at Klragrama, Kangra District. (From a Photograph

300 ARCHITECTURE IN THE HIMALAYAS. BOOK II.

architrave-bearing capital in a graceful and pleasing manner.
We all know the manner in which the Ionic and Corinthian
capitals effect this ; pleasingly, it is true, but not without effort
and some little clumsiness, which it required all the skill and
taste of classical architects to conquer. To effect this object,
the Hindus placed a vase on the top of their column, the bowl
of which was about the same diameter as that of the pillar on
which it was placed, or rather larger ; but such an arrangement
was weak, because the neck and base of the vase were necessarily
smaller than the shaft of the pillar, and both were still circular.
To remedy these defects, they designed a very beautiful class
of foliaged ornament, which appears to grow out of the vase,
on each of its four faces, and, falling downwards, strengthens
the hollows of the neck and foot of the vase, so as to give them
all the strength they require, and at the same time to convert
the circular form of the shaft into the required square for the
abacus of the capital. The Hindus, of course, never had
sufficient ability or constructive skill to enable them to produce
so perfect a form as the Corinthian or Ionic capitals of the
Greeks or Romans ; but it is probable that if this form were
taken up at the present day, a capital as beautiful as either of
these might even now be produced. It is, indeed, almost the
only suggestion that Indian architecture seems to offer for
European use.

It is by no means clear when this form of capital was first
introduced. It first appears, but timidly it must be confessed,
in such Buddhist caves as were excavated after the end of the
2nd century : — as, for instance, in the Sri Yajna cave at Nasik
(Woodcut No. 105) ; in the courtyard of the VLrwakarma, at
Elura (Woodcut No. 83) ; and in some of the later caves at
Ajanta — the twenty-fourth for instance. It is found at Eran
(Woodcut No. 1 66), among some fragments that I believe to be
of the age of the Guptas, about A.D. 400, and it is currently
employed in the middle group of Hindu caves at Elura, such
as the Ravana-ka Khai, and other caves of that age, say about
A.D. 600. It afterwards became frequent, almost universal, with
the Jains, down to the time of the Muhammadan conquest.
The following representation of one (Woodcut No. 167), from
a half column of a temple in Orissa, shows it in a skeleton form,
and therefore more suited to explain its construction than a
fuller capital would do. On its introduction, the bell-shaped or
Persepolitan capital seems to have gone out of fashion, and does
not again appear in Indian art.

To return from this digression : there can be no doubt that
the temple of Vaidyanatha is dedicated to Siva, not only from
the presence of the bull in front of it, in a pavilion of the same

CHAP. II.

TEMPLES AT KANGRA.

301

architecture as the porch, but also because Ganeja appears
among its integral sculptures. In the niche in the back

166. Pillar in porch of a
Temple at Eran, of the
Gupta age.

167. Capital of Half Column from a Temple in
Orissa. (From a Lithograph.)

however, is the base of what has been a marble image of
Mahavira, with an inscription in two lines telling that it was
consecrated in A.D. 1240. This base must have been transferred
to the Siva, temple after the destruction of the original Jaina
shrine, and probably owes its preservation and that of a figure
placed over it, to the ignorance of the priests of Vaidyanatha.1

The temple second in interest is that of Siddhanath at the
west end of the town. It consists of a four-pillared hall and a
shrine, measuring 33 ft. by 20 ft. over all, and with a jikhara
about 35 ft. in height. It faced to the east, and had doorways
on each side of the shrine leading into a pradakshina or
circumambulatory passage.2 In a niche in the south wall there
had been an inscription, long since illegible ; but in the back or
west wall was a figure of Surya ; and the temple was most
probably dedicated to the sun.

1(Epigraphia Indica,' vol. i. pp. 98
and 118-119, and compare p. 120. On
it was placed a sandstone figure of Surya
— perhaps from the Siddhanath temple.
A similar transference has occurred at
Kot Kangra, where an image of Parr-

wanath was found in the porch of a small
Hindu temple — Ibid. p. 1 20.

2 The drawings of this temple in
Cunningham's 'Archaeological Reports,'
vol. v. plate 44 are not correct.

BOOK III.
DRAVIDIAN STYLE.

CHAPTER I.

INTRODUCTORY.

THE limits within which the Dravidian style of architecture
prevailed in India are not difficult to define or understand.
Practically they are those of the Madras Presidency, or, to
speak more correctly, they are identical with the spread of
the people speaking Tamil, or the cognate tongues. Dr.
Caldwell, in his 'Grammar,' estimated these, in 1874, at forty-
five or forty -six millions,1 but he includes among them a
number of tribes, such as the Tudas and Gonds, who, it is
true, speak dialects closely allied to the Tamil tongues, but
unless we know their history, language is only a poor test
of race, and in this instance architecture does not come to
our aid. And, so far as we at present know, these tribes
are in too rude a state to have any architecture of their own
in a sufficiently advanced state for our purposes. Putting
them aside, therefore, for the present, we still have, according
to the census of 1901, over fifty-two millions of people speaking
Tamil, Telugu, Kanarese, and Malayalam,2 whom we have no
reason for doubting are practically of the same race, and who,
in so far as they are Hindus — not Jains, but followers of 5iva
and Vishnu — practise one style of architecture, and that known
as the Dravidian. On the east coast the boundaries of the

1 ' Comparative Grammar of the
Dravidian Languages,' London, 2nd
edition, 1875, p. 42.

2 This total includes the Madras
Presidency, Mysore, Travankor, part of
the Bombay Presidency and Haidarabad.
In the last named there are 5,148,000
Telugu people and 1,562,00x2 Kanarese.

The totals in each language are : —
Tamil . . . 16,299,000
Malayalam . . 5,278,000
Telugu . . . 20,409,000
Kanarese . . 10,234,000

52,220,000

CHAP. I.

INTRODUCTORY.

3°3

style extend as far north as the mouth of the Krishna, and
it penetrates sporadically and irregularly into the Nizam's
territories, but chiefly through the Telugu speaking districts.

On the west coast its natural boundary northwards is the
Krishna to the Dharwar district, and thence south-east, past
Vijayanagar and to the east of Sravana-Belgola and north of
Mysore city westwards to the coast. Much of the Kanarese
country lies to the north and west of this, and a large
part of the Telugu area is to the north of it ; but, of course,
examples of the style are to be found beyond this line, and
of other styles within it. At Elura in latitude 20° N. we have
it, but this is most probably due to the Rashtrakuta kings
having employed architects from their capital of Malkhed,
200 miles to the south. It took no permanent root there,
however, while the reflex wave brought the northern styles into
Mysore or other southern countries, where their presence was
as little to be expected as that of the Dravidian so far north.

Although considerable progress has lately been made in
the right direction, no satisfactory solution has yet been
arrived at of the problem of the origin of the Dravidians. So
far as history is concerned, in such glimmerings of tradition
as we possess, at the earliest time at which we find any
mention of them the most civilised and important of their
communities occupied the extreme southern point of the
peninsula.1 North of them all was forest, but between the
Christian Era and the Muhammadan invasion we find the
jungle gradually disappearing, and the southern races pushing
northwards, till, in the I4th century, they were checked and
driven back by the Moslim.

Till we know more about the origin of these Dravidian
races, however, it seems expedient for the present to assume
that the Tamil-speaking races are practically aboriginal. As
far back as their traditions reach, we find the Dravida Deja,
or southern part of India, divided into kingdoms or states,
of which three are frequently mentioned — the Pandyas, the
Cholas, and the Cheras,2 forming a little triarchy of powers,
not often interfered with by the other nations of the peninsula,
nor interfering with those beyond their limits. During the
greater part of their existence their relations of war and

1 See H. H. Wilson's historical sketch
in the 'Journal of the Royal Asiatic
Society,' vol. iii. (1836), p. 199 et seqq.

2 In Aroka's second edict reference is
made to Chola, Pandya, Satiyaputra,
Keralaputra and Ceylon ; and, as Kerala
is well known as the name of the Chera

country, we have here mention of all
three states together with ' Satiyaputra,'
which may possibly be that of the
JTatakarnis of Banavasi. The Ajoka
inscriptions discovered in 1892 in Mysore
are indicative of the wide influence of
that emperor.

3°4

DRAVIDIAN STYLE.

BOOK III.

peace have been chiefly among themselves, and they have
grown up a separate people, as unlike the rest of the world
as can well be conceived.

Of the three the most southern was the Pandya kingdom,
which occupied the extreme south of the peninsula from
Cape Comorin to the Vellar river1 in the Pudukottai state
on the east coast, and to Achchankovil Pass2 on the west,
including the southern part of Travankor. It seems to have
been of sufficient importance about the time of the Christian
Era to have attracted the special attention of Greek and
Roman geographers and merchants. How much earlier it
became a state, or had a regular succession of rulers, we
know not,3 but it seems to have attained to some consistency
as early as five or six centuries before the Christian Era,
for we find its princes referred to in the earliest Singhalese
traditions of the ' Mahawansa.' 4 Their early capital was
probably Korkai or Kolkai, at the mouth of the Tamraparni
close to Kayal, and known to the early geographers as the
seat of the pearl fishery ; but the Pandya princes, at an early
date, either removed to Madura, or possibly they had another
capital there. This continued to be the seat of government
of the later rulers of the country from some time in the I2th
century till its absorption in the middle of the i8th.

During the long period of their rule, the Pandyas had
several epochs of great brilliancy and power, followed by
long intervening periods of depression owing to frequent
invasions from their neighbours the Chola kings of
Trichinopoly. The 1st century, and afterwards the 5th
or 6th, seem to have been those in which they especially
distinguished themselves. The large number of gold, silver
and copper coins, chiefly of the Roman emperors from
Augustus to Nero, that have been found at different places
in the region, indicate a considerable commerce with the West
at that age. We have lists of kings, but how far mythical
we know not, the times at which any of them lived being
quite unknown before the pth or loth century, and such as
have been approximately ascertained between the pth and
1 3th century are those mentioned in Chola inscriptions. If
buildings of the first ten or twelve centuries exist, which is by

1 It falls into Palks Strait in 10° 8' N.
latitude.

2 In N. latitude 9° 6'.

3 See Bishop Caldwell's ' Political and
General History of the District of Tinne-
velly' (Madras, 1881); and 'Indian
Antiquary,' vol. xxii. (1893), pp. 57ffg.

4 Vijaya sends to the Pcandya king for
his daughter. — ' Mahawansa,' ch. vii.
And the second and fourth Sinhalese
princes are named Pandu - vasa and
Pandukabhaya, as being of Pandya
descent.

CHAP. I. INTRODUCTORY. 305

no means improbable, they are still unknown to us, and must
remain so till the full results of the Archaeological Surveys
are adequately published. In the early ages of their history the
Pandya kings are believed to have been Buddhists, but one of
them, perhaps in the I2th century, became a vSaiva, and
persecuted the heterodox sect mercilessly.

The Muhammadans, led by Malik Kafur, conquered Madura
in 1311, and for about half a century they held the country,
till dispossessed by a new line of Pandya princes, who were
probably more or less under the supremacy of Vijayanagar.
About 1525, however, the Nayyaks — officers of Vijayanagar —
usurped supreme power at Madura, and ruled there till 1736.
By far the most distinguished prince of the Nayyak dynasty
was Tirumalai Nayyak, who ruled from 1623 till 1659. This
prince adorned the capital city of Madura with many splendid
edifices, some of which have been drawn by Daniell and others.
What more ancient remains there are will not be known till
they have been carefully and scientifically surveyed, and the
results published.

The Chola kingdom extended northward from the border
of the Pandya country and the valley of the Kaveri and
Kolerun rivers, whose banks seem always to have been its
principal seat, to the Pal£r river or nearly to Madras, all
along the eastern coast, called after them Cholamandalam or
Coromandel. Westwards their kingdom extended into Mysore,
but the boundaries varied at different periods, and, after the
fall of the Pallavas of Kanchi, they advanced northwards to
the limits of the Eastern Chalukyas of Vengi. The date of
the origin of their kingdom is not known, but the mention of
Chola as a state in the Aj-oka edicts is proof of its antiquity.
Their early capital was Uraiyur, now a suburb of Trichinopoly.
The earliest princes of the dynasty, whose position we can
assign, belong to the 9th century, but it is only with Rajaraja I.,
who became king in 983, that any connected chronology
commences. Their epoch of greatest glory was between the
loth and I2th centuries, when they seem to .have conquered
not only their neighbours the Pallavas, Pandyas and Cheras,
but even to have surpassed the bounds of the triarchy, and
carried their arms into Ceylon, and to have maintained an equal
struggle with the Chalukyas in the north. Their capital during
this period was at Kanchi, now Conjivaram, which they had
wrested from the Pallavas. By the middle of the I3th century,
however, their power had waned, and they sank step by step,
first under the Muhammadans, from whom it passed to the
Nayyaks of Madura, and then to the Marathas.

The Cheras occupied the country northward of the Pandya
VOL. I. U

306

DRAVIDIAN STYLE.

BOOK III.

kingdom, and westward of the Chola, stretching along the west
or Malabar coast to Honawar: whether they ever occupied
any considerable portion of Mysore or of Koimbatur is not so
certain. The capital was Karur or Vanji, probably adjacent
to Cranganor in the Cochin state.1 Tradition assigns to the
state a series of kings styled Perumal, who seem to have been
elective, and are mentioned as Pandyas and Cholas, but the
list ended with Cheruman Perumal, in the pth century, who is
said to have become a Moslim. No Chera inscriptions have
as yet come to light to help us to unravel their history, and the
territory was probably divided into principalities, whose main
defence was the mountain range, separating their country from
their eastern neighbours.

But however this may be, the old Chera region is best
defined by its architecture, for the style of Hindu temple,
Jaina basti, and Moslim mosque, is distinctly one throughout
the country from the north of Kanara to the south of Travankor.
Jainism probably penetrated into the country at an early date,
and till late in the i6th century the Tuluva rajas were Jains,
whilst the colossal statues at Kcirkala and Venur in South
Kanara and at Sravan - Belgola in Mysore, with the bastis at
Mudabidri, Bhatkal and elsewhere testify to their zeal. On the
other hand, the Malayalis are snake worshippers to the present
day ; in their gardens the Hindus usually provide a " Naga-
kotta" or snake shrine.2

A fourth dynasty, already mentioned, appears at an early
date in the history of the Dekhan ; the Pallavas 3 possibly rose
to power on the decay of the Andhra power in the third century,
and they seem to have secured much of the Chola country,
probably before the 5th. About A.D. 400 Samudragupta, among
other princes, claimed to have overthrown Vishnugopa of Kanchi,
who was doubtless a Pallava king, and recent epigraphical research
has now brought to light details respecting this dynasty which
ruled the country called Dravida from their capital at Kanchi,
now Conjivaram. In the /th century they were at war with
the Chalukyas of Badami, and Narasimhavarman I. claimed to
have destroyed Badami during the reign of the Chalukya king
Pulike^in II. (A.D. 609-642), whilst his father Mahendravarman
had, at an earlier date, defeated the same king near Kanchi when

1 Ptolemy (' Geographia,' vii. 86) men-
tions Karoura as the capital of Kero-
bothras, probably intended for Keraput-
ran, ' King of Chera.' — 'Indian Anti-
quary,' vol. xxxi., pp. 343-344; Logan's
' Malabar,' vol. i. pp. 252, 253.

2 Ante, p. 43, note 3.

1 The 'Mahabharata," Vishnu Purana,'

and ' Varaha Mihira' (in the 6th century),
frequently mention the Pahlavas among
the tribes in the north-west, and some
scholars have tried to identify them with
the Pallavas, who were in the south-east
of India as early as the 2nd century, but
the similarity of names alone will hardly
justify the assumption of identity.

CHAP. I.

INTRODUCTORY.

he invaded Dravida.1 Probably about this period the Pallavas
extended their rule over the Bellari district and parts of Mysore ;
but in the following century we learn that Vikramaditya II., the
Chalukya king, about 740, defeated Nandivarman Pallava and
entered Kanchi, bestowing gifts on the shrine of Rajasirnhej-vara
— now the Kailasanatha temple — built by Narasimhavarman II.
A century later they were attacked by the Rashtrakutas, and their
power seems to have been broken, and they gradually succumbed
to the Cholas, who re-asserted their power in the loth century.

Parantaka or Viranarayana (cir. A.D. 907-946) advanced the
Chola power and boasts of taking Madura from the Pandya
king and the invasion of Ceylon,2 as also of gilding the "Golden
Hall " at Chidambaram, the then famed temple of his race. His
son Rajaditya was killed in battle by the Rashtrakuta king
Krishna III., and a period of civil war followed till 985, when
Rajaraja secured the throne and carried his conquests as far
as Kalinga on the north, to Kollam or Quilon on the west, and
to Ceylon in the south.8 Though a J>aiva, we have a long
copperplate grant of the 2ist year of his reign granting a
village to a Buddhist temple at Negapattam built by a king
of Kataha — apparently in the eastern peninsula * ; but his great
architectural monument was the Tanjor temple. For a century
after Rajaraja I. the Cholas maintained the commanding position
he had gained for his dynasty, but from the time of Vikrama
Chola (1118-1135), their power gradually waned, and after this
the rise of the Ballalas in Mysore, and the revival of the Pandyas
in the south, seem to have checked them to such an extent that
they never regained their previous position.

Although, politically, these states always remained distinct,
and generally antagonistic, the people belonged to the same
race. Their architecture is different from any other found in
India, but united in itself, and has gone through a process of
gradual change from the earliest times at which we become
acquainted with it, until we lose sight of it altogether in the
last century. This change is invariably for the worse, the earlier
specimens being in all instances the most perfect, and the degree
of degradation forming, as mentioned above, a tolerably exact
chronometric scale, by which we may measure the age of the
buildings.

Buddhism does not seem to have ever gained such a footing

1 Hultzsch, ' South Indian Inscrip-
tions,' vol. iii. pp. 34offg. ; and Fleet,
' Bombay Gazetteer,' vol. i. pt. ii. pp.

2 The ' Mahawansa ' seems to place this
invasion in the time of Udaya III.,

A.D. 964-972; but the flight of the
Pandu king from MadurS to Ceylon is
mentioned under Kasyapa IV., A.D.
929-939« 'Mahawansa,' chh. Hi., liii.

' Mahawansa,' ch. lv.
4 Ante, p. 206.

3o8 DRAVIDIAN STYLE. BOOK III.

among the Dravidian races generally as it did in northern and
western India ; yet, in the 7th century, when Hiuen Tsiang
visited Kanchipuram, most probably when Narasimhavarman I.
was the Pallava sovereign, he reckoned "some hundred of
sangharamas with ten thousand priests " in the Dravida country,
all belonging to the Sthavira school of the Mahayana, with
eighty Hindu temples and many Jaina heretics.1 In Malakuta
or the Pandya country he reports from hearsay that the
monasteries were mostly ruinous, whilst Jains were numerous ;
and for the Chola province, which he probably passed through,
he makes a like statement. Like their temples at Negapattam,
Buddhist monasteries would mostly be of brick, and when the
sect disappeared, whether from persecution — as tradition asserts
— or through absorption into Vaishnava or other sects, their
buildings would be pulled down or altered for other purposes.

The Jaina religion long continued to flourish at Conjivaram
and in Mysore; but, though influential from their intelligence,
the Jains never formed more than a small numerical fraction of
the people among whom they were located.

The Hindu religion, which was probably always supreme
in the Dravidian districts, now commonly designated the
Brahmanical, is divided into the worshippers of Siva and
Vishnu, which are quite distinct and almost antagonistic ;
but both are so overloaded with absurd fables and monstrous
superstitions that it is very difficult to ascertain what they
really are or ever were. Nor are we yet in a position to speak
confidently of their origin.

Both these religions have borrowed an immense amount of
nomenclature from the more abstract religions of the Aryan
races, and both profess to venerate the Vedas and other scriptures
in the Sanskrit language. Indeed it is all but impossible that
the intellectual superiority of that race should not make itself
felt on the inferior tribes, but it is most important always to
bear in mind that the Sanskrit-speaking Aryan was a stranger
in India. It cannot indeed be too often repeated that all that
is intellectually great in that country — all, indeed, which is
written — belongs to them ; but all that is built — all, indeed,
which is artistic — belongs to other races, who were either
aboriginal or immigrated into India at earlier or subsequent
periods, and from other sources than those which supplied the
Aryan stock.

There does not seem to be any essential difference either in
plan or form between the Saiva and Vaishnava temples in the

1 Beal, ' Buddhist Records,' vol. ii. p. 229.

CHAP. I.

INTRODUCTORY.

309

south of India. It is only by observing the images or emblems
worshipped, or by reading the stories represented in the
numerous sculptures with which a temple is adorned, that we
find out the god to whom it is dedicated. Whoever he may
be, the temples consist almost invariably of the four following
parts, arranged in various manners, as afterwards to be explained,
but differing in themselves only according to the age in which
they were executed : —

1. The principal part, the actual temple itself, is called the
Vimdna.1 It is always square in plan, and surmounted by a
pyramidal roof of one or more storeys ; and it contains the cell
in which the image of the god or his emblem is placed. .

2. The porches or Mantapams^ which always cover and
precede the door leading to the cell.

3. Gate-pyramids, Gopurams? which are the principal features
in the quadrangular enclosures that surround the more notable
temples.

4. Pillared halls or Chaultris — properly Ckdwadts* — used
for various purposes, and which are the invariable accompani-
ments of these temples.

Besides these, a temple always contains tanks or wells for
water — to be used either for sacred purposes or the convenience
of the priests — dwellings for all the various grades of the priest-
hood are attached to it, and numerous other buildings for state
or convenience.

1 Vimdna is generally used to designate
"a chariot" or vehicle of the gods, a
moving palace ; hence it includes the
shrine and spire.

a In Sanskrit — Mandapa, a pavilion or
open porch, thence a hall, and a temple.

3 Gopura means a town gate, hence an
entrance, applied to the lofty towers over

the entries to southern temples. The
later style of gopurams dates from the
1 6th century, and do not properly belong
to the original Dravidian temples. They
were probably intended for purposes of
defence against invasion and plunder.

4 ChdwattoT Chdwadi is a public lodging
place, a shelter for travellers.

3io DRAVIDIAN STYLE. BOOK III.

CHAPTER II.
HINDtJ CONSTRUCTION.

CONTENTS.
Arches — Domes — Plans — 5ikharas.

ARCHES.

BEFORE proceeding to describe the arrangements of Hindu or
Jaina temples, it may add to the clearness of what follows on
the various styles if we first explain the peculiar modes of
constructing arches and domes which they invariably employed.

As remarked above, although we cannot assert that the
Buddhists never employed a true arch, this at least is certain —
that, except in the roofs of one or two small chaityas recently
discovered, no structural example has been found in India, and
that all the arched or circular forms found in the caves are
without exception copies of wooden forms, and nowhere even
simulate stone construction. With the Hindus and Jains the
case is different : they use stone arches and stone domes which
are not copied from wooden forms at all ; but these are
invariably horizontal arches, never formed or intended to be
formed with radiating voussoirs.

It has been explained, in speaking of Pelasgic art,1 how
prevalent these forms were in ancient Greece and Asia Minor,
and how long they continued to be employed even after the
principles of the true arch were perfectly understood. In India,
however, the adherence to this form of construction is even more
remarkable. As the Hindus quaintly express it, " an arch never
sleeps " ; and it is true that a radiating arch does contain in
itself a vis viva which is always tending to thrust its haunches
outwards, and goes far to insure the ultimate destruction of
every building where it is employed : while the horizontal forms
employed by the Hindus are in stable equilibrium, and, unless
disturbed by violence, might remain so for ever.

1 Fergusson's 'Ancient and Mediaeval Architecture' (3rd ed.), vol. i. pp. 243
et seqq.

CHAP. II.

ARCHES.

There can be no doubt that the Hindus carried their horror
of an arch to an excess which frequently led them to worse
faults on the other side. In city walls for instance, where there
is a superabundant abutment on either hand to counteract any
thrust, the horizontal principle is entirely misplaced. If we take,
for instance, one of the city gates at Vijayanagar (Woodcut No.
1 68), we cannot help perceiving that with much smaller stones

168.

View of City Gateway, Vijayanagar. (From a Photograph.)

and less trouble a far more stable construction could have been
obtained, so long as the wall on either hand remained entire.
What the Hindu feared was that if the wall were shattered, as
we now find it, the arch would have fallen, though the horizontal
layers still remain in their places.

Instead of a continuous bracket like that shown in the last
example, a more usual form, in modern times at least, is that of
several detached brackets placed a little distance apart the one
from the other. When used in moderation this is the more
pleasing form of the two, and in southern India it is generally
used with great success. In the north they are liable to
exaggerate it, as in the gateway from Jhinjhuwada in Gujarat
(Woodcut No. 169, p. 312), when it becomes unpleasing, though
singularly characteristic of the style.1

1 Other examples of the same style may be seen in the gateways of Dabhoi. —
Burgess and Cousens, 'The Antiquities of Dabhoi in Gujarat,' plates 10, 13, and 16.

312

DRAVIDIAN STYLE.

BOOK III.

It is this horizontal or bracket mode of construction that is
the formative principle of the Dravidian or Southern style of

Hindu architecture,
every form and every
ornament depending
almost wholly upon it.
In the north, however,
another development of
the same principle is
found in the horizontal
dome, which is scarcely
known in the south, but
which has given a new
character to the style,
and, as one of its most
beautiful features, de-
mands a somewhat de-
tailed explanation.

DOMES.

It is to be regretted
that, while so much has
been written on the
history of the pointed
arch, so little should
have been said regarding
the history of domes :
the one being a mere
constructive peculiarity

that might very well have been dispensed with ; the other being
the noblest feature in the styles in which it prevails, and perhaps
the most important acquisition with which science has enriched
the art of architecture.

The so-called Treasuries of Mycenae and Orchomenos, as
well as the chambers in Etruscan tombs, prove that as early
as ten or twelve centuries before Christ the Pelasgic races had
learned the art of roofing circular chambers with stone vaults,
not constructed, as we construct them, with radiating vaults,
on the principle of the common arch, but by successive layers
of stones converging to a point, and closed by one large stone
at the apex.

Whoever invented the true or radiating arch, the Romans
were the first who applied it as a regular and essential archi-
tectural feature, and who at the same time introduced its com-
plement, the radiating dome, into architectural construction ;

169. Gateway, Jhinjhuwada.

(From Kinloch Forbes' Ras Mala.)

CHAP. II.

DOMES

at what period it is not now known. The earliest example, the
Pantheon, is also the finest and largest ; but we have lost
entirely the innumerable steps by which the architects must
have slowly progressed to so daring an experiment.

There is, however, a vast difference between these two classes
of domes, which it is necessary to bear in mind in order to
understand what follows.

The Roman arch and Roman dome are always constructed
(Woodcut No. 170) on the principle of voussoirs, or truncated
wedges, radiating from a centre. This enabled the Romans to
cover much larger spaces with their domes than perhaps was
possible on the horizontal principle ; but it involved the incon-
venience of great lateral thrusts, continually tending to split the
dome and tear the building in pieces, and requiring immense
and massive abutments to counteract their destructive energy.

170. Radiating Arch.

171. Horizontal Arch.

The Indian or horizontal dome never can be made circular
in section, except when used on the smallest scale, but almost
always takes a form more or less pointed (Woodcut No. 171).
From the time of the building of the Treasury of Mycenae l to
the birth of Christ, we have a tolerably complete series of arches
and vaults constructed on this principle, but few domes properly
so called. After the Christian Era the first example is found
in a singular tomb at Mylassa,2 near Halicarnassus in Caria,3
where the dome exhibits all the peculiarities of construction
found in the Jaina temples of India. After this we almost lose
the thread of its history till the form reappears in porches like
those of the I ith century on Mount Abu, where it is a perfectly
established architectural feature, that must have been practised
long before it could be used as we find it in that building.
Whether we shall ever be able to recover the lost links in this
chain is more than doubtful, but it would be deeply interesting
to the history of art if it could be done. In the meantime, there
is no difficulty in explaining the constructive steps by which the
object is now attained in India. These may also throw some

1 ' History of Ancient and Mediaeval
Architecture,' vol. i. p. 243.

* Ibid. p. 371 ; and ante, p. 209.

3 Fully illustrated in vol. ii. of the
:i~..i — t: Society's ' Antiquities of

Dilettanti
Ionia.'

DRAVIDIAN STYLE.

BOOK III.

light on the history of the invention, though this is not, of
course, capable of direct proof.

The simplest mode of roofing a small square space supported
by four pillars is merely to run an architrave or stone beam from
each pillar, and cover the intermediate opening by a plain stone
slab. Unless, however, slabs of great dimensions are available,
this mode of construction has a limit very soon arrived at. The
next step therefore is to reduce the extent of the central space
to be covered by cutting off its corners ; this is done by
triangular stones placed in each angle of the
square, as in Woodcut No. 172, thus employ-
ing five stones instead of one. By this means,
the size of the central stone remaining the
same, the side of the square space so roofed
is increased in the ratio of ten to seven, the
actual area being doubled. The next step
in the process (Woodcut No. 173) is by em-
ploying three tiers and nine stones, instead
of two tiers and five stones, which quadruples
17*- the area roofed. Thus, if the central stone

Diagram of Roofing. - , . . c ,

is 4 ft., by the second process the space roofed
will be about 5 ft. 8 in. ; by the third 8 ft. square ; by a fourth
process (Woodcut No. 174) — with four tiers and thirteen stones

V V

173-

Diagrams of Roofing

— the extent roofed may be 9 ft. or 10 ft, always assuming the
central stone to remain 4 ft. square. All these forms are still
currently used in India, but with four pillars the process is
seldom carried further than this ; with another tier, however,
and eight pillars (as shown in Woodcut No. 175), it may be
carried a step further — exactly the extent to which it is carried
in the tomb at Mylassa above referred to. In this, however, as

CHAP. II.

DOMES.

in all instances of octagonal domes in this style, instead of the
octagonal form being left as such, there are always four external
pillars at the angles, so that the
square shape is retained, with twelve
pillars, of which the eight internal
pillars may be taken as mere inser-
tions to support the long architrave
between the four angular pillars.

It is evident that here again we
come to a limit beyond which we
cannot progress without using large
and long stones. This was sometimes
met by cutting off the angles of the
octagon, and making the lower course
of sixteen sides. When this has been
done an awkwardness arises in getting
back to the square form. This was
escaped from, in all the instances I
am acquainted with, by adopting cir- w- Diagram of Roofing,
cular courses for all above that with

sixteen sides. In many instances the lower course with sixteen
sides is altogether omitted, and the circles placed immediately

Diagram of Indian Construction.
B. Form of bracket capital in the angle of an octagonal dome.

on the octagon, as in the temple of Vimala at Abu (Woodcut
No. 284, vol. ii. p. 39). It is difficult to say how far this system

3i6 DRAVIDIAN STYLE. BOOK III.

might be carried constructively without danger of weakness.
The Indian domes seldom exceed 30 ft. in diameter, but this
may have arisen more from the difficulty of getting architraves
above 12 ft. or 13 ft. in length to support the sides, than from
any inability to construct domes of larger diameter in them-
selves. This last difficulty was to some extent got over by a
system of bracketing, by which more than half the bearing of
the architrave was thrown on the capital of the column, as
shown in Woodcut No. 176. Of course this method might have
been carried to any extent, so that a very short architrave
would suffice for a large dome ; but whether this could be done
with elegance is another matter. The Indians seem to have
thought not ; at least, as far so I know, they never carried it to
any extent. Instead of bracketing, however, they sometimes
used struts, as shown in Woodcut No. 176, but it is questionable
whether that could ever be made a really serviceable con-
structive expedient in stone architecture.

The great advantage to be derived from the mode ol
constructing domes just described was the power it gave
of placing them on pillars without having anything to fear
from the lateral thrust of the vault. The Romans never
even attempted this, but always, so to speak, brought their
vaults down to the ground, or at least could only erect them
on great cylinders, which confined the space on every side.
The Byzantine architects cut away a great deal of the sub-
structure, but nevertheless could never get rid of the great
heavy piers they were forced to employ to support their domes ;
and in all ages were forced to use either heavy abutments
externally, or to crowd their interiors with masses of masonry,
so as in a great measure to sacrifice either the external effect
or the internal convenience of their buildings to the constructive
exigencies of their domes. This in India never was the case ;
all the pressure was vertical, and to ensure stability it only
required sufficient strength in the support to bear the down-
ward pressure of the mass — an advantage the importance of
which is not easily over-estimated.

One of the consequences of this mode of construction was,
that all the decoration of the Indian domes was horizontal,
or, in other words, the ornaments were ranged in concentric
rings, one above the other, instead of being disposed in vertical
ribs, as in Roman or Gothic vaults.1 This arrangement allows
of far more variety without any offence to good taste, and
practically has rendered some of the Indian domes the most

1 The tendency of the Indian con-
struction, however, was to make the
section of the dome nearly conical as

each course or ring of stone, after the
first two or three, had about the same
amount of projection inwards.

CHAP. II.

DOMES.

exquisite specimens of elaborate roofing that can anywhere be
seen. Another consequence of this mode of construction was
the employment of pendants from the centres of the domes,
which are used to an extent that would have surprised even
the Tudor architects of our own country. With them, however,
the pendant was an architectural tour de force, requiring great
constructive ingenuity and large masses to counterbalance it,
and is always tending to destroy the building it ornaments ;
while the Indian pendant, on the contrary, only adds its own
weight to that of the dome, and has no other prejudicial
tendency. Its forms, too, generally have a lightness and
elegance never even imagined in Gothic art ; it hangs from
the centre of a dome more like a lustre of crystal drops than
a solid mass of marble or of stone.

As before remarked, the eight pillars that support the dome
are almost never left by themselves, the base being made square
by the addition of four others at the angles. There are many
small buildings so constructed with only twelve pillars, as shown
in the annexed diagram (No. I77)>
but two more are oftener added on

each face, making

twenty altogether,

as shown on the

upper side of the

annexed diagram

(No 178) ; or four

on each face, mak-
177. ing twenty-eight ;

Diagram of the arrange- or again, two in
JSa Dome" * front of these four, Diagram Plau'of Jaina Porch.

or six on each face,

so as to make thirty-six ; and the same system of aggregation
is carried on till the number of pillars reaches fifty-six (Wood-
cut No. 179), which is the largest number I ever saw surrounding
one dome ; but any number of these domes may surround one
temple, or central dome, and the number consequently be multi- •
plied almost ad infinitum. When so great a number of pillars
is introduced as in the last instance, it is usual to make the
outmost compartment on each face square, and surmount it with
a smaller dome. This is occasionally though rarely done even
with the smallest number.

The first result of this arrangement is, that the Hindus
obtained singularly varied outline in plan, producing the
happiest effects of light and shade with every change in the
sun's position. Another result was, that by the accentuation of
the salient and re-entering angles, they produced those strongly-

DRAVIDIAN STYLE.

BOOK III.

marked vertical lines which give such an appearance of height
to Gothic designs. To accomplish this, however, the Western

architects were obliged to
employ buttresses, pinnacles,
and other constructive ex-
pedients. The Hindus obtained
it by a new disposition of the
plan without anywhere inter-
rupting the composition. This
form of outline also expresses
the internal arrangements of
the porch better than could be
done by the simpler outline of
either a square or circle, such
as is usually employed in
Europe. Its greatest merit,
however, is, that the length of

179. Diagram of jaina Porch. the greater aisles is exactly

proportioned to their relative

width as compared with that of the subordinate aisles. The en-
trance being in the angle, the great aisle forms the diagonal,
and is consequently in the ratio of 10 to 7, as compared
to what it would be if the entrance were in the centre of
the side, where we usually place it. From the introduction
of the octagonal dome in the centre the same proportion
(correctly 1000 to 707) prevails between the central and side
aisles, and this again is perhaps the most pleasing that has yet
been introduced anywhere. In Gothic churches the principal
aisles are generally twice as wide as the side ones, but they are
also twice as high, which restores the proportion. Here, where
the height of all is the same, or, nearly so, this gradation just
suffices to give variety, and to mark the relative importance
of the parts, without the one overpowering the other: and
neither has the appearance of being too broad nor too narrow.
It is, of course, difficult for those who have never seen a
building of the class just described to judge of the effect of these
arrangements ; and they have seldom been practised in Europe.
There is, however, one building in which they have accidentally
been employed to a considerable extent, and which owes its
whole beauty to the manner in which it follows the arrangement
above described. That building is Sir Christopher Wren's
church of St. Stephens, Walbrook. Internally its principal
feature is a dome supported on eight pillars, with four more in
the angles, and two principal aisles crossing the building at right
angles, with smaller square compartments on each side. This
church is the great architect's masterpiece, but it would have

CHAP. II. PLANS. 319

been greatly improved had its resemblance to a Hindti porch
been more complete. The necessity of confining the dome and
aisles within four walls greatly injures the effect as compared
with the Indian examples. Even the Indian plan of roofing,
explained above, might be used in such a building with much
less expense and less constructive danger than a Gothic vault
of the same extent.

PLANS.

Till the discovery of the small Buddhist chaitya halls at Ter
and Chezarla and elsewhere, already described (p. 1 26), there was
only one temple in India which gave us any hint of how the
plans of such halls were related to those of Hindu and Jaina
temples. Fortunately, however, its evidence is so distinct that
there could be very little doubt about the matter. The temple in
question is situated in the village of Aihole, in Bijapur district,
in western India, not far from the place where the original
capital of the Chalukyan sovereigns was situated, and near the
caves of Badami on the one hand and the temples of Pattadakal
on the other. Its date is uncertain to some extent, since an
inscription on its outer gateway recording a grant to the temple,
during the reign of Vikramaditya-Satya-sraya, is undated;1 and
there were two Chalukya kings of this name — one ruling between
A.D. 655 and 680, and the other between 733 and 746. But the
grant was to a temple already established, and even if made in
the 8th century the fane might well be of fifty or eighty years
earlier date, as its architecture would indicate. It is thus not
only one of the oldest structural temples known to exist in
western India, but in fact one of the only three yet discovered
that can with any certainty be said to have been erected before
the beginning of the 8th century.

This temple, as the sculptures testify, was dedicated to
Vishnu — the special divinity of the Chalukyas ; but the words
carved in Kanarese on the basement — " the holy Jaina temple " —
seems to indicate that at one time it had been claimed or
appropriated by the Jains, and this, with some misconception as
to the character of the sculptures, has led to the mistake of
its being supposed that it was originally Jaina. Its original
dedication is fortunately, however, of very little importance for
our present purposes. The age when this temple was erected
was the age of toleration in India. The Chinese traveller
Hiuen Tsiang has left us a most vivid description of a great

1 ' Indian Antiquary,' vol. viii. pp. 285-286; and 'Archaeological Survey of
Western India,' vol. i. pp. 40 et seqq.

320

DRAVIDIAN STYLE.

BOOK III.

quinquennial festival, at which he was present at Allahabad in
A.D. 643, at which the great King Siladitya presided, and dis-
tributed alms and honours, on alternate days, to Buddhists,
Brahmans, and heretics of all classes, who were assembled there
in tens of thousands, and seem to have felt no jealousy of each
other, or rivalry that led, at least, to any disturbance.1 It was
on the eve of a disruption that led to the most
violent contests, but up to that time we have no
reason to believe that they did not all use
similar edifices for their religious purposes, with
only such slight modifications as their different
formulae may have required (Woodcut No. 180).
Be this as it may, any one who will compare
the plan of the chaitya at Sancht (Woodcut
No. 47), which is certainly Buddhist, with that of
this temple at Aihole, which is Vaishnava, can
hardly fail to perceive how nearly identical they
must have been when complete. In both in-
stances, it will be observed, the apse is solid,
and it appears that this always was the case in
structural free-standing chaityas. At least, in
all the rock-cut examples, so far as is known,
the pillars round the apse are different from
those that separate the nave from the aisles ;
they never have capitals or bases, and are mere plain makeshifts.
From the nature of their situation in the rock, light could not
be admitted to the aisle behind the apse from the outside, but
must be borrowed from the front, and a solid apse was con-
sequently inadmissible ; but in free-standing examples, as at
Aihole, it was easy to introduce windows there or anywhere.
Another change was necessary when, from an apse sheltering a
relic-shrine, it became a cell containing an image of a god ; a
door was then indispensable, and also a thickening of the wall
when it was necessary it should bear a tower or .dkhara to mark
the position of the cella on the outside. Omitting the verandah,
the other changes introduced between the erection of these two
examples are only such as were required to adapt the points of
support in the temple to carry a heavy stone roof, instead of the
light wooden superstructure of the primitive Buddhist chaitya
(Woodcut No. 181).

It may be a question, and one not easy to settle in the
present state of our knowledge, whether the Buddhist chaityas
had or had not verandahs, like the Aihole example. The rock-

180.

OldTempleatAihole.
(From a Plan by J.
Burgess.) Scale 50
ft. to i in.

1 Hiouen Thsang, ' Memoires sur les Contrees Orientales,' tome i. pp. 253 et seqq, ;
or Beal, ' Buddhist Records,' vol. i. pp. 214 et seqq.

CHAP. II.

PLANS.

321

cut examples naturally give us no information on this subject,
but the presumption certainly is, looking at their extreme
appropriateness in that climate, that they had this appendage,
sometimes at least, though not perhaps usually.

If from this temple at Aihole we pass to the neighbouring

Saiva one of Papanatha at Pattadakal, built probably not very
much later, we find that we have passed the boundary line that
separates the ancient from the mediaeval architecture of India,
in so far at least as plans are concerned (Woodcut No. 182).
The circular forms of the Buddhists have entirely disappeared,
and the cell has become the base of a square tower, as it
VOL. I. X

322

DRAVIDIAN STYLE.

BOOK III.

remained ever afterwards. The nave of the chaitya has become
a well defined mandapa or porch in front of, but
distinct from, the cell, and these two features in
an infinite variety of forms, and with various
subordinate adjuncts, are the essential elements
of the plans of the Jaina and Hindu temples of
all the subsequent ages.

The procession-path round the cell — called
pradakshina — as that round the apse, remained
for some centuries as a common but not a uni-
versal feature. The verandah disappeared.
Round a windowless cell it was useless, and the
pillared porches contained in themselves all the
elements of shelter or of shadow that were
required.

SlKHARAS.

182. Plan of Papa-
natha Temple at
Pattadakal. Scale
Soft, to i in.

There is one other peculiarity common to
both Hindu and Jaina architecture in the north of India that
requires notice, before proceeding to describe particular ex-
amples. It is the form of the towers or spires called ^Sikharas
or Vimanas, which invariably surmount the cells in which the
images are placed. It is probably correct to assert that the
images of the Tirthankaras are invariably placed in oblong
or square cells, and those of Hindu deities in square — gener-
ally cubical cells, of no great dimension, and that these cells
receive their light from the doorway only. It seems also an
invariable rule that the presence and position of the cell
should be indicated externally by a tower or spire, and that
these towers, though square or nearly so in plan, should
have a curvilinear outline in elevation. If the tower at Bodh-
Gaya (ante, p. 78) retains unaltered the original form given
to it when erected about the 5th or 6th century, this dictum
would not apply to Buddhist architecture. As it is, however,
the only Buddhist .yikhara yet discovered it is hardly fair
to draw any decided inference from one single example, while
with Jaina or Hindu towers I know of no exception. Take,
for instance, the tower represented in the following woodcut
(No. 183), which purports to be an elevation of the celebrated
Black Pagoda at Kanarak in Orissa, and may be looked
upon as a typical example of the style, and of which it may
be considered of a fair medium example. The upper part
of the tower, to some extent, overhangs its base. It bends
inward towards the summit, and is surmounted by what is
called an Amalaka — a massive circular coping stone which
supports a vase called amritakalasa or amritakaraka, i.e. " dew

CHAP. II.

^IKHARAS.

323

vessel." l Its peculiar corrugated form occurs frequently in old
examples as a sort of blocking course dividing the jikharas

horizontally into numerous
small compartments, and it
seems as if what is used there
in a straight-lined form was
employed as a circular orna-
ment at the summit. It is
a very beautiful architectural
device, and was, as far as
I can see, adopted only
because it was so, and con-
trasted brilliantly with the
flat ornaments with which
it was employed. At present

183. Restored Elevation of the Sun-temple at Kanarak.

(From a Drawing by the Author.) No Scale.

we do not seem to be in a position to explain its origin, or
that of a great many other details that are frequently met with
in Hindu architecture.

1 The "amalaka" has been popularly-
supposed to be derived from dmalaka — the
Phyllanthus emblica, Emblica officinalis
or Emblica myrobalan ; but, though an
article of Hindu materia medica, it is so
insignificant a berry that it could hardly
be thought of as an architectural model.

May it not be from aniala — "pure,"
" spotless " ? Amalanla — " pure stone," is
applied to this crowning member. — Beal,
' Buddhist Records,' vol. ii pp. 136-137 ;
Foucher, ' L'Art Greco-Bouddhique du
Gandhara,' tome i. p. g6n.

324

DRAVIDIAN STYLE.

BOOK III.

Whatever its origin, this amalaka is generally surmounted
by a flat dome of reverse curvature, in the centre of which stands
the kalasa, karka, or pinnacle, in the form of a vase, generally of
very beautiful and graceful design.

The great and at first sight puzzling question is, from what

original is this curious combina-
tion of forms derived ? It is
like nothing found anywhere out
of India, and like no utilitarian
form in India, that we now
know of. It cannot be derived
from the dome -like forms of
the topes. They are circular

184.

Diagram Plan and Section of the Temple at Kanarak, designed to
explain the construction of Hindti Temples.

CHAP. II.

SIKHARAS.

325

both in plan and elevation. The 5ikharas are straight-lined in
plan, and their section is never a segment of a circle ; it is not
derived from any many-storeyed buildings, as the .rikharas or
vimanas of the Dravidian architecture of the south of India,
which seem certainly to have been copied from the many
storeyed viharas of the Buddhists, and we cannot fancy any
class of domestic building which could have formed a model
out of which they could have been elaborated. One curious
thing we do know, which is that all the ancient roofs in India,
whether represented in the bas-reliefs or copied in the caves,
were invariably curvilinear — generally circular or rather ogee —
having a ridge added externally to throw off the rain from
that weakest part ; but nothing on any bas-relief or painting
gives us a hint of any building like these .rikharas.

Another curious and perplexing circumstance regarding
the .rikharas is that when we first meet them, at Bhuvane^war, for
instance, on the Bay of Bengal, or at Pattadakal in the 7th
century, near the west coast of India, the style is complete
and settled in all its parts. There was no hesitation then,
nor has there been any since. During the twelve or thirteen
centuries that have elapsed since the erection of these earliest
known examples, they have gone on becoming more and more
attenuated, till they are almost as pointed as Gothic spires,
and their degree of attenuation is no bad test of their age ;
but they never changed in any essential feature of the design.
All the parts found in the oldest examples are retained in
the most recent, and are easily recognisable in the buildings
of the present time.

The one hypothesis that occurs to me as sufficient to
account for this peculiarity is to assume that it was a con-
structive necessity. If we take for instance an assumed
section of the diagram (Woodcut No. 184, p. 324), it will be
seen how easily a very tall pointed horizontal arch, like that
of the Treasury at Mycenae referred to above, p. 312, would
fit its external form. In that case we might assume that
the tower at Bodh-Gaya took a straight-lined form like the
doorway at Missolonghi and the ' Gate of Lions ' at Mycenae,
while the Hindus took the more graceful curvilinear shape, which
certainly was more common in remote classical antiquity,1 and
as it is found in Persia may have reached India at a remote
period.

This hypothesis does not account for the change from

1 See Woodcuts Nos. 102, 114, 124,
126, 129, 172, 177 and 178 of vol. i. of
the author's ' History of Ancient and
Mediaeval Architecture,' 3rd edn. ; and

for the Missolonghi doorway and Mycenae
Gate of Lyons, — Ibid., Nos. 130 and 131
on p. 247 .

3*6

DRAVIDIAN STYLE.

BOOK III.

the square to the circular form in the upper part, nor for
its peculiar ornamentation ; but that may be owing to our
having none of the earlier examples. When we first meet
with the form, either in Dharwar or Orissa, it is complete in
all its parts, and had evidently reached that state of per-
fection through long stages of tentative experience. The dis-
covery of some earlier examples than we now know may one
day tell us by what steps that degree of perfection was reached,
but in the meanwhile I fear we must rest content with the
theory just explained, which, on the whole, may be considered
sufficient for present purposes at least.1

1 In his work on the ' Antiquities of
Orissa,' vol. i., Babu Rajendralal Mitra
suggests at page 31 something of this

sort, but if his diagram were all that is
to be said in favour of the hypothesis, I
would feel inclined to reject it.

CHAP. III. MAMALLAPURAM. 327

CHAPTER III.

DRAVIDIAN ROCK-CUT TEMPLES.

CONTENTS.
Mamallapuram — Kailas, Elura.

ALTHOUGH it may not be possible to point out the origin of
the Dravidian style, and trace its early history with the same
precision as we can that of Buddhist architecture, there is nothing
so mysterious about it as there is regarding the styles of northern
India, nor does it burst on us full blown at once as is the case
witn the architecture of the Chalukyas. Hitherto, the great
difficulty in the case has been, that the temples of southern India
have almost all been found to be of so modern a date. The great
building age there was the i6th and I7th centuries of our era.
Some structural buildings, it is true, could be traced back to the
nth or 1 2th with certainty, but beyond that all was to a great
extent conjecture ; and if it were not for rock-cut examples, we
could hardly go back much further with anything like certainty.
Recent investigations, however, combined with improved know-
ledge and greater familiarity with the subject, have now altered
this state of affairs to a great extent. It seems hardly doubtful
now that the Kailas at Elura, and the great temples at Pattadakal,
are anterior to the loth century.1 In fact, it has been ascertained
that they date from the 8th, and the " raths," as they are called,
at Mamallapuram or the " Seven Pagodas " on the Madras coast,
are as early as the /th century, and are in reality the oldest
examples of their class known, and the prototypes of the style.

One circumstance which prevented the age of the Mamalla-
puram raths being before detected is, that being all cut in granite
and in single blocks, they show no sign of wearing or decay,
which is so frequently a test of age in structural buildings, and
being all in the same material produces a family likeness among
them, which makes it at first sight difficult to discriminate
between what is old and what new. More than this, they all

Burgess, ' Report on Belgam and Kaladgi,' 1875, plates 39, 40.

328

DRAVIDIAN STYLE.

BOOK III.

possess the curious peculiarity of being unfinished, whether
standing free, as the raths, or cut in the rock, as caves, or on
its face, as the great bas-relief; they are all left with one-third
or one-fourth merely blocked out, and in some instances with
the intention merely indicated. It looks as if the workmen had
been suddenly called off while the whole was in progress, and
native traditions, which always are framed to account for what
is otherwise most unintelligible, have seized on this peculiarity,
and make it the prominent feature in their myths. Add to this
that it is only of late we have acquired that knowledge of the
subject and familiarity with its details, which enable us to check
the vagaries of Indian speculation. From all these causes it is
not difficult to understand how easily mistakes may be made in
treating of such mysterious objects.

If we do not know all we would wish about the antiquities
of Mamallapuram,1 it is not because attempts have not been
made to supply the information. Situated on an open beach,
within 32 miles of Madras, it has been more visited and oftener
described than any other place in India. The first volume of
the 'Asiatic Researches' (1788) contained an exhaustive paper
on them by Wm. Chambers. This was followed in the fifth
(1798) by another by Mr Goldingham. In the second volume
of the ' Transactions of the Royal Asiatic Society ' (1830) there
appeared what was then considered a most successful attempt
to decipher the inscriptions there, by Dr. Guy Babington,
accompanied by views of most of the sculptures. Before this,
however, in 1816, Colonel Colin Mackenzie had employed his
staff to make detailed drawings of all the sculptures and
architectural details, and he left a collection of about forty
drawings, which are now, in manuscript, in the India Office.
Like all such collections, without descriptive text, they are
nearly useless for scientific purposes. The 'Madras Journal,' in
1844, contained a guide to the place by Lieutenant J. Braddock,
with notes by the Rev. G. W. Mahon, the Rev. W. Taylor,
and Sir Walter Elliot;2 and almost every journal of every
traveller in these parts contains some hint regarding them, or
some attempt to describe and explain their peculiarities or
beauties. With the exception of the Mackenzie MS. — the

1 The name of this place, among
English writers, has been subject to
various changes ; a century ago it
usually figured as Maha Balipuram, as
in Southey's ' Curse of Kehama ' ; Dr.
Babington stated that in the inscrip-
tions it was called Mahimallaipur ; the
Rev. W. Taylor made it Mamallapuram,
which is now accepted ; other forms

were Mavalivaram, Mahavallipur, etc.

2 ' A Guide to the Sculptures, Excava-
tions and other remarkable subjects
at Mamallaipur, generally known to
Europeans as "the Seven Pagodas"
by the late Lieut. John Braddock, etc.'
in ' Madras Journal of Literature and
Science,' vol. xiii. (June, 1844) pp. 1-56.

CHAP. III.

MAMALLAPURAM.

329

most of these were collected in a volume in 1869 by a
Lieutenant Carr, and published at the expense of the Madras
Government,1 but, unfortunately, as too often happens, the
editor selected had no general knowledge of the subject, nor
had he apparently much local familiarity with the place. His
work in consequence added nothing to our previous stores.2

In 1883 the editor called attention to the temple of
Kailasanathaswamin at Conjivaram as a Pallava temple of
probably about the 7th century, containing a number of early
inscriptions.3 On examination by Dr. Hultzsch, these were
found partly to belong to the same period as those of the
raths at Mamallapuram. The temple, now seemingly a
Vaishnava shrine, was erected by the Pallava king Rajasimha

185.

Raths, Mamallapuram. (From a Sketch by the Author.

alias Narasimhavarman II., towards the end of the /th century
and dedicated to I^wara or Siva ; and the names on the
Dharmaraja rath agree with those on this temple, confirming
the date arrived at in 1880 for these works.4

1 It included also a short account of
the place written in Kanarese for Col.
Mackenzie in 1803, with a translation.
The publication was issued in two
forms — in atlas folio, 96 pp. , and also in
octavo, 246 pp. , with folding plates, and
with the same errata.

- A survey of the monuments at
Mamallapuram was made a considerable
time ago, by the Archseological Survey
of Madras.

3 Hultzsch, ' South Indian Inscrip-
tions,' vol. i. pp. 8ff.

4 'Cave Temples of India,' p. no.

DRAVIDIAN STYLE.

BOOK III.

The oldest and most interesting group of these monuments
are the so-called five " Raths " l or monolithic temples standing
on the seashore to the south of the other rock excavations.
One of these, having an apsidal termination, appears in the
centre of the preceding woodcut (No. 185) a little detached
from the rest. The other four stand in a line running from
north-north-east to south-south-west, and look as if they had
been carved out of a single stone or rock, which originally, if
that were so, must have been between 35 ft. and 40 ft. high
at its southern end, sinking to half that height at its northern
extremity, and its width diminishing in like proportion.

The first on the north is the Draupadi Rath — a mere pansala

or cell II ft.
square externally,
and with a curvi-
linear roof rising
to about 18 ft.
high (Woodcut
No. 1 86). Appar-
ently it was once
crowned by a
finial of some
sort, but its form
cannot now be
ascertained. This
rath is the most
completely fin-
ished of the five,
and is now unique
of its kind, but
must have be-
longed to an ex-
tensive class of
buildings when it
was executed, and
their form conse-
quently becomes
important in the
history of the

style. The cell inside measures 6 ft. 6 in. in depth by 4 ft. 6 in.
across, on the back wall of which is a four-armed 5akti or female
divinity, probably Lakshmi, with some attendants : the dwarpalas
also are females, as are the figures on the north, east, and south
sides.

186. Draupadi's Rath. (From a Photograph.)

1 Ratha has much the same meaning as Vim&na — a chariot or covered car.

CHAP. III.

MAMALLAPURAM.

33*

The next is known as Arjuna's rath, and is a small copy
of Dharmaraja's — the last to the southwards — the only
difference being that Arjuna's is very much smaller than the
other, measuring 11 ft. 6 in. by 16 ft. in plan, and 20 ft. in
height. A cell has been excavated inside, only 4 ft. 6 in. by 5 ft.,
but it contains no image ; the figures on the outside walls,
however, seem to show that it was dedicated to Siva. It is
cracked from top to bottom, and part of the finial has fallen
off. The roofs of the lower and first stories are ornamented
with those ranges of little simulated cells which became the
distinguishing characteristics of Dravidian architecture, and it
is surmounted by an octagonal dome which is an equally
universal feature of the style.

The third — Bhima's rath — seen partially in the Woodcut
No. 185, is very re-
markable : it is an
oblong building hav-
ing a curvilinear
shaped roof with a
straight ridge. Its
dimensions are nearly
48 ft. long, 25 ft. wide,,
and 25 ft high. Ex-
ternally, it seems to
have been completely
carved, but internally
only partially exca-
vated, the works
being apparently

stopped by some accident. It is cracked completely through,
so that daylight can be seen through it, and several masses of
the rock have fallen to the ground — this has been ascribed to
an earthquake and other causes. My impression is, the explana-
tion is not far to seek, but arose from unskilfulness on the part
of workmen employed in a first attempt. Having completed
the exterior, they set to work to excavate the interior so as
to make it resemble a structural building of the same class,
leaving only such pillars and supports as were sufficient to
support a wooden roof of the ordinary construction (Woodcut
No. 187). In this instance it was a mass of solid granite
which, had the excavation been completed, would certainly
have crushed the lower storey to powder. As it was, the
builders seem to have taken the hint of the crack and stopped
the further progress of the works.

1 The dotted hatching on this plan represents a suggested mode in which the rath
might have been completed if finished as intended.

187. Plan of Bhima's Rath. (From a Plan by Mr.
R. F. Chisholm.) Scale 20 ft. to i in.1

332

DRAVIDIAN STYLE,

BOOK III,

It is a little difficult to say how it was intended to have
been completed. The centre was occupied by a hall about

9 or 10 ft wide by 30 ft.
open on one or probably both

^s"^- sides, and intended to be

closed at both ends. This
central hall was surrounded
by a verandah measuring 5 ft.
3 in. in the clear on the sides,
but only 3 ft. at the ends.
There would then have been
ten or twelve pillars in the
centre and two at each end.
One of these is represented
in the annexed woodcut (No
1 88), and they are all of the
same pattern, which, in fact,
with very slight modifications,
is universal at Mamallapuram.
They all have bases repre-
senting Vyalis or conven-
tional lions, with spreading
capitals, and of proportions
perfectly suited to a building
of the dimensions of this one,
if executed in wood.

The fourth and most
southerly, however — Dhar-
maraja's rath — is the finest
and most interesting of the
group. A view of it has already
been given (Woodcut No. 89)

and it is shown on the right hand of Woodcut No. 185. As will
be seen from the annexed plan (No. 189), its dimensions are 26 ft.
9 in., by 28 ft. 8 in., and its height is rather more than 35 ft. It
is consequently much larger than Arjuna's rath, but even with
these dimensions it can only be considered as a model. The
three upper storeys are ornamented with those little simulated
cells mentioned above, and which are so universal in the south of
India ; the front of each of these cells with their connecting links,
is adorned with a representation of one of those semi-circular
dormer windows which are so usual in Buddhist architecture.
Here each has a human head looking outwards. Behind these
cells the walls are divided by slender pilasters into narrow com-
partments, and in each is placed the statue of a deity of the
Hindu Pantheon, among which are found representations of

188.

Pillar in Bhima's Rath. (From a
Drawing by Mr. R. F. Chisholm.)

CHAP. III.

MAMALLAPURAM.

333

5iva, Brahma, and Vishnu in various characters as Arddhanari
Narasimha, Varaha, etc. — none of them with more than four

arms.1 Over these figures are seven- ^ ....

teen short epigraphs containing
epithets that were long misunder-
stood ; but from a comparison of
them with the very similar series
round the inside of the enclosure
of the Kailasanatha temple at
Kanchipuram, a clue is obtained
that enables us to fix the date of
these monuments. Stated briefly,
it stands thus : — On the east side
of the third storey of this rath is
the epigraph — " The temple of the
holy Atyantakama - Pallave^vara : ;T,l89- £lan ?f Dharmaraja Rath.

-n • » j i-u- At 4.1* (From a Drawing by Mr. R.F.Chisholm.)

Ranajaya, and this Atyantakama, Scale 20 ft. to i in.

in the other storeys is styled Nara-
simha Srinidhi, Sribhara, etc. On the monolithic temple of
Gan&ra, to be noticed below, and in the Dharmaraja Mantapa
cave are identical inscriptions of the same Atyantakama->Srinidhi-
Sribhara ; and at .Saluvankuppam cave is an inscription of
King Atiranachanda with the names of Atyantakama, Sribhara,
Ranajaya, Kalakala, etc. — which are also epithets of Rajasimha
in the Kanchi inscriptions. Now we learn from copperplate
grants that the Pallava king, Rajasimha, bore the names of
Kalakala, Narasimhavishnu and Narasimhavarman, and must
have reigned about the last quarter of the 7th century. We
can hardly escape the conclusion, then, that Rajasimha,
Atyantakama, Atiranachanda, etc., are all names (or birudas) of
one king, the son and successor of Ugradanda - Lokaditya or
Parame^varavarman.2 His dedications are all 6aiva, and
their occurrence on so many of the Mamallapuram shrines
supports the testimony, previously founded on the style alone,
that they belong to one time, and were all excavated within
a short period about A.D. 670 to 7°°-

As stated above, we have on this rath many of the gods of
the Hindu Pantheon, but in forms more subdued than are to
be found elsewhere. The one extravagance is that they gener-
ally have four arms — never more — to distinguish them from
mortals ; but none of those combinations or extravagances we
find in the caves at Eliira, Elephanta, and elsewhere. It is the

1 ' Transactions of the Royal Asiatic
Society,' vol. ii. plates 16, 17 ; or Carr's
compilation.

2 Hultzsch, ' South Indian Inscriptions,'
vol. i. pp. I -24 ; and Fleet, in ' Bombay
Gazetteer,' vol i. pt. i. pp. 322-326.

334

DRAVIDIAN STYLE,

BOOK III.

soberest and most reasonable version of the Pantheon yet dis-
covered, and, consequently, one of the most interesting, as well
probably, as the earliest1

The upper three storeys are entirely finished externally with

190.

Elevation of Dharmaraja Rath. Scale 13 ft. to i in.
Mr. R. F. Chisholm.)

(From a Drawing by

their sculptures, the lower one merely blocked out ; and it is
difficult to say how far it was intended to excavate the interior.
A cell was excavated to a depth of 5 ft. in the third storey,
and it may have been intended to enlarge it. A similar attempt
has been made in the second storey, but only to the depth of
4 ft. From there being six pilasters on the outside of the
third storey, we may gather that in a structural building its roof

1 These figures, properly drawn or photographed, would be almost indispensable
for the illustration of Hindu mythology.

CHAP, III.

MAMALLAPURAM,

335

would have been supported by thirty-six wooden posts, and in
like manner that the second storey would have had sixty-four
supports, but, of course, with some of these omitted in the centre.
From its extreme irregularity it is not easy to suggest what may

r
<

«» j-.

A. h

<» t

4 Bs.

\ — ,— ,— ^ ^ — i-r~"/ ^ — r-r~ ** if

---^-----"•- .-^-.A----^-. «-r-^-jr_--^- M

Section of Dharmaraja Rath (through A.B. fig. 189), with the suggested internal
arrangements dotted in. Scale 10 ft. to i in.

have been the intended arrangement of the lowest storey ; but,
from the wider spacing of the pillars externally, it is evident
that in a structural building stone, and not wood, would have
been employed. From the arrangement of the exterior we
gather that there would have been four free-standing pillars in
the centre, as shown in dotted lines in the plan and section
(Nos. 189 and 191). It is not clear, however, how many of the

336

DRAVIDIAN STYLE,

BOOK III.

eight piers that surrounded these four were free-standing or
attached as pilasters to thick external walls. What stopped the
completion of this and the other raths, we shall never learn.
They are certainly very like Buddhist buildings, as we learn to
know them from the early caves, and it seems hardly to admit
of doubt that we have here petrifactions of the later forms of
Buddhist architecture,1 and of the first forms of that of the
Dravidians.

The want of interiors in these raths makes it sometimes
difficult to make this so clear as it might be. We cannot, for
instance, tell whether the apsidal rath, called Sahadeva and
Nakula's, to the west of the line of the others, and
forming the fifth of the group, was intended to
reproduce a chaitya hall or a vihara like that in
Woodcut No. 63. Though small, it is one of the
most interesting of the whole ; but like the others,
it is very unfinished, especially on the east side. Its
dimensions are 18 ft. in length — north and south,
by ii ft. across, and about 16 ft. in height.
It faces north, on which side there is a small pro-
jecting portico (Woodcut No. 193), supported by
PRath!Sa(Faromaa two pillars, and within is a small empty cell. Ex-
Drawing by Mr. ternally the back end is apsidal, and so perhaps,

1'] if on a larger scale, its interior might have been :

1 ••«• «« -i , — »

in. as it is, it is too small, and the square form is

more convenient in such an apartment.
The interest of this rath lies in the fact that it represented,
on a small scale, the exterior of one of those chaitya halls, which
form so important a feature in the western groups of Buddhist
caves, of which, until the discovery of the Chezarla and Ter
structures, we had no other instance from which to judge of
what the external appearance may have been of the structural
chaityas, from which the cave-temple examples were copied.
But this rath being in several storeys, and the whole so con-
ventionalised by the different uses to which they are applied
for the purposes of a different religion, that we must not stretch
analogies too far.2

1 Among the remains found at Bharaut
is a bas-relief representing a building so
exactly like the long rath here, that there
can be no doubt that such buildings were
used in the north of India two centuries
at least before Christ, but to what
purpose they were applied is not so clear.
The one at Bharaut seems to have con-
taineH the thrones or altars of the four
last Buddhas.

2 Among the sculptures of the Gan-

dhara monasteries are several representing
fa£ades of buildings. They may be cells
or chaitya halls, but, at all events, they
are almost exact reproductions of the
fafade of this rath (see Woodcut No.
123, page 216). Being used as frame-
works for sculpture, the northern
examples are, of course, conventional-
ised ; but it is impossible to mistake
the identity of intention.

CHAP. III. MAMALLAPURAM. 337

The sixth — the Gane^a rath — situated at a distance of three-

quarters of a mile north of the others — is represented in the
woodcut (No. 194), which, strange to say, is the most nearly
finished of any, and gives a fair idea of the form these oblong
temples took. Though small, it is a singularly elegant little
temple. In plan, its dimensions are 19 ft. by n ft. 3 in., and
28 ft. in height. It is in three storeys w'th very elegant details,
and of a form very common afterwards in Dravidian architecture
VOL. I. Y

338

DRAVIDIAN STYLE.

BOOK III.

for gopurams or gateways. The roof has a straight ridge,
adorned at the ends by Saiva tri-mlas, and similar emblems

194.

Ganera Rath, M^mallapuram. (From a Photograph.)

crowned the dormer windows. The ridge was ornamented by
nine small pinnacles, which also continue to be employed.
Though entering in the side, this temple was never intended
to be pierced through.

On the back wall of the verandah is an inscription in old
and very florid characters, dedicating the shrine to Siva by
King Atyantakama- Ranajaya,1 now identified with Rajasimha
Pallava of the end of the 7th century.

What interests us most here, however, is that the square
raths are the originals from which all the vimanas in southern
India were copied, and continued to be copied nearly unchanged
to a very late period. Woodcut No. 195, for instance, repre-
sents one from Madura, erected in the i8th century. It is
changed, it is true, and the cells and some of the earlier
features are hardly recognisable ; but the wonder rather is that
eleven centuries should not have more completely obliterated all
traces of the original. There is nothing, however, in it which
cannot be easily recognised in intermediate examples, and

1 Hultzsch's 'South Indian Inscriptions,' vol. i. pp. 4-8.

CHAP. III.

MAMALLAPURAM.

339

their gradual transformation detected by any one familiar with
the subject. On the other hand, the oblong raths were halls
or porticos with the Buddhists, and became the gateways
or gopurams which are frequently — indeed generally — more
important parts of later Dravidian temples than the vimanas
themselves. They, too, like the vimanas, retain their original

195-

Perumal Temple at Madura. (From a MS. Drawing in the possession of
the late General Monteith, Madras Engineers.) No scale.

features very little changed to the present day, as may be seen
from the annexed example from a modern Tamil temple on
the opposite shore of the Gulf of Manar (Woodcut No. 196).
To all this, however, we shall have frequent opportunities of
referring in the sequel, and it will become much plainer as we
proceed.

The other antiquities at Mamallapuram, though very

340

DRAVIDIAN STYLE.

BOOK. III.

interesting in themselves, are not nearly so important for our
history as the raths just described. The caves are generally
small, and fail architecturally, from the feebleness and tenuity
of their supports. The southern cave-diggers had evidently

196. Entrance to a Hindu Ten

(From Sir J. E. Tennent's ' Ceylon.')

not been grounded in the art, like their northern compeers,
by the Buddhists. The long experience of the latter in the
art taught them that ponderous masses were not only necessary
to support their roofs, but for architectural effect ; and neither
they nor the Hindus who succeeded them in the north ever
hesitated to use pillars of two or three diameters in height,
or to crowd them together to any required extent. In the
south, on the contrary, the cave-diggers tried to copy literally
the structural pillars used to support wooden roofs. Hence,
I believe, the accident to the long rath, and hence certainly
the poor and modern look of these southern caves, which
has long proved such a stumbling-block to all who have tried
to guess their age. Their sculpture is better, and some of
their best designs rank with those of Elura and Elephanta,
to which they were anterior. The Badami sculptures, executed
in the 6th century (A.D. 579), are so similar in style with the
best examples in the Mamallapuram caves, that we had con-
cluded they could not be far distant in date, and must be placed
in the preceding century ; and this has since been supported
by the contents of the inscriptions in the Dharmaraja mantapam
at the Saluvankuppam cave.

CHAP. Ill,

MAMALLAPURAM.

341

The great bas-relief on the rock, 90 ft. by 30 ft., is perhaps
the most remarkable thing of its class in India. It is close to
the Gane^a ratha, and is locally known
as ' Arjuna's Penance,' but what it was
meant to represent is still a puzzle.
It is in two sections divided by a
crack or split in the rock, in which
are placed a great Naga wearing a
crown surrounded by a seven - fold
hood (Woodcut No. 197); under him is
a Nagini with the usual triple hood ;
and below this a large cobra's head.
The figures on both sides are directed
towards this recess. On the left there
is a small shrine below and a devotee
worshipping at it, whilst behind are
several wild beasts ; above the shrine
is a yogi standing on one foot before
a deva, four armed, with a sort of I97>
sceptre and attendants, and behind
and above him all the figures are in
pairs — male and female — with only two arms each.

Head of the Naga figure at
Mamallapuram.

On the

rock to the right, there are two large and some smaller elephants,
and above them three tiers of figures, represented as floating
through the air, all two-armed, mostly in pairs, some with birds'
legs and wings (Gandharvas), and wild beasts behind. There
seems nothing to enable us to fix its age with absolute certainty ;

342 DRAVIDIAN STYLE. BOOK III.

it can hardly, however, be doubted that it belongs to about the
8th century.

There is one other antiquity at a place called Saluvan-
kuppam, half a mile north of Mamallapuram village, which
deserves notice as a descendant of the tiger cave at Udayagiri
near Katak (Woodcut No. 272). Here, not one, but a dozen
tiger or lion heads welcome the anchorite to his abode, or rather,
the devotee to his shrine (Woodcut No. 198). Here, too, they
are conventionalised as we always find them in Chalukyan art ;
and this example serves, like every other, to show how the Hindu
imagination in art runs wild when once freed from the trammels
of sober imitation of natural things, which we found to be its
characteristic in the early stages of Buddhist art.

Here is an inscription in two different alphabets of King
Atiranachanda, who has also the birudas or epithets of Atyanta-
kama, etc., from which he appears to be identical with the
Rajasimha-Narasimha who executed the Dharmaraja rath and
probably most of the excavated shrines at Mamallapuram.

KAILAS, ELtJRA.

From the raths at Mamallapuram to the Kailas at Elura
the transition is easy, but the step considerable. At the first-
named place we have manifest copies of structures intended
originally for other purposes and used at Mamallapuram in
a fragmentary and disjointed manner. At Elura, on the
contrary, the whole is welded together, and we have a perfect
Dravidian temple, as complete in all its parts as at any future
period, and so far advanced that we might have some difficulty
in tracing the parts back to their originals without the fortunate
possession of the examples on the Madras shore.

Independently, however, of its historical or ethnographical
value, the Kailas is itself one of the most singular and interesting
monuments of architectural art in India. Its beauty and singu-
larity always excited the astonishment of travellers, and, in
consequence, it is better known than almost any other structure
in that country, from the numerous views and sketches of it that
have been published. Unlike the Buddhist excavations we have
hitherto been describing, it is not a mere interior chamber cut in
the rock, but is a model of a complete temple, such as might
have been erected on the plain. In other words, the rock has
been cut away, externally as well as internally. The older caves
are of a much more natural and rational design than this temple,
because, in cutting away the rock around it to provide an
exterior, the whole has necessarily been placed in a pit. In the
cognate temples at Mamallapuram (Woodcut No. 185) this dim"-

CHAP. III.

KAILAS, ELURA.

343

culty has been escaped by the fact that the boulders of granite
out of which they are hewn were found lying free on the shore ;
but at Elura,no insulated
rock being available, a
pit was dug around the
temple in the sloping
side of the hill, about
1 06 ft deep at its in-
most side, and half that
height at the entrance
or gopuram, the floor of
the pit being 160 ft. wide
and 280 ft. in length. In
the centre of this rect-
angular court stands the
temple, as shown in the
accompanying plan
(Woodcut No. 199), con-
sisting of a vimana, 96 ft.
in height, preceded by
a large square porch,
supported by sixteen
columns (owing prob-
ably to the immense
weight to be borne) ;
before this stands a
detached porch for the
Bull Nandi, reached by
a bridge ; and in front
of all stands the gate-
way, which is in like

manner connected with the last porch by a bridge, the whole
being cut out of the native rock. Besides these there are two
pillars, or dhwajastambhas (literally banner staves) left standing
on each side of the detached porch, and two elephants about the
size of life. Round the court there is a peristylar cloister with
cells, and above are some halls (not shown in the plan), which
give to the whole a complexity, and at the same time a complete-
ness, which never fail to strike the beholder with astonishment
and awe.

. As will be seen from the view (Woodcut No. 200), the out-
line of the vimana or .rikhara is at first sight very similar to that

199.

Kailas at Eldra. (From ' Cave Temples ot
India.') Scale 100 ft. to i in.1

1 This plan represents the temple and
surrounding shrines at the level of their
upper floors, but the surrounding court at
the lower level of the entrance. At the

upper level on the north side is the
Lanke^\ara temple, the hall 01 which is
about 75 ft. long by 50 ft. wide, exclusive
of the shrine.

344

DRAVIDIAN STYLE.

BOOK III.

of the raths at Mamallapuram, but on closer inspection we find
everything so modified at Elunl as to make up a perfect and

200. Kailas Elura. (From a Sketch by the Author.)

well-understood design. The vimana with its five surrounding
cells and the porch in front of it, with its side balconies, make
a complete Hindu temple, such as are found in hundreds in
India, and instead of the simulated cells that surround the
upper storey in the Madras example, they become realities, but
used for widely different purposes. Instead of being simulated
residences, the five cells that surround the central temple, on
the same platform, are each devoted to a separate divinity, and
complete the whole as a Sivalaya or " abode of Siva," whilst
they group most pleasingly with the central vimana.1 Here,
too, they are independent, and separated from the temple itself.

1 These shrines are now empty, but
their purposes are thus explained : North
and south of the shrine are doorways
leading out upon the platform, on which
they stand, and which forms a pra-
dakshina round the Linga shrine; and

passing out by the south door the first
shrine on the soulh was appropriated to
the Matris or seven mothers, arranged
along the back wall with Kartikaswamin
or .Siva at the left end, and Ganeja with
Bhringi at the right ; the next — at the

CHAP. III.

KAILAS, ELURA.

345

In the west corner of the north side of the court is a small
chapel (Woodcut No. 201) that is probably of an early date.
The guards or dwarpals, at the door of the cella inside the

201. Shrine of the River Goddesses. (From a Photograph.

principal temple, were the river goddesses Ganga and Yamuna ;
and this chapel was dedicated to the trio — Sarasvati, Ganga,
and Yami or Yamuna — the first, on the left, standing on a

south-east corner — was dedicated to the
disgusting Chanda, to whom the refuse
of the offerings are thrown ; on the east
is the shrine of Parvati, whose place is
just behind her lord's ; that on the north-
east belongs to Bhairava or Rudra,
"the terrible"; and the fifth, on the
north side, opposite to the Somasfttra,
or outlet for the washings of the Linga —
which it is unlawful to pass in performing
the pradakshina ritual — is the shrine of

Ganera. The Matris are often repre-
sented in .Saiva sculpture. They occur at
Eluri again in cave temples 14, 21, and
22, as also under the bridge leading to
the Nandi shrine here ; and are found at
Elephanla, Gulwada, and elsewhere. —
' Cave Temples of India,' pp. 428, 434,
453, and plate 72 ; ' Archaeological Survey
of Western India,' vol. v. pp. 39, 40,
and plate 34.

346

DRAVIDIAN STYLE.

BOOK III.

lotus flower, with foliage and birds behind her ; the central one
on a makara or conventionalised alligator ; and the third on a
tortoise, with water plants represented behind them, and richly-
carved torans above. All three
are in almost entire relief. These
figures are scarcely found in later
sculptures, whilst they are frequent
in cave temples of the 5th and later
centuries. The apartment is about
23^ ft. long by 9 deep, and n ft.
high, and has two free - standing
pillars and two in antis, of a pattern
differing entirely from those of the
temple, but of a much earlier type
than those in the Lanke^wara.
They have low bases and shafts of
sixteen sides, with cushion capitals
and bracket abaci. The frieze
above has been divided into small
compartments, with seven small
groups of figures in them. The
excavation above has only the
front of its balcony completed.

Though much damaged by
Moslim violence, the lower part
of the entrance to the court shows
a considerable advance on any-
thing found at Mamallapuram, and
an approach to what the gopurams
afterwards became, in so far at
least, as the perpendicular parts
are concerned ; instead, however,
of the tall pyramids which were
so universal in later times, the
entrance to the Kailas exhibits
only what may be called the gen
of such an arrangement. It is
the upper member of a gopuram

Scale 10 ft. to i in. way, and so small as not to

visible except from above.1

On each side of the Nandi Mandap stands a square pillar or
dhwajastambha, bearing the trisula or ensign of Siva, now much

1 In Daniell's 'Oriental Scenery' it. Being cut in the rock, no addition or
(1816), pt. iv. plate 12, the gateway is alteration could afterwards have been
shown, and in plate 1 5 the upper part of intended.

CHAP. III.

KAILAS, ELURA.

347

defaced (Woodcut No. 202). They are 49 feet high, inclusive
of the trident on the top and of very elegant proportions.1
Their analogy to the lion pillars of the Buddhists (Woodcut
No. 7) and the Chaumukh pillars of the Jains must be at
once recognised, each bearing a symbol of the creed.

In the south of India, among the Jains, as mentioned in a
later chapter, such pillars are very common, usually standing

203.

Dipclan in Dharwar. (From a Photograph.)

singly in front of the temples, and were apparently intended to
carry quadruple figures of Tirthankaras — known as Chaumukhs.
They generally consist of a single block of granite, square at
base, changing to an octagon, and again to a figure of sixteen

1 There must have been an inscription
just above the base of this one, but it
has long since disappeared. For a fuller
account of KailSs and its accessories see

' The Cave Temples of India,' pp. 448-
462, and plates 80-84 > aRd ' Archaeo-
logical Survey of \\estern India,' vol. v.
pp. 26-37, and plates i, 4, 24-31.

348 DRAVIDIAN STYLE. BOOK III.

sides, with a capital of very elegant shape. Some, however, are
circular, and, indeed, their variety is great. They range from
30 ft. to 40 ft. and even 50 ft. in height, and, whatever their
dimensions, are among the most elegant specimens of art in
southern India. Typical Jaina examples of these exist at
Sravana-Belgola and Venur in Mysore and at Guruvayankeri in
South Kanara (Woodcut, vol. ii., No. 308).

One class of the Stambhas at Hindu temples was intended
to carry lamps at festivals, of which Woodcut No. 203 represents
a specimen ; but another class — the dhwaja-stambhas — like the
above at Elura, are frequently in pairs, and bear the symbol of
the sect — the tri^ula or Garuda. Besides the well-known pillar
at Eran, a fine example of a Vaishnava Stambha, consisting of
a very lofty square tapering monolith, carved with a creeper
pattern up each face and standing on an elaborately carved
base, is found at Sompalle in South Arkat district. At
Balagami is one 35 ft. high, erected by Some^vara I.
Chalukya, about 1047, crowned by a human figure having
two birds' heads, called a Gandabherunda, — a form of Garuda.
There is a well - known example also at Jajpur (Woodcut,
vol. ii., No. 321), and another, removed from the Kanarak
temple, stands in front of the great Purt temple.

Fortunately we have now the means of determining the
age of the Kailas with some precision. Kirtivarman II. of
the Western Chalukyas was overthrown by Dantidurga the
Rashtrakuta king of Malkhed before A.D. 754, and soon after
the subjugation was completed by his successor Krishna I.
who ruled from about 757 till 783. One of the achievements
ascribed to this latter king is the construction of a wonderful
Saiva temple in " the hill Elapura," which is to be identified with
this Kailas temple.1

After the temple itself was finished fresh additions appear
to have been made from time to time in the rock walls of
the surrounding court. The large cave temple, known as
Lanke^vara, cut in the scarp on the north side of the court,
differs so markedly from the style of the central temple, that
it must be ascribed to fully one if not two centuries later.2
Some excavations, too, are quite unfinished, and works may
have gone on till the I2th century.

Considerable misconception exists on the subject of cutting

1 ' Indian Antiquary,' vol. xii. p. 229.
Thirty years ago, on the remains of paint-
ing on the roof of the small porch, was

a fragment of an inscription in which the
name "Kannara," i.e. Krishna, was still

legible in old characters.

2 For a view of the interior of Lank-
ervara, see Fergusson's ' Illustrations of
the Rock-cut Temples of India/ I7th
plate.

CHAP. III.

KAILAS, ELURA.

349

temples in the rock. Almost every one who sees these temples
is struck with the apparently prodigious amount of labour
bestowed on their excavation, and there is no doubt that their
monolithic character is the principal source of the awe and
wonder with which they have been regarded, and that, had the
Kailas been an edifice of masonry situated on the plain, it would
scarcely have attracted the attention of European travellers. In
reality, however, it is considerably easier and less expensive to
excavate a temple than to build one. Take, for instance, the
Kailas, the most wonderful of all this class. To excavate
the area on which it stands would require the removal
of about 100,000 cubic yards of rock, but, as the base of the
temple is solid and the superstructure massive, it occupies in
round numbers about one-half of the excavated area, so that
the question is simply this — whether it is easier to chip away
50,000 yards of rock, and shoot it to spoil (to borrow a railway
term) down a hillside, or to quarry 50,000 cubic yards of stone,
remove it, probably a mile at least to the place where the temple
is to be built, and then to raise and set it. The excavating
process would probably cost about one-tenth of the other. The
sculpture and ornament would be the same in both instances,
more especially in India, where buildings are always set up in
block, and the carving executed in situ. Nevertheless, the
impression produced on all spectators by these monolithic
masses, their unalterable character, and appearance of eternal
durability, point to the process as one meriting more attention
than it has hitherto received in modern times ; and if any rock
were found as uniform and as easily worked as the Indian
amygdaloidal traps, we might hand down to posterity some
more durable monument than many we are now erecting at far
greater cost.

3so DRAVIDIAN STYLE. BOOK III.

CHAPTER IV.

DRAVIDIAN TEMPLES.

CONTENTS.

Pattadakal and Dharwar Temples — Conjivaram — Tanjor — Tiruvalur —
.SYirangam — Chidambaram — Rame^varam — Madura — Tinnevelly —
Kumbakonam — Vellor and Perur — Vijayanagar.

WHEN we turn from these few scattered rock-cut examples
to the great structural temples of the style, we find their
number is so great, their extent so vast, and their variety
so perplexing, that it is extremely difficult to formulate any
distinct ideas regarding them, and still more so, as a matter
of course, to convey to others any clear idea on the subject.
To any one at all familiar with the present status of the
population of the province, the greatest wonder is how such
a people could ever have conceived, much less carried out,
such vast undertakings as these, and that so recently that
some of the greatest and boldest were only interrupted by
our wars with the French scarcely a hundred and fifty years
ago. The cause of this, however, is not far to seek. Ever
since we took possession of the country, our countrymen
have been actuated by the most beneficent intentions of
protecting the poor against the oppression of the rich. By
every means we have sought to secure the ryot in his holding,
and that he should not be called on to pay more than his
fair share of the produce of his land ; while to the landowner
we have offered a secure title to what belonged to him,
and a fixed income in money in lieu of his portion of the
produce. To a people, however, in the state of civilisation
to which India has reached, a secure title and a fixed income
only means the power of borrowing on the occasion of a
marriage, a funeral, or some great family festival, ten times
more than the borrower can ever pay, and our courts as
inevitably give the lender the power of foreclosing his mortgage
and selling the property. During the century in which this
process has been going on, the landed aristocracy have gradually
disappeared, and the wealth of the country has passed into the

CHAP. IV. TEMPLES. 351

hands of the money-lenders of the cities. The aim of the
government may have been beneficent, and may produce the
greatest happiness to the greatest number ; but in such a
community neither science, nor literature, nor art, have any
place, and religion itself becomes degraded by the status
of its votaries.

Before we interfered, the practical proprietorship of the
land was in the hands of a few princes or feudal lords, who
derived from it immense revenues they had no means of
spending, except in works of ostentation, which in certain
stages of civilisation are as necessary for the employment
of the masses as for their own glorification. In such a country
as India the employment of one-half of the population in
agriculture is sufficient to produce food for the whole, while
the other half are free for any employment that may be
available. A similar state of affairs prevailed apparently
in Egypt in the time of the Pharaohs, but with very different
results. The Egyptians had great and lofty ideas, and a
hankering after immortality, that impressed itself on all
their works. The southern Indians had no such aspirations ;
they had no history to which they could look back with pride,
and their religion was an impure and degrading fetishism. It is
impossible that anything very grand or imposing should come
out of such a state of things. What they had to offer to their
gods was a tribute of labour, and that was bestowed without
stint. To cut a chain of fifty links out of a block of granite and
suspend it between two pillars, was with them a triumph of art.
To hollow deep cornices out of the hardest basalt, and to leave
all the framings, as if of the most delicate woodwork, standing
free, was with them a worthy object of ambition, and their
sculptures are still inexplicable mysteries, from our ignorance
of how it was possible to execute them. All that millions
of hands working through centuries could do, has been done,
but with hardly any higher motive than to employ labour
and to conquer difficulties, so as to astonish by the amount
of the first and the cleverness with which the second was
overcome — and astonished we are ; but without some higher
motive true architecture cannot exist. The Dravidians had
not even the constructive difficulties to overcome which enabled
the mediaeval architects to produce such noble fabrics as our
cathedrals.

The aim of architects in the Middle Ages was to design
halls which should at the same time be vast, but stable, and
suited for the accommodation of great multitudes to witness
a lofty ritual. In their struggle to accomplish this they
developed intellectual powers which impress us still through

352 DRAVIDIAN STYLE. BOOK III.

their works. No such lofty aims exercised the intellectual
faculties of the Hindu. His altar and the statue of his god
were placed in a dark cubical cell wholly without ornament,
and the porch that preceded that was not necessarily either
lofty or spacious. What the Hindu architect craved for, was a
place to display his powers of ornamentation, and he thought
he had accomplished all his art demanded when he covered
every part of his building with the most elaborate and most
difficult designs he could invent. Much of this ornamentation,
it is true, is very elegant, and evidences of power and labour do
impress the human imagination, often even in defiance of our
better judgment, and nowhere is this more apparent than in
these Dravidian temples. It is in vain, however, we look among
them for any manifestation of those lofty aims and noble results
which constitute the merit and the greatness of true architectural
art, and which generally characterise the best works in the true
styles of the western world.

Turning from these generalities to the temples themselves,
the first great difficulty hitherto experienced in attempting
either to classify or describe them was that so very few plans of
them had been published. There are probably upwards of thirty
great Dravidian temples, or groups of temples, any one of which
must have cost as much to build as an English cathedral, some
a great deal more; but of all these there were few of which, till
lately, trustworthy plans were available. This is, of course,
irrespective of some early examples of village temples, and, it
may be, of some groups which have been overlooked. If these
temples had been built like those of the Greeks, or even as the
Christian churches in the Middle Ages, on one uniform plan,
changing only with the progress of time, one or two plans
might have sufficed ; but the fact is that, in nine cases out of ten,
the larger Dravidian temples are a fortuitous aggregation of
parts, arranged without plan, as accident dictated at the time
of their erection ; and, without plans, no adequate idea could
be conveyed to those who have not seen them.

In the south of the Bombay Presidency are some of the
earlier examples of this style. Among these the great temple
at Pattadakal, now known as Virupaksha, we learn from in-
scriptions upon it, was built by Lokamahadevi, one of the
queens of the Chalukya king Vikramaditya II., who ruled
from 733 to 747. It belongs, therefore, to the date formerly
assigned to it on archaeological grounds, as having been erected
during the 8th century.1 In plan it is almost exactly a duplicate

1 ' Archaeological Survey of Western India,' vol. i. ( 1874) pp. 31 et seqq., and plates
38-40; 'Cave Temples of India,' pp. 450, 451 ; 'Indian Antiquary,' vol. x. pp. 162-169.

CHAP. IV.

PATTADAKAL.

353

annexed wood-

of the Kailas, as may be gathered from the
cut (No. 2O4).1 Antecedently to the
discovery of their relative dates we could
readily believe the southern temple to be the
older of the two, but certainly not distant
in date, and this has been fully confirmed
by the results of more recent research.

Though not the oldest, this is the most
important as well as the largest temple of
this style in the Kanarese districts. It is
the only one now in use in Pattadakal ;
the others, mostly of great age and interest,
are used as dwellings or cattle - sheds.
Four of the larger are all of the same
style — the jikharas being all square pyra-
mids, divided into distinct storeys, and
each storey ornamented with imitation
cells, alternately oblong and square. Their
ornamentation is coarser or more archaic
than that of the later Chalukyan style, and
the domical termination of the spires is
less graceful. They are wanting, too, in
that general elegance of form and detail
which are so characteristic of the latter, but

are not without a purpose-like boldness of form, expression of
stability, and a certain amount of grandeur, though this is more
readily observable in the larger temples in the south of India
than in those of Pattadakal. If we compare it with the more
modern temples, however, it will be seen how much the latter
lost by the gradually growing steepness of outline and attenua-
tion of details. The more modern forms are not without a
certain degree of elegance which is wanting in the more ancient,
but in all the higher characteristics of design, the older are by
far the finest examples.

The Virupaksha temple stands in an enclosure 224 ft. in
length from east to west, and varying in width from 105 ft. in
front with a large gateway on the east and a smaller one on
the west (Woodcut No. 204). This court has been surrounded
by small shrines or cells, some of which remain. The temple
itself faces the east and has entrance porches also on the north
and south sides. The hall or mandap measures 50 ft. 8 in.
from north to south by 45 ft 10 in. from east to west, and its
roof is supported by sixteen massive pillars, each of one block

204. Plan of Great Temple

at Pattadakal.
Scale 100 ft. to i in.

1 There are four photographs of this temple in the ' Architectural Antiquities of
Dharvvar and Mysore,' plates 54-57 ; and one in ' Archaeological Survey of Western
India,' vol. i. plate 38.

VOL. I. Z

354 DRAVIDIAN STYLE. BOOK III.

2 ft. 5 in. square and 7 ft. 5 in. high, crowned by deep bracket

capitals. These piers are arranged in fours, as in Kailas temple,

CHAP. IV.

PATTADAKAL.

355

with pilasters against the walls. The lintels over them and the
slabs of the roof, as well as the faces of the pillars, are covered
with archaic sculptures, and the central square of the roof is
filled by a great coiled Nagaraja with five hoods, protected by a
chhattra, and two Naginis-with triple hoods are intertwined with
his tail. This mandapa is lighted by twelve perforated stone
windows — four in each wall, — an arrangement not found in
modern temples. On the inner side of the hall stand two more
square piers before the shrine, the doorway of which projects
forward, forming a passage 10 ft. in length, into the cella
which is 12 ft. square and contains the 5aiva altar. A circum-
ambulatory passage goes round this, lighted by two perforated
windows in each outside wall.

Like all the early Dravidian shrines, it is built of very large
blocks of stone closely jointed and without any cement. The
representation of the south elevation in Woodcut No. 205, will
convey a better idea than any description of the style and
appearance of the structure which, though dilapidated, is still
a striking and imposing example of the class. The base is
elaborately carved ; in large panels in the walls are numerous
representations of vSiva in various forms, and of other gods ;
and the horseshoe arch is as abundantly represented as in a
Buddhist temple.

A little to the north-west of this is the temple of Sangame^var
— now much dilapidated, but quite similar
in plan and detail, if somewhat smaller and
not so carefully finished, the mandapa is
much ruined, as shown in Woodcut No, 206.
Its interest lies in the fact that it is older
by perhaps thirty or forty years than the
preceding, having been erected in the reign
of Vijayaditya (697-733). A third large
Dravidian temple stands somewhat to the
east of north from this last, and is still
plainer — probably unfinished in its sculp-
tured ornamentation ; and to the north-
west of it is one of several deserted temples
here built in the northern Hindu style. 206.
These last two are represented on Woodcut
No. 309 (vol. ii.), which places the two forms
in vivid contrast. The building there shown on the left is this
Sangame^var temple — a storeyed pyramid of Dravidian archi-
tecture— and that on the right a tower in the northern style.
In both the base is generally of a cubical form, but in the
northern with a slight projection on each face.

In a field outside the same village is an ancient Jaina temple

Sangamejvar Temple
at Pattadakal.
Scale 50 ft. to i in.

356

DRAVIDIAN STYLE.

BOOK III.

also in this style, and probably belonging also to the 8th
century. On the east it has an open portico, 19 ft. wide by
20 ft. 3 in., the roof of which is supported on eight pillars
standing on a low screen wall with four on the floor. The
mandap or hall is 32 ft. 3 in. wide and 30 ft. 8 in. from east
to west, having four square pillars on the floor and half pillars
against the walls, with two round columns at the entrance to
the lobby or ante-chamber of the shrine — all about equally
spaced apart. The shrine is 8 ft. 11 in. wide by 9 ft. 9 in.
deep, and is surrounded by a pradakshina passage. A peculiar
feature is a stone ladder in the north aisle of the mandap leading
to the roof, in the tower on which is an upper shrine, a common
feature in most of the Jaina temples.1

Though the temple of Malegitti Sivalaya,
outside the town of Badami is but small, it is a
characteristic example of the early Dravidian
style, and is probably one of the earliest, if
not the only one, now existing in so com-
plete preservation. On this account it seems
deserving of representation (Plate VIII.). The
storeyed character of the jikhara is well defined,
its whole style is that of the Mamallapuram

207. Malegitti Temple raths, with a few sculptures in the panels,
scaie^ff to'i in. whilst the mandap is lighted by perforated

windows on each side ; and, as will be seen
from the plan, Woodcut No. 207, it has a porch in front, facing
east, supported on four very massive square
pillars. The type of these latter indicates very
distinctly that in age it is not far removed from
the period when the caves were executed,
whilst the similarity of pattern to those of the
Virupaksha temple at Pattadakal indicates a
date about the 7th century — perhaps a century
earlier than that of the great temple there.

At Aihole, the old Jaina temple known as
the Meguti temple, has lost its jikhara, but is
one of the oldest shrines in the Dravidian style
for which we have a date, since it was com-

208. Meguti Temple pleted in the reign of Pulikejin II. by Ravikirti,

at Aihole. a Jain, in A.D. 634-635.2 The arrangement

Scale 50 ft. to i in. shown jn the plarjj WOOdcut No. 2O8, is SOme-

what peculiar, the shrine being surrounded by
eight small rooms, 8 ft. wide, in place of a pradakshina passage.3

1 ' Archaeological Survey of Western
India ' vol. i. p. 35, and plate 45.

2 'Indian Antiquary,' vol. viii. p. 237.

3 Perhaps the three divisions in front
of the shrine might be regarded as one
apartment.

B

CHAP. IV. CONJIVARAM. 357

There is an antarala or vestibule, between the shrine and the
Maha-mandapa ; and the roof of the latter was supported by
sixteen square piers. It faces north, and has had an upper
shrine, as many Jaina temples have. These three examples
illustrate the different styles of early southern temples.

CONJIVARAM.

Conjivaram, or Kanchipuram, is a city where tradition would
lead us to expect more of antiquity than in almost any other
city of the south. About the middle of the 4th century, or soon
after, Samudragupta claims to have overcome Vishnugopa, the
Fallava king of Kanchi ; and about A.D. 640 Hiuen Tsiang visited
Kanchipura, the capital of Dravida, which he describes as a large
city with ruins near it ascribed to A^oka. In the kingdom he
speaks of "some hundred sangharamas and 10,000 priests of the
Sthavira school " of the Mahayana, and there were some eighty
Hindu temples and many Jaina heretics.1 Epigraphical research,
as stated above (page 306), has brought to light the names of
the Pallava sovereigns ruling here about the /th century, and
their contests with the Chalukyas of Badami and other powers.
Pulike^in II. of Badami had invaded the Pallava kingdom about
A.D. 630; Vikramaditya II., fully a century later, took Ranch?,
and has left an inscription in the Kailasanatha temple ; and
again, about 870, it was taken by Krishna III., the Rashtrakuta
sovereign.2 When the rule finally passed out of the hands of
the Pallavas into those of the Cholas, probably about the end
of the 1 2th century, Kanchipuram became the capital of the
latter dynasty. Be this as it may, the two towns, Great and
Little Conjivaram, possess groups of temples as picturesque
and nearly as vast as any to be found elsewhere.

But by far the most interesting of the Conjivaram temples
is a very early one known as Kailasanatha standing in the
fields to the west of the town.3 From its style when first
seen by the editor in 1883, he at once placed it as about coeval
with the " Seven Pagodas." Fortunately it contains several
original inscriptions, the translation of which has established
the fact that it was erected by the Pallava king Rajasimha or
Narasimhavishnu, the son of Ugradanda - Lokaditya, who was
a contemporary of Vikramaditya I. of Badami (A.D. 655-680),
so that Rajasimha must have ruled at Kanchi in the second

1 Beal's ' Buddhist Records,' vol. ii. I Indeed the notice there given of the Con-

pp. 228-230, and ' Life of Hiuen Tsiang,'
pp. 138-139.

2< Epigraphia Indica,' vol. iii. pp. 36ofT,
and vol. iv. pp. 280-281.

3 Mr Sewell did not notice this temple
in his 'Lists of Antiquities' (p. 178).

jivaram antiquities is unusually meagre.
From Sir Walter Elliot's MS. copies of
inscriptions, four are noted (Ibid. p. 180)
but without mention of the many earlier

ones.

358

DRAVIDIAN STYLE.

BOOK III.

half of the 7th century. The temple is a Saiva shrine originally
styled Rajasimhe^vara, consisting of a vimana and separate

mantapam in a court sur-
rounded by small cells, as
shown in the plan, Woodcut
No 209. At a late date the
mantapam has been joined to
the vimana by an ardhaman-
tapam, roofed on six pillars
with the entrance on the
south side, and the east face
of the old mantapam is shut
off by a wall with only a
window through it. Further,
an additional shrine has been
inserted in the east wall of
the court with a small porch
and surrounding court ; but
this is evidently a very early
addition, as are also, prob-
ably, the eight small shrines
in front of the enclosure.
The vimana contains the
usual Linga cell surrounded
by a pradakshina passage ;
and round it are seven small
attached shrines — with large

Dl%5 ajni 'jjih1 Ujl1 l«|lilWffl N and is between them — the
O three on the back facing

209. Plan of Kailasanatha Temple, Conjivaram. west, and the Others facing
(From Mr A Rea's survey.) east There are algo two

Scale 50 ft. to i in. , .

somewhat larger in front, by

the sides of the entrance to the main cell. All these are
occupied inside and sculptured outside by forms of .Siva,
Parvatt, and other devatas of the same cult, and form the
temple into a complete Sivalaya — differing mainly in details
of arrangement from what we have met with in the Kailas
temple at Elura.1 At the latter, too, the east end of the
court is surrounded by a gallery or corridor containing fully
forty .Saiva sculptures, which are here represented by a series
of small shrines quite round the court — all, except those on
the east side, facing to the east and covered inside and out
with Saiva imagery — mostly of Siva and Parvati. The little
shrines had all jikharas over them appearing over the walls

1 Ante, pp. 342ff.

X

CHAP. IV.

CONJIVARAM.

359

of the court, and with a bull or Nandi and an elephant placed
alternately between the spires. On the fronts next the court
these cells have each two pillars supported by rampant Vyalis
or horned lions in varied attitudes. The attached pillars at
the corners of the cells round the main shrine are similarly
supported ; and the outer walls are ornamented by scores of
pilasters, supported by the same conventionalised animals with
riders on them. On the west side is a gopuram or gateway with
a tower over it, but the entrance is blocked up — that on the
east being alone used now. The roof of the old mantapam
has been destroyed, but the style of the temple may be under-
stood from the Plate No. IX., which, as was to be expected, is
the same as that of the Mamallapuram Raths, already described,
and gives us a typical example of the Dravidian style about
A.D. 680.

A second example of this early Dravidian style is pre-
sented in the Vaishnava temple to the east of Conjivaram,

210. Section of Vaikuntha Perumal Temple, Conjivaram. (From a Drawing by
Mr. A. Rea.) Scale 25 ft. to i in.

known as the Vaikuntha Perumal temple.1 It stands in a
court 79 ft. 2 in. wide and 108 ft. from east to west, but at
29 ft. from the east end it is contracted to 63 ft. 3 in. wide.
To this is attached an entrance hall or portico 44 ft. 4 in.
by 50 ft. — more modern than the rest, and roofed on twenty-
four pillars. A pillared verandah runs quite round the inside

1 Vaikuntha is the heaven of Vishnu, who is usually styled Perumal (the great
one) in the Tamil country.

36°

DRAVIDIAN STYLE.

BOOK III.

of the court, the walls of which are covered with sculpture.
The outsides of the court walls are divided by pilasters on
rampant Vyalis into bays in which are niches with sculptured
pediments. The vimana in the centre of the west portion
of the court measures 47 ft. square over the walls, with a
mandapam in front, having two rows of four pillars each
across the floor. The shrine is of unusual arrangement, being
of three storeys (Woodcut No. 210); round the lower are two
passages, in the inner of which is a stair leading to the upper
floor, and the outer has a door and two windows on each of
the three exterior sides. On the first floor is a somewhat larger
shrine with one covered passage or pradakshina, round it, and
an open balcony on the roof over the outer passage of the
ground floor. In the third storey is also a shrine with an
open balcony round it over the roof of the first passage below.
In the fourth storey there is again a chamber under the large
octagonal dome that crowns the jikhara. The walls of the
lower storey of the temple are divided by pilasters into panels
filled with Vaishnava sculptures.

There are three or four other old temples of the same
style at Conjivaram, but they are small and considerably
ruined, and hardly come under our notice here.1

The more modern great Saiva temple of Ekamranatha at
Great Conjivaram possesses some first-class gopurams, though
no commanding vimana. The largest gopuram is on the south
side of the outer enclosure, and has ten storeys and a large top ;
it was built by Krishnadeva Raya of Vijayanagar (1509-1530)
and is 188 ft. high.2 It has, too, a hall of about 540 columns,
several large and fine mantapas, large tanks with flights of stone
steps, and all the requisites of a first-class Dravidian temple, but
all thrown together as if by accident. No two gopurams are
opposite one another, no two walls parallel, and there is hardly
a right angle about the place. All this creates a picturesque-
ness of .effect seldom surpassed in these temples, but deprives
it of that dignity we might expect from such parts if properly
arranged.

In Little Conjivaram is the Vaishnava temple of Varada-
raja-swamt or Arulala-Perumal, which — though smaller than the
Saiva shrine — is the wealthier, being the principal seat of the
Virishtadwaitya school of Ramanuja. The principal gopuram

1 The editor is indebted to Mr. A. Rea
of the Archaeological Survey, who, with
the consent of the Government, has
favoured him with a proof of his volume
on Pallava Architecture now in the press,
on which the above remarks are based.

1 " The like model of these lofty towers

elsewhere introduced, procured for them
the designation of Rayer gobaram, or a
tower after the Rayer's fashion — that is,
a large and lofty tower." — W. Taylor's
'Oriental Historical Manuscripts,' vol. ii.
p. 125; Campbell's 'Teluga Grammar,'
introd. p. xii.

CHAP. IV.

CONJIVARAM.

of seven storeys and about 100 ft. high is plain in style and
not plastered over like so many others. Within, to the left,
is a hall of pillars, carved in the style of the Vellor and other
temples, with figures riding on horses or hippogriffs. North
of this is the usual Teppa-kulam or sacred tank and other
buildings. Inscriptions of the beginning of the I3th century
show that the temple was then in existence, and it is probably
of still earlier date.

The Kamakshi x temple, by its architectural style, suggests

211. The Shore Temple at Mamallapuram, (From a Plan by Mr. A. Rea.)
Scale 50 ft. to i in.

that it is of considerable age, and an inscription of Harihara II.
of Vijayanagar mentions the gift of a copper door for the

1 Kamakshi is a name of DurgS. or Parvati, in the character of Manmatha or Muruga :
in some parts of southern India the worship is connected with the Holi festival.

36*

DRAVIDIAN STYLE.

BOOK III.

shrine in 1393, but till a survey with sufficient illustrations
is published we cannot speculate on its age. Two miles
south, in the hamlet of Tiruppadikunram, is a Jaina temple,
dedicated to Vardhamana, of which the shrine has an apsidal
back — which is covered with sculptures of a somewhat superior
character. It probably dates from the nth century, and con-
tains inscriptions of Chola and early Vijayanagar kings — one
commemorates a grant to it by the general of Bukka II. in 1387.
Another structure of some interest in connection with these
early Dravidian temples is what is known to Europeans as the
" Shore temple " at the " Seven Pagodas." Standing on a point
of land that juts out into the sea, it has suffered much from
high tides and sea air, and is seriously damaged. The surround-
ing walls of the court have been much ruined, but excavations
made in 1884 revealed the lines of these where they had fallen,
and of other buildings in the west half of the court. The plan
(Woodcut No. 21 1) will show the disposition of the whole, and
the photograph of the temple itself from the north-west (Plate
No. X., Frontispiece] gives a view that makes the style intelligible.
It may be somewhat later than the two earliest temples at
Conjivaram, described above, but if so it can hardly be ascribed
to a later date than the Qth century A.D. At present, it contains
two shrines unsymmetrically disposed — the smaller, facing the
west, being dedicated to Vishnu, and the larger, entered' from
the east, now contains the Saiva Lingam.1

TANJOR.

One great exception to the rule that the larger Dravidian
temples are arranged as accident dictated is to be found at
Tanjor. The great Pagoda there was commenced on a well-
defined and stately plan, which was persevered in till its
completion. As will be seen from the annexed diagram
(Woodcut No. 212) it consists of two courts,2 one a square of
about 250 ft., originally devoted to minor shrines and residences ;
but when the temple was fortified by the French in I7773 it was
converted into an arsenal, and has not been re-appropriated to
sacred purposes. The temple itself stands in a courtyard
extremely well proportioned to receive it, being about 500 ft.
long by half that in width, the distance between the gateway

1 The temple is sometimes called
Sthalajayanaswami, perhaps because in a
chamber about 12 ft. long, behind the
larger shrine, but accessible from it, is
a gigantic figure of Vishnu lying on the
floor.

2 As the plan is only an eye-sketch,
and the dimensions obtained by pacing,
it must not be too much relied on. It
is sufficient to explain the text, and that
is all that is at present required.

3 Inscription on gateway.

CHAP. IV.

TANJOR.

363

and the temple being broken by the shrine of the Bull Nandi,1

which is sufficiently important for its purpose, but not so much

so as to interfere with the effect of the

great vimana, which stands near the inner

end of the court. The perpendicular part

of its base measures 82 ft. square, and is

two storeys in height, of simple outline,

but sufficiently relieved by niches and

pilasters. Above this the pyramid rises

in thirteen storeys to the summit, which

is crowned by a dome said to consist of a

single stone, and reaching a height of

190 ft. The porch in front is kept low,

and as will be seen from the woodcut

(No. 213) the tower dominates over the

gopurams and surrounding objects in a

manner that imparts great dignity to the

whole composition.

The great temple is dedicated to the
worship of Siva, as Brihatfavara, in the
form of an enormous Linga, and every-
thing in the inner courtyard belongs to
the S"aiva cult. The outer gopuram is of
much later date, and probably belongs to
the early part of the i6th century, when
Vaishnava ideas were more prevalent,
and the mythological representations had
become mixed ; the smaller shrines in the
court also belong to later dates. But, as
one of the oldest and best preserved
examples of Dravidian art, its date is
of much archaeological interest. Strange
to say, however, this date was long
obscured by assertions that had no

scientific basis. Thanks to Dr. Hultzsch's careful translations of
the long inscriptions, in old Tamil, that cover the base of the
central shrine, this is no longer doubtful. The inscriptions
enumerate the many gifts of gold images, vessels, and ornaments
made to the temple by Rajaraja or Ko-Rajakejarivarman, his
eldest sister Kundavaiyar and others, in the 25th, 26th, and

212.

Diagram Plan of Tanjor
Temple. (From a Sketch
by the Author.) Scale
200 ft. to i in.

1 The dimensions of this image are
16 ft. from muzzle to rump, by above
7 ft. across, 12 ft. 2 in. to top of head,
10 ft. 4 in. to top of hump, and 7 ft. 5 in.
to top of back. It is composed of a
single block of stone, I believe granite,
but it has been so frequently and so

thoroughly coated with oil, which is
daily applied to it, that it looks like
bronze. I tried to remove a portion of
this epidermis in order to ascertain what
was beneath, but was not successful. No
other kind of stone, however, is used in
any other part of the temple.

364

DRAVIDIAN STYLE.

BOOK III.

2Qth years of his reign ; and these gifts were presented to " the
stone temple of Rajaraje-rvara " which, we are told, he " caused
to be built at Tanjor."1 Now Rajarajadeva Chola, the son of

213.

View of the Great Temple at Tanjor. (From a Photograph by Middleton
Rayne, Esq., C.E.)

Parantaka II., became king in A.D. 985, and must have completed
the temple some time before 1012 ; as it would take some years
to build, we may safely assume that it was begun about the
end of the loth century, and completed about ioi2.2

1 Hultzsch's ' South Indian Inscrip-
tions,' vol. ii. pp. 8 and 68. Conf.
'Archaeology in India,' p. 25.

2 It is only within the last few years
that we have been able to obtain any-
thing like a correct list of the Chola

CHAP. IV.

TANJOR.

365

Besides the great temple and the Nandi porch there are
several other smaller shrines in the enclosure, one of which,
dedicated to Subrahmanya or Karttikeya, a son of Siva's, is as
exquisite a piece of decorative architecture as is to be found
in the south of India, and though small, almost divides our
admiration with the temple itself (Woodcut No. 214). It is
built behind an older shrine, which may be coeval with the great
temple as originally designed. But this is evidently of more

214.

Temple of Subrahmanya, Tanjor. (From a Photograph.)

recent date, — probably two centuries more modern than
the principal temple. The woodcut No. 215 of one of the
piers in the verandah in front of the temple, when compared
with that given below (p. 387), from Tirumal Nayyak's chaultri,
shows at a glance that they belong to about the same period,

kings ; the researches of the late Professor
Kielhorn, C.I.E., of Gottingen, based on
the epigraphical labours of Dr. Hultzsch
and Mr. B. L. Rice, C.I.E., supply the
following list of the principal rulers : —
A.D. 907 Parantaka I.

Rajaditya Muvadi Chola.
Parantaka II.
985 Rajaraja I.
1012 Rajendra Chola I.

1018 Rajadhiraja I.
1052 Rajendradeva.
1063 Virarajendra.
1070 Kulottunga Chola I.
1118 Vikrama Chola.

Kulottunga Chola II.
1146 Rajaraja II.
1178 Kulottunga Chola III.
1216 Rajaraja III.
1246-1267 Rajendra Chola III.

366

DRAVIDIAN STYLE.

BOOK III.

and this may be assigned with confidence to the i;th century.
The central pillar is alternately square and octagon, with shafts

attached on the two
side faces, and the
whole very richly
ornamented.

The Tanjor temple
>in 1758 was besieged
in vain by Lally from
1 where now the district
Court stands, and in
1771 the English be-
sieged and took it.
Many of the sculp-
tures show traces of
these sieges, and after
the last the temple
was turned into a
camp for thirty years.
In 1801 - 1802 Raja
Sarfoji had it purified
and re-consecrated.1

There is another
temple at the village
of Gangaikonda-
puram, 38 miles to
the north - east of
Tanjor, and 20 miles
south-west from
Chidambaram, that,
so far as is yet known,
must be at least as
old as that at Tanjor, if not older, and of which it would be
very desirable to have a complete survey, as it is said to retain
even more of its original purity of design than the latter.3

TlRUVALUR.

The temple at Tiruvalur in Tanjor district,4 about 1 5 miles
west of Negapattam, contrasts curiously with that at Tanjor
in the principles on which it was designed, and serves to
exemplify the mode in which, unfortunately, most Dravidian
temples were aggregated.

Pier in Subrahmanya Temple, Tanjor.2

1 Dr. A. Burnell in an article, 1 2th
November 1877.

2 From 'Technical Art Series,' 1894.

3 'Journal of the Asiatic Society of

Bengal,' vol. xlix. (1880) pp. 1-4.

4 At Tiruvallur, in Chir.galpat district,
is a Vaishnava temple dedicated to Vira-
raghava.

CHAP. IV.

TIRUVALUR.

367

The nucleus here was a small village temple (Woodcut No.
216), drawn to the same scale as the plan of Tanjor in Woodcut
No. 212. It is a double shrine, dedicated to
Valmikesvara or Siva and his consort, standing
in a cloistered court which measures 191 ft. by
156 ft. over all, and has one gopuram in front.
On the south is a shrine of Tyagaraja-swami.
The central shrine is said to belong to the early
years of the I5th century; but there are some
defaced inscriptions of Rajaraja I. and Rajendra
Chola (A.D. 985-1018) on the small shrine of 2I6. innerTempie
Achale^vara, which may be the oldest portion at Timvaiur.

. , . ,1 • ,i . . /. Scale 200 ft. to i in.

now existing ; otherwise there is nothing to
distinguish it from the ordinary temples found in every village.
It, however, at some subsequent period became sacred or rich,
and a second court was added, measuring about 470 ft. each way,
with two gopurams, higher than the original one, and containing
within its walls numberless little shrines and porches. Additions

217. Temple at Tiruvalur. (From a Drawing in Ram Raz's ' Hindu Architecture.')

were again made at some subsequent date, the whole being
enclosed in a court 957 ft. by 726 ft. — this time with five
gopurams, of which those on the west and east or front are
respectively 101 and 118 feet high, and containing several

368 DRAVIDIAN STYLE, BOOK III,

subordinate shrines. When the last addition was made, it was
intended to endow the temple with one of those great halls
which were considered indispensable in temples of the first
class. Generally they had — or were intended to have — 1000
columns ; this one has only 807, and almost one-half of these
mere posts, not fitted to carry a roof of any sort. There can,
however, be very little doubt that, had time and money been
available, it would have been completed to the typical extent.
The general effect of such a design as this may be gathered
from the bird's-eye view (Woodcut No. 217). As an artistic
design, nothing can be worse. The gateways, irregularly spaced
in a great blank wall, lose half their dignity from their posi-
tions ; and the bathos of their decreasing in size and elaboration,
as they approach the sanctuary, is a mistake which nothing can
redeem. We may admire beauty of detail, and be astonished at
the elaboration and evidence of labour, if they are found in such
a temple as this, but as an architectural design it is altogether
detestable.

SRIRANGAM OR SERINGAM.

The temple which has been most completely marred by
this false system of design is the great Vaishnava temple at
Srirangam, 2| miles north of Trichinopoly, which is certainly
the largest, and, if its principle of design could be reversed,
would be one of the finest temples in the south of India (Wood-
cut No. 219, p. 371). Here the central enclosure is quite as
small and as insignificant as that at Tiruvalur, and except that
its dome is gilt has nothing to distinguish it from an ordinary
village temple. The plan (Woodcut No. 218) of the inner four
courts will explain its arrangements. The fourth enclosure,
however, is the most magnificent. It encloses the hall of 1000
columns (S), which measures some 500 ft. by 138 ft. The
number of columns is eighteen in front by sixty-three in depth,
and 953 altogether.1 They consequently are not spaced more
than 10 ft. apart from centre to centre ; and as at one end the
hall is hardly over 10 ft. high, and in the loftiest place only
20 ft, and most of the pillars being spaced nearly evenly over

1 The plan is from Capt. Cole's, re- mantapam ; I, 6ri-pandara mantapam ;
produced in ' India : Photographs and J, Surya - pushkarani ; K, Ramasvami

Drawings of Historical Buildings' (Griggs,
1896), plate 53. References to the plan,
Woodcut No. 218:— A, South, or Kurat
Alwar Gopuram ; B, East, or Vellai
Gopuram ; and C, North, or Nachiyar-
sanadi Gopuram, of the fourth court ;
D, Vaikuntha Gopuram of the second
court. E, the shrine, which, curiously
enough, is circular ; F, Chandan man-

temple; L, Vaikunth temple ; M, Chandra-
pushkarani ; N, Dhanvantari temple ;
O, O, Vasudeva Perumal temples ; P,
Narasimha Perumal temple ; Q, Q,
Granaries ; R, Ranganayaki temple ;
S, The Thousand Pillar mantapam ; T,
•Seshagiri-rao mantapam ; U, Ranga-
vilasam mantapam; V, Kurat -Alwar
temple ; W, W, Krishna temples ; X,

tapam ; G, Yajna Sala ; H, Garuda Elephant stabl es ; Y, Kada-kili mantapam.

2i8. Plan of Srirangam Temple — the four inner courts. Scale 200 ft. to i in.
VOL. I. 2 A

37°

DRAVIDIAN STYLE.

BOOK III.

the floor, it will be easily understood how little effect such a
building really produces.1 They are, however, each of a single
block of granite, and all carved more or less elaborately. A
much finer portico stretches across the third court of the
Jambuke^vara temple from gopuram to gopuram ; the pillars
in it are much more widely spaced, and the central aisle is
double that of those on the sides, and crosses the portico in
the centre, making a transept ; its height, too, is double that
of the side aisles. It is a pleasing and graceful architectural
design ; the other is only an evidence of misapplied labour.
The Seshagiri-rao Mantapam (T), to the south of the preceding,
is the most elaborately carved hall in the temple. It is
supported by pillars with rearing horses and other figures in
front, similar to those in the Madura temple, and is probably
of about the same age — or the middle of the i/th century.
The Gopuram (B), on the east side of this court is 146^ feet
in height, and is one of the finest in the temple.

The three outer enclosures have nothing very remarkable in
them, being generally occupied by the Brahmans and persons
connected with the temple. Each, however, has, or was intended
to have, four gopurams, one on each face, the superstructures
are of brick, and some of these are ornamented in plaster of
considerable magnificence. The outer enclosure is, practically,
a bazaar, filled with shops, where pilgrims are lodged, and
fed, and fleeced. The wall that encloses it measures 2,521 ft.
by 2,865 ft,2 and, had its gopurams been finished, they
would have surpassed all others in the south to the same
extent as these dimensions exceed those of any other known
temple. The unfinished southern gopuram, leading to the
river and Trichinopoly, measures 130 ft. in width by 100 ft.
in depth; the opening through it measures 21 ft. 6 in., and
twice that in height. The four jambs or gateposts are each
of a single slab of granite, more than 40 ft. in height, and
the rooftng-slabs throughout measure from 23 ft. to 24 ft. Had
the ordinary brick pyramid of the usual proportion been added
to this, the whole would have risen to a height of nearly 300
ft Even as it is, it is one of the most imposing masses in
southern India, and probably — perhaps because it never was

1 A plan and section of the Thousand -
pillared mantapam and six plates of
pillars are given in the 'Journal of
Indian Art and Industry,' vol. viii.
(1899), plates 89-95, and the same with
two others, in ' India: Photographs,' etc.
ut supra, plates 54-62.

2 The innermost court, enclosing the
temple, measures 240 ft. from north to
south by 181 ft. from east to west ; the

second 426 ft. by 295 ft. ; the third 767 ft.
by 503 ft. ; the fourth 1,235 ft. by 849 ft. ;
the fifth which, with the remaining two,
is occupied by houses, measures 1,653 ft.
by 1,270; the sixth 2,108 ft. by i, 846ft.;
and the seventh is 2,521 ft. over all at the
south end and about 2,485 at the north, by
2,865 ft- in length. — ' Madras Manual of
Administration,' vol. iii. p. 833 ; and
Major Cole's plan.

CHAP. IV.

5RIRANGAM.

372 DRAVIDIAN STYLE. BOOK III.

quite finished — it is in severe and good taste throughout.1 Its
date, fortunately, is perfectly well known, as its progress was
stopped by its being occupied and fortified by the French
during our ten years' struggle with them for the possession of
Trichinopoly ; and if we allow fifty years for its progress, even
this would bring the whole within the limits of the i8th century.
The other three gopurams of this enclosure are in the same
style, and were commenced on the same scale, but not being
so far advanced when the work was stopped, their gateposts
project above their walls in a manner that gives them a very
singular appearance, and has led to some strange theories as
to their design.

Looked at from a distance, or in any direction where the
whole can be grasped at once, these fourteen or fifteen great
gate towers cannot fail to produce a certain effect, as may be
gathered from the view in Woodcut No. 219; but even then it
can only be by considering them as separate buildings. As
parts of one whole, their arrangement is exactly that which
enables them to produce the least possible effect that can be
obtained either from their mass or ornament. Had the four
great outer gopurams formed the sides of a central hall or court,
and the others gone on diminishing, in three or four directions,
to the exterior, the effect of the whole would have been increased
in a surprising degree. To accomplish this, however, one other
defect must have been remedied : a gateway even 1 50 ft. wide
in a wall nearly 2000 ft. in extent is a solecism nothing can
redeem ; but had the walls been broken in plan or star-shaped,
like the plans of Chalukyan temples, light and shade would
have been obtained, and due proportions of parts, without
any inconvenience. But if the Dravidians ever had it in them
to think of such things, it was not during the i/th and i8th
centuries, to which most things in this temple seem to belong.
The shrines and inner prakaras, however, must be of much
earlier date, for we find Jatavarman Sundara Pandya, about
1254, making large gifts and additions to the temple; and in
1371 Kampana Udaiyar of Vijayanagar was engaged repair-
ing it.8

As mentioned above, the great Vaishnava temple of 5rirangam
owes all its magnificence to buildings erected during the reign of
the Nayyak dynasty, whose second capital was Trichinopoly, and
where they often resided. Within a mile to the east, however,
of that much-lauded temple is another, dedicated to Siva, under

1 A drawing of it was published in i 2 See the inscriptions from the temple

my ' Picturesque Illustrations of Indian j in ' Epigraphia Indica,' vol. iii. pp. 7^- >

Architecture.' It has since been frequently vol. vi. pp. 324, 330; and vol. vii. pp.

photographed. 176, 177.

CHAP. IV.

CHIDAMBARAM.

373

the title of Jumbuke^wara, which, though not so large as that
dedicated to Sri Ranganatha, far surpasses it in beauty as an
architectural object. The east gateway of the outer enclosure
is not large, but it leads direct to the centre of a hall containing
some 250 pillars. On the right these open on a tank fed by a
perpetual spring, which is one of the wonders of the place.1
The corresponding space on the left is occupied by about
470 columns : these together form the Thousand - pillared
mantapam. Between the gopurams of the second enclosure
is a very beautiful portico of cruciform shape, leading to the
door of the sanctuary, which, however, makes no show externally,
and access to its interior is not vouchsafed to the profane. The
age of the courts of this temple is somewhat earlier than that of
its great rival, and being all of one design, they probably were
begun and completed at once, and from the simplicity of its
parts and details must be earlier than the great buildings of
Tirumalai Nayyak. In fact an inscription on the south wall of
the second prakaram is dated in the tenth year of Jatavarman
Sundara Pandya, who began to reign in 1251 ;2 and, though
the outer enclosures were subsequently added — probably as
late as A.D. 1600 — the nucleus of the buildings must be about as
old as the I2th century, and possibly even two centuries earlier.3
One of the great charms of this temple, when I visited it,
was its purity. Neither whitewash nor red nor yellow paint had
then sullied it, and the time-stain on the warm-coloured granite
was all that relieved its monotony ; but it sufficed, and it was a
relief to contemplate it thus after some of the vulgarities I had
seen. Now all this is altered. Like the pagodas at Rame^varam,
and more so those at Madura, barbarous vulgarity has done its
worst, and the traveller is only too fully justified in the contempt
with which he speaks of these works of a great people which
have fallen into the hands of such unworthy successors.

CHIDAMBARAM.

The Saiva temple at Chidambaram in South Arkot district
is one of the most venerated, and has also the reputation of
being one of the most ancient temples in southern India. It
was there, therefore, if anywhere, that I at one time hoped to
find some remains that would help to elucidate the history of
the style. It was, besides, so far removed from any capital
city or frequented haunt of man that one might hope to find
its original form unaltered.

1 The view in this temple in my ' Pic-
turesque Illustrations of Indian Archi-
tecture,' No. 21, is taken from the corner
of this tank.

8 ' Indian Antiquary,' vol. xxi. p. 121
and vol. xxii. p. 221.

3 Hultzsch's ' South Indian Inscrip-
tions,' vol. ii. p. 253.

374 DRAVIDIAN STYLE. BOOK III.

It is old, but, probably, not older than the Tanjor and
Gangaikondapuram temples. The Kongade^a Rajakal relates
a legend of Vira Chola Rayar who " one day saw on the sea-
shore the Sabhapati of Chidambara (Siva), attended by Parvati,
dancing and beating the damaraka (a kind of drum) ; he there-
fore expended great sums of money in building the Kanaka,
or Golden Sabha." l A little further on, it is said, " Arivari-
deva (A.D. 985) observing that his grandfather had built only
a Kanaka-Sabha to the Chidambara deity, he built gopurams,
maddals (enclosures), madapanas (image-houses), sabhas (holy
places or apartments), and granted many jewels to the deity."
Though this work is not at all trustworthy, yet from an inscrip-
tion we learn that Viranarayana or Parantaka I., early in the
loth century, covered the hall with gold, or erected the
Kanaka-sabha. There is also in the temple an inscription of
Rajadhiraja I. (1018 to 1052) which may be regarded as
evidence that the two inner enclosures, BB, at the west end
of the tank (Woodcut No. 220) were in existence in the loth
century.2 They, indeed, measuring about 320 ft. square, appear
to have been the whole of the original temple, at least in the
loth and nth centuries. On the whole, I am inclined to
believe that this inner temple is really the one referred to in
the above extract as far as supported by the inscription. The
temple of Parvati, C, on the north of the tank, was added
afterwards, most probably in the I4th or I5th century, and
to that age the great gopurams and the second enclosure also
belong. The north gopuram, 140 feet high, is ascribed to
Krishnadeva, about 1520, whilst that on the east has inscriptions
of Sundara Pandya about 1250. The hall of 1000 columns,
E, was almost certainly erected between 1595 and 1685, at
which time, we learn from the Mackenzie MSS., the kings of
the locality made many donations to the fane.3 It was then,
also, in all probability, the outer enclosure was commenced ;
but it never was carried out, being in most places only a few
feet above the foundation.

The oldest thing now existing here is a little shrine in the
inmost enclosure (opposite A in the plan). A porch of fifty-
six pillars about 8 ft. high, and most delicately carved, resting

1 'Journal of the Royal Asiatic Society,'
vol. viii. p. 7. The Chidambaram temple
is dedicated to one of the Pancha-lingams
or five notable symbols of Siva in southern
India. These are — (i) at Conjivaram,
the " prithvi-lingam," made of earth, —
claimed also by the Tiruvalur temple ;
(2) at Jambukejvaram, the "apa-lingam,"

exuding water; (3) at Tiruvannamalai in j No. 20, p. 15.
S. Arkat, the " tejo-lingam," sparkling

with light ; (4) at Kalahasti in N. Arkat,
the " vayu-lingam," of which the lamp
vibrates with the wind ; and (5) this at
Chidambaram is the " akaja-lingam," of
ether — having no material representation.

2 ' Epigraphia Indica,' vol. iii. pp.
280-281.

3 ' Madras Journal,' vol. viii. (1838),

CHAP. IV.

CHIDAMBARAM.

375

on a stylobate, ornamented with dancing figures, more graceful
and more elegantly executed than any others of their class, so

Sca/e of Fe-t 260 ft. to i in.

220. Plan of Temple of Chidambaram.

(From a Plan by Admiral Paris, in ' Tour du Monde,' vol. xvi. p. 35.)

far as I know, in southern India. At the sides are wheels and
horses, the whole being intended to represent a car, as is
frequently the case in these temples. Whitewash and modern
alterations have sadly disfigured this gem, but enough remains
to show how exquisite, and consequently how ancient, it was.
It is the Nritya or Nritta Sabha, the hall of the dance, in
allusion, probably, to the circumstance above mentioned as

DRAVIDIAN STYLE.

BOOK III.

leading to the foundation of the temple,1 as well as to the
chief idol of the temple — Nate^a.

In front of the central shrine is one of very unusual
architecture, with a tall copper roof, which, I have no doubt,

221. View of Porch of Chidambaram. (From Drawings by the Author.)

represents or is the golden or Kanaka-sabha above referred
to, and in front of this is a gopuram and pillared porch, making
up what seems to have been the original temple. The central

1 The Nathukottai Chettis, who manage (or mismanage) the temples, propose to
move this Nritya Sabha to make room for the extension of a new cloister they are now
making.

CHAP. IV.

CHIDAMBARAM.

377

temple is dedicated to Nate^a or Siva as god of the dance,
and is a plain wooden building standing on a stone pavement ;
but behind it an apartment of polished black stone has recently
been added, part of the roof of which is formed of gilt plates.
The outer enclosure, with the buildings it contains, are, it
appears, of later date.

The temple of Parvati, C, is principally remarkable for its
porch, which is of singular elegance. The preceding woodcut
(No. 221) gives some idea of its present appearance, and the
section (Woodcut No. 222) explains its construction. The

222. Section of Porch of Temple at Chidambaram. (From a Sketch by the Author.)

No Scale.

outer aisles are 6 ft. in width, the next 8 ft, but the architect
reserved all his power for the central aisle, which measures
21 ft. 6 in. in width, making the whole 50 ft or thereabouts.
In order to roof this without employing stones of such dimen-
sions as would crush the supports, recourse was had to vaulting,
or rather bracketing, shafts, and these brackets were again tied
together by transverse purlins, all in stone, and the system
was continued till the width was reduced to a dimension that
could easily be spanned. As the whole is enclosed in a
court surrounded by galleries two storeys in height, the effect
of the whole is singularly pleasing.

Opposite to this, across the Sivaganga tank, is the Raja-
sabha or hall of 1000 columns, E, similar in many respects to
that at Srirangam, above described, but probably slightly more
modern. It is about 197 ft. wide by 338 ft. in length. Here
the pillars are arranged twenty-four in front by forty-one in
depth, making 984 ; but in order to get a central space, four in
the porch, then twenty-eight, then two, and again twenty-four,
have been omitted, altogether fifty-eight ; but, on the other

DRAVIDIAN STYLE

BOOK III.

CHAP. IV. CHIDAMBARAM. 379

hand, those of the external portico must be added, which nearly
balances the loss, and makes up the 1000. It must be confessed
this forest of granite pillars, each of a single stone, and all more
or less carved and ornamented, does produce a certain grandeur
of effect, but the want of design in the arrangement, and of
subordination of parts, detract painfully from the effect that
might have been produced. Leaving out the pillars in the
centre is the one redeeming feature, and that could easily have
been effected without the brick vaults, formed of radiating
arches, which are employed here — another certain proof of the
modern age of the building. These vaults are certainly integral,
and as certainly could not have been employed till after the
Muhammadans had settled in the south, and taught the Hindus
how to use them.

Although this temple has been aggregated at different ages,
and grown by accident rather than design, like those at
Tiruvalur and ^Srirangam just described, it avoids the great
defect of these temples, for though like them it has no tall
central object to give dignity to the whole from the outside,
internally the centre of its great court is occupied by a tank,
round which the various objects are grouped without at all
interfering with one another. The temple itself is one
important object, to the eastward of it : the Parvatt temple,
another on the north, and forms a pleasing pendant to the
looo-columned chauari on the south. On the north side of
the Parvati another temple (D) was commenced (Woodcut
No. 220), called the Pandyanayaka-kovil, with a portico of
square pillars, four in front, and all most elaborately ornamented,
but in such a manner as not to interfere with their outline
or solidity. For long it stood in an unfinished and ruined
state, but has of late been restored. This temple was dedicated
to Shanmukha or Subrahmanya ; but we cannot feel sure of
its age. From its position, however, and the character of its
ornamentation, there seems little doubt that it belongs to the
end of the i/th and first half of the i8th century. From its
style, however, I would be inclined to ascribe it to the earlier
date. The main buildings are enclosed within high walls
of dressed granite ; outside are four " car streets," 60 feet
wide ; and the whole area belonging to the temple is about
39 acres.

A large portion of the innermost area, which is historically
the most important, has of recent years been undergoing
elaborate restoration and important additions at the hands
of the^Nathukottai Chettis, thafunfortunately will quite obliterate
much that is most important for archaeology. They are adding
a wide cloister intended to run all round the enclosure, but the

380 DRAVIDIAN STYLE. BOOK III.

Vaishnava Amman temple stands in the way of this, and the
Chettis have been interdicted from removing it. This court
contains also the Chita-sabha or central shrine — a plain wooden
building, as already mentioned, on a stone platform. In it Siva
is represented as Nataraja or Nate^a. Behind is the bedroom
for the idol — a new structure of polished black stone.

The special Lingam of the temple is supposed to be the
Akasa or " ether " lingam, which is said to stand behind the idol
with a curtain and string of Bilva leaves suspended in front of it.
In front of the Chita-sabha is the Kanaka-sabha, also of wood,
where the daily worship of the Nataraja is performed. Almost
adjoining is the temple of Vishnu or Ranganatha. Besides
these and the Nritta-sabha already mentioned, is the Deva
Sabha used as an office, and near it a small building or old
shrine, believed to be the original structure, round which the rest
of the temple has grown.1

RAMESVARAM.

If it were proposed to select one temple which should
exhibit all the beauties of the Dravidian style in their
greatest perfection, and at the same time exemplify all its
characteristic defects of design, the choice would almost
inevitably fall on that at Rame^varam, in the island of
Pambam2 (Woodcut No. 224). In no other temple has the
same amount of patient industry been exhibited as here, and
in none, unfortunately, has that labour been so thrown away
for want of a design appropriate for its display. It is not
that this temple has grown by successive increments like
those last described ; for it was finished on a settled plan,
as undeviatingly carried out as that at Tanjor, but on a
principle so opposed to it, that while the temple at Tanjor
produces an effect greater than is due to its mass or detail,
this one, with double its dimensions and ten times its elabora-
tion, produces no effect externally, and internally can only be

1 Francis, 'Gazetteer of South Arcot,' erections in a nondescript style of build-
vol. i. pp. 270, 271. ing. Mr A. R. Gopalaiyar manfully

2 Strictly speaking — the temple that opposed the manager in the Courts, but
till recently was here — for, like Chid- in vain ; they would not interfere with
ambaram and other 5aiva temples in the manager's plans, though he leased
southern India, the Nathukottai Chettis out the temple property to his own rela-
(one of their number having got the lives, or destroyed the statues of former

management into his hands), have set
about demolishing it, and it is reported
that already almost the whole of the
interior or oldest portions have been
pulled down, and are to be replaced by

patrons and benefactors of the temple to
substitute those of his wealthy caste
fellows, that they may have the merits
accruing to temple building hereafter.

CHAP. IV.

RAMESVARAM.

w.

*

381

224. Plan of the Great Temple at Ramejvaram, before 1905. Scale 168 ft. to i in.

A Gandhamadana old temple.

B Great Gopuram and entry to the second Prakaram.

C Small Gopuram for the Amman shrine.

N North, S, South, and W, West gopurams.

38=

DRAVIDIAN STYLE.

BOOK III.

seen in detail, so that the parts hardly in any instance aid one
another in producing the effect aimed at.

The only part of the temple, outside the central prakaram, at
least, which is of a different age from the rest, is a small
vimana, known as Gandhamadhane-rvara (A), of very elegant
proportions, that stands in the garden, on the right hand
of the visitor as he enters from the west. It has, however,
been so long exposed — like the temple on the shore at
Mamallapuram — to the action of the sea-air, that its details
are so corroded they cannot now be made out, and its age
cannot consequently be ascertained from them. It is safe,
however, to assert that possibly it may be of the nth or
1 2th century. Its dimensions may be roughly 50 ft. in
height, by 34 ft. in plan, so that it hardly forms a feature in
so large a temple. From the bulls that occupy the platform
in front, and from its name, it is evident it was originally
dedicated to Siva, as the central temple apparently is, though
the scene of Rama's most celebrated exploit, and bearing his
name.1

Externally the temple is enclosed by a wall 20 ft. in height,
and possessing four gopurams, one on each face, which have this
peculiarity, that they alone, of all those I know in India, are
built wholly of stone from the base to the summit. The western
one (W) alone, however, is finished, and owing apparently to
the accident of its being in stone, it is devoid of figure-sculpture
— the plaster casts that now adorn it having been added in
recent times : it is 78 ft. in height. Those on the north and
south (N and S) are hardly higher than the wall in which
they stand, and are consequently called the ruined gateways.
They are, however, of comparatively modern date, and, in
fact, have never been raised higher, and their progress was
probably stopped in the beginning of the last century, when
Muhammadan, Maratha, and other foreign invaders checked the
prosperity of the land, and destroyed the wealth of the priest-
hood. The eastern facade of the second prakaram has two
entrances and two gopurams, the smaller, C, to the south of the
large one and leading into the Parvati temple, is finished. The
larger one (B in the plan) is ascribed to 6adayaka Tevar
Dalavay, who was the Setupati about 1640, but it never was

1 In the north, the affix sw&min to the
names would indicate a Vaishnava cult,
but not so here where the chief images
— Ramalingervaraswamin and Parvata-
vardhini in the central temples, and
Vijvanathaswamin and Vnalakshi in the
secondary shrines, are all essentially

.Saiva, and the Linga occupies the chief
cell. The shrine of Sabhapati — a name
given to 5iva at Chidambaram — is
ascribed to Dalavay Setupati. For an
account of the ritual of the Ramervaram
temple, see ' Indian Antiquary,' vol. xii.
(1883), pp. 315-326.

CHAP. IV. RAMESVARAM. 383

carried higher than we now see it. Had it been finished,1 it
would have been one of the largest of its class, and being wholly
in stone, and consequently without its outline being broken by
sculpture, it would have reproduced more nearly the effect of an
Egyptian propylon than any other example of its class in India.
The glory, however, of this temple resides in its corridors.
These, as will be seen by the plan, extend to nearly 4000 ft.
in length. The breadth varies from 17 ft. to 21 ft. of free
floor space, and their height is apparently about 30 ft. from
the floor to the centre of the roof. Each pillar or pier is

225. Central Corridor, Rame^varam. (From a Photograph.)

compound, 12 ft. in height, standing on a platform 5 ft. from
the floor, and richer and more elaborate in design than those
of the Parvati porch at Chidambaram (Woodcut No. 221), and
are certainly more modern in date.

The general appearance of those corridors may be gathered
from the annexed woodcut (No. 225), but no engraving, even
on a much more extended scale, can convey the impression
produced by such a display of labour when extended to an
uninterrupted length of 690 ft. None of our cathedrals are

1 There is a view of it in the Atlas of plates that accompanies Lord Valentia's
travels ; not very correct, but conveying a fair idea of its proportions.

DRAVIDIAN STYLE.

BOOK III.

more than 500 ft., and even the nave of St Peter's is only 600
ft. from the door to the apse. Here the side corridors are
almost 700 ft. long, and open into transverse galleries as rich
in detail as themselves. These, with the varied devices and
modes of lighting, produce an effect that is not equalled certainly
anywhere in India. The side corridors are generally free from
figure-sculpture, and consequently, from much of the vulgarity
of the age to which they belong, and, though narrower, produce
a more pleasing effect. The central corridor leading from the
inner enclosure is adorned on one side by portraits of the
Setupati rajas of Ramnad in the i/th century, and opposite
them, of their ministers. Even they, however, would be
tolerable, were it not that within the last few years they have
been painted with a vulgarity that is inconceivable on the part
of the descendants of those who built this fane. Not only they,
however, but the whole of the architecture has first been dosed
with repeated coats of whitewash, so as to take off all the
sharpness of detail, and then painted with blue, green, red, and
yellow washes, so as to disfigure and destroy its effect to an
extent that must be seen to be believed. Nothing can more
painfully prove the degradation to which the population is
reduced than this profanity. No upper class, and no refinement,
now remains, and the priesthood are sunk into a state of
debasement.

Assuming, however, for the nonce, that this painting never
had been perpetrated still the art displayed here would be very
inferior to that of such a temple as, for instance, Halebid in
Mysore, to be described further on. The perimeter, however, of
that temple is only 700 ft. ; here we have corridors extending
to 4000 ft, carved on both sides. It is the immensity of the
labour here displayed that impresses us, much more than its
quality, and that, combined with a certain picturesqueness and
mystery, does produce an effect which is not surpassed by any
other temple in India, and by very few elsewhere.

But for the wilful destruction of the inscriptions (less than
fifty years ago), we might have had the whole history of this
temple.1 The central shrines are built of a dark hard lime-

1 The Pandaram or manager of the
temple raised a suit against the Zamindar
of Ramnad to deprive him of the
hereditary right of patronage and super-
vision of the temple. It was conducted
on the Pandaram's part by one Apparu
Pillai, who destroyed the old inscriptions
and forged others, inserting them in the
walls, and then produced copies and
translations of them as evidence against

the claims of the Ramnad Setupatis.
The suit was appealed to the Privy
Council, but, on such evidence, was given
in favour of the forgers, and the
Zamindars were deprived of their right to
appoint the Dharmakartas or have any
share in the management of the temple
which their ancestors built and had so
richly endowed.

CHAP. IV. MADURA. 385

stone, differing from that employed in the rest of the building,
and are ascribed to a Setupati named Udaiyan, early in the
1 5th century.1 To him also is ascribed the west gopuram and
surrounding walls,2 but they probably belong to the time of
Sadayaka Tevar Udaiyan, early in the i/th century. The
first prakaram or enclosure, containing the older shrines,
measures about 190 ft. by 307 ft., and outside this — excepting
the old vimana — the style is so uniform and unaltered that its
erection could hardly have lasted beyond a hundred years ;
and if so, it must have been during the i/th century, when
the Ramnad rajas were at the height of their prosperity, and
when their ally or master, Tirumalai Nayyak, was erecting
buildings in the same identical style at Madura, that the
second prakaram, measuring 386 ft. by 314 ft, was completed
by Raghunath Tirumalai, about 1658, the south half of it
having been built by his predecessor. Vijaya Raghunath Tevar
(1709-1723), erected some buildings in the Amman temple,
with the mandapa in front, in which are statues of himself and
of his father Kadamba Tevar; and in 1740 Muttu Raghunatha
began the third prakaram, 690 ft. in length by 435 ft. from east
to west, which was completed by Muttu Ramalinga Tevar in 1769.
The central shrine may probably belong to the I5th or i6th
century, but all the enclosing courts had been erected within
170 years after the end of the i6th century, and endowed
exclusively by the family of the Setupati chiefs of Ramnad.
It is one of the last great works of Hindus, and the last addition
to it is the finest of all.

MADURA.

If the native authorities consulted by the late Professor
Wilson in compiling his Historical Sketch of the Kingdom of
Pandya could be relied upon, it would seem that the founda-
tion of the dynasty ought to be placed some centuries before
the Christian Era.3 Even, however, if this is disputed, the fact
of the southern part of the Peninsula being described as the
" Regio Pandionis " by classical authorities, is sufficient to prove
that a kingdom bearing that name did exist there in the early
centuries of the Christian Era. Their first capital, however,
seems to have been Korkai, near Cape Comorin.4 The story
of Kula^ekhara founding the dynasty, and the fabulous incidents

1 It is said he was aided in the work [ this work by a rich merchant and his

by a Singhalese king or chief named
Pararaja Sekhara, under whose super-
vision the stones were hewn and fitted
at Trinkonamalai.

2 He is said to have been assisted in

wife from Nagiir, whose statues surmount
the eastern wall.

3 ' Journal of the Royal Asiatic Society,'
vol. iii. p. 202.

4 Ante, p. 304.

VOL. I. 2 B

386 DRAVIDIAN STYLE. BOOK III.

with which the tale is adorned, is one of the favourite legends of
the south, and is abundantly illustrated in sculptures of Tirumalai
Nayyak's chaultri and in other buildings of the capital.

For our present purposes it is not worth while to attempt to
investigate the succession of the dates of the seventy-three kings
who are said to have succeeded one another before the accession
of the Nayyak dynasty, in 1559, inasmuch as no building is now
known to exist in the kingdom that can claim, even on the most
shadowy grounds, to have been erected by any of these kings.
It may have been that, anterior to the rise of the great Chola
dynasty, in the loth and I ith century, that of Madura may have
had a long period of prosperity and power ; but whatever they
did build has been destroyed or so altered that its existence
cannot now be identified. After that, for a while they seem
to have been subjected to the Ballala dynasty of Mysore, and
the same Muhammadan invasion that destroyed that power in
1310 spread its baneful influence as far as Ramnad, and for two
centuries their raids and oppressions kept the whole of southern
India in a state of anarchy and confusion. Their power for
evil was first checked by the rise of the great Hindu state of
Vijayanagar, on the Tungabhadra, in the I4th century, and by
the establishment, under its protection, of the Nayyak dynasty
by VLrwanath Nayyak, in the i6th. After lasting 177 years,
Mtnakshi, the last sovereign of the race — a queen — was first
aided, and then betrayed, by Chanda Sahib the Nawab of the
Karnatik, who plays so important a part in our wars with the
French in these parts.

It may be — indeed, probably is the case — that there are
temples in the provinces that were erected before the rise of
the Nayyak dynasty, but all those in the capital, with the great
temple at 5rirangam, described above, were erected or extended
during the two centuries of their supremacy, and of those in the
capital nine-tenths at least were erected during the long and
prosperous reign of the seventh king of this dynasty, Tirumalai
Nayyak, or as he is more popularly known, Trimal Nayyak, who
reigned from 1623 to f6$g.1

Of his buildings, the most important, for our purposes at
least, is the celebrated Vasanta or Pudu Mantapam,2 known as
' Tirumalai's chaultri,' which he built for the reception of the

1 'Journal of the Royal Asiatic Society,' the ' Archseologia,' vol. x. p. 457; and

vol. iii. pp. 230 et seqq. \ by Wilson, ' Journal of the Royal Asiatic

a Fortunately this chaultri is also one Society,' vol. iii. p. 232. Volumes of

of the best known of Indian buildings. I native drawings exist in some collections

It was drawn by Daniell in the end of < containing representations of every pillar,

the 1 8th century, and his drawings have A model in bronze of a porch exists at

been repeated by Langles and others. I South Kensington Museum, and it has

It was described by Mr. Blackadder in been abundantly photographed.

CHAP. IV.

MADURA.

387

presiding deity of the place, who consented to leave his dark
cell in the temple and pay the king an annual visit of ten
days' duration in the hot month of May, on condition of his
building a hall worthy of his dignity, and where he could
receive in a suitable manner the homage of the king and his
subjects. As will be seen from the plan (Woodcut No. 226)

?5 '

& & . G v?
- .- jg V- .

P~^

3

•

*^*

C

»

& • • c-

C

•

* 0

: «s

»

& «

; e

I

* :

•3

*

«. e

i *

&

*

•

*
»

3> C
& «

*

•i

» I

' I

&

e. e

«

•

if 0.

&

i. <&.

•CJ

8"

& o

>»

•

» "S

*

»

TT & 0 *

*

'^ S

* Q 68:

fe

•

226.

Plan of Tirunialai Nayyak's
Chaultri. (From a Draw-
ing in the possession of
the Royal Asiatic Society. )
Scale loo ft. to i in.

227.

Pillar in Tirumalai Nayyak's Chaultri.

(From a Drawing in the possession of the Royal

Asiatic Society. )

the hall is 333 ft. long by 105 ft. in width, measured on the
stylobate, and consists of four ranges of columns, all of which
are different, and all most elaborately sculptured. An elevation
of one is given (Woodcut No. 227), but is not so rich as those
of the centre, which have life-sized figures attached to them,
and are even more elaborate in their details. In this instance
it will be observed that the detached bracketing shaft at

388 DRAVIDIAN STYLE. BOOK III.

Chidambaram has become attached to the square central pier,
and instead of the light elegance that characterised that
example, has become a solid pier, 5 or 6 ft. in depth —
richer certainly, but far from being either so elegant or so
appropriate as the earlier example.

The view of the interior (Woodcut No. 228) gives some, but
only a faint, idea of the effect. The sides are now closed with
screens ; but in effect, as in detail, it is identical with the
corridors at Rame^varam, where the light is abundant.

As the date of this hall is known — it took twenty -two
years to erect it, 1623 to 1645 — it becomes a fixed point in
our chronology of the style.1 We can, for instance, assert
with perfect certainty that the porch to Parvati's shrine at
Chidambaram (Woodcut No. 221) is certainly anterior to this,
probably by a couple of centuries, and, with equal certainty
that the corridors at Ramejvaram are contemporary.

From the history of the period we learn that the rajas of
Ramnad were at times independent, at others at war with the
Nayyaks ; but in Tirumalai Nayyak's time they were either his
allies or dependants ; and the style and design of the two
buildings are so absolutely identical that they must belong to
the same age. If the king of Madura had indeed been allowed
any share in making the original design, that temple would
probably have been a nobler building than it is ; for, though the
details are the same, his three aisled hall leading to the
sanctuary would have been a far grander feature architecturally
than the single - aisled corridors that lead nowhere. The
expense of one of the single-aisled corridors at Rame^varam,
almost 700 ft. long, would have been about the same as the
triple - aisled chaultri at Madura, which is half their length.
Consequently the temple must have cost between three and
four times as much as the chaultri ; and the actual cost must
have been immense when we consider the amount of labour
expended on it, and that the material in both is the hardest
granite.

The facade of this hall, like that of almost all the great halls
in the south of India, is adorned either with Vyalis — monsters
of the lion type trampling on an elephant — or, even more
generally, by a group consisting of a warrior sitting on a rearing
horse, whose feet are supported on the shields of foot soldiers,
sometimes slaying men, sometimes tigers. These groups are

1 According to Wilson the mantapam Oriental Historical Manuscripts states

was begun in the second year of I the cost of it at a lakh of Pons or

Tirumalai's reign, and completed in ^"20,000, and that it was finished in

twenty-two years, at a cost of upwards seven years, 1626-1633.
of a million sterling. But one of the i

CHAP. IV.

MADURA.

389

390

DRAVIDIAN STYLE.

BOOK III.

found literally in hundreds in southern India, and, as works
exhibiting difficulties overcome by patient labour, they are un-
rivalled, so far as I know, by anything found elsewhere. As
works of art, they are the most barbarous, it may be said the most
vulgar, to be found in India, and do more to shake one's faith
in the civilisation of the people who produced them than any-
thing they did in any other department of art. Where these
monstrosities are not introduced, the pillars of entrances are
only enriched a little more than those of the interior, when the
ornamentation is in better taste, and generally quite sufficiently
rich for its purpose.

Immediately in front of his chaultri, Tirumalai Nayyak com-
menced his Raya gopuram, which, had he lived to complete
it, would probably have been the finest edifice of its class in
southern India. It measures 174 ft. from north to south, and
117 feet in depth. The entrance through it is 21 ft. 9 in. wide ;
and if it be true that its gateposts are 57 ft. in height, that would
have been the height of the opening.1 It will thus be seen that
it was designed on even a larger scale than that at 5rirangam,
described above, and it certainly far surpasses that celebrated
edifice in the beauty of its details. Its doorposts alone are
single blocks of granite, carved with the most exquisite scroll
patterns of elaborate foliage, and all the other carvings are
equally beautiful. Being unfinished, and consequently never
consecrated, it has escaped whitewash, and alone, of all the
buildings of Madura, its beauties can still be admired in their
original perfection.

The great temple at Madura is a larger and far more
important building than the Vasanta mantapam or Chawadi.
It possesses all the characteristics of a first-class Dravidian
temple, and, as its date is well known, it forms a landmark of
the utmost value in enabling us to fix the relative date of
other temples (Woodcut No. 229).

The sanctuary (A) is dedicated to .Siva, under the name of
Sundarejvara, and this clearly points to its having been founded
by one of the Pandya princes who bore the name of Sundara, of
whom there were four during the thirteenth century. The
shrine of the goddess (B) is consecrated to Minakshi.2

The city and temple fell to the Muhammadans in 1324, who
plundered it mercilessly, and it was only recovered by the
Vijayanagar sovereign in 1372 when Hindu worship was

1 Most of these particulars, regarding
the temples, are taken from Capt. Lyon's
description of his photographs of the
places. He devoted twenty-four photos
to this temple, of which the negatives
are now in the India Office.

2 Minakshi means "fish-eyed"; but,
with the first vowel short, it would mean
" bright-eyed." Traditionally she was the
wife of Sundara, but as a goddess is the
same as Parvati, the goddess wife of
.Siva.

CHAP. IV.

!

229. Plan of Madura Temple. Scale 200 ft. to i in.1

1 Key to the plan:— A Temple of Sundarejvar ; AA Pudu mantapam ; B Shrine of

59a DRAVIDIAN STYLE. BOOK III.

restored. The temple itself certainly owes all its magnificence
to Tirumalai Nayyak, A.D. 1623-1659, or to his elder brother,
Muttu Virappa, who preceded him, and who built the Minakshi
Nayyak mantapam, said to be the oldest thing now existing,
outside the shrines. The Kalyana mantapam is said to have
been built in A.D. 1707, and the Tatta Suddhi in 1770. These,
however, are insignificant parts compared with those which
owe their origin to Tirumalai Nayyak.

The temple itself is a nearly regular rectangle, two of its
sides measuring 720 ft. and 729 ft, the other two 834 ft. and
852 ft. It possessed four gopurams of the first class (I. II. III.
IV.), and six smaller ones (VI. to XI.) ; a very beautiful tank (L),
surrounded by arcades ; and a hall of 1000 columns (Q), whose
sculptures surpass those of any other hall of its class I am
acquainted with. There is a small shrine (J), dedicated to the
god Sabhapati, which occupies the space of fifteen columns, so
the real number is only 985 ; but it is not their number but
their marvellous elaboration that makes it the wonder of the
place, and renders it, in some respects, more remarkable than
the chaultri about which so much has been said and written.
I do not feel sure that this hall alone is not a greater work
than the chaultri ; taken in conjunction with the other buildings
of the temple, it certainly forms a far more imposing group.

TlNNEVELLY.

Though neither among the largest nor the most splendid
temples of southern India, that at Tinnevelly will serve to give
a good general idea of the arrangement of these edifices, and
has the advantage of having been built on one plan, and at one
time, without subsequent alteration or change. Like the little
cell in the Tiruvalur temple (Woodcut No. 216), it has the
singularity of being a double temple, the great square being
divided into equal portions, of which the north one is dedicated
to the god Siva, the south half to his consort Parvati. The
following plan (Woodcut No. 230) represents one of the halves,

Minakshidevi ; C Small shrine of Ganeia,
and D of Subrahmanya ; E Vedi or Altar ;
F Nandi Pavilion ; G Javandi^vara man-
tapam ; H Navagraha or nine planets ;
I Large Gane^a ; JJ Shrines of Natejvar ;
K Poet's College; L Tank of Golden
Lillies ; M Mudali Pillai mantapam ; O
Ashta Lakshmi hall ; P 16 pillar man-
tapam ; Q Thousand pillar mantapam ;
R Viravasantaraya mantapam ; S Katyana

tapam ; U Lingam ; V Tiruvachi Gopu-
ram ; W Chitra mantapam ; Y Ellam-
vatta-siddha; Z Madura Nayyak temple;
I, II, III, IV, Four outer Gopurams;
VI, VIII, IX, Three Gopurams of
the second Prakaram; X, XI, Gates
to the Minakshi prakaram ; VII, Gate
between the temples ; XII, Ashta-Sakti
mantapam ; The plan is reduced from
one in Francis's ' Madura Gazetteer,'

Sundara mantapam ; T Servaikaran man- i vol. i.

CHAP. IV.

TINNEVELLY.

393

TTTTT?TT;rTtTT f TT'TTT

230. Half-plan of Temple at Tinnevelly. (From a Plan in the possession of the Royal
Asiatic Society. ) Scale 100 ft. to i in.

394 DRAVIDIAN STYLE. BOOK III.

which, though differing in arrangement from the other, is still
so like it as to make the representation and description of one
sufficient for both.

The general dimensions of the whole enclosure are 580 ft.
by 756 ft., the larger dimension being divided into two equal
portions of 378 ft. each. There are three gateways to each
half, and one in the wall dividing the two ; the principal gate-
way faces the east entrance to the temple, and the lateral ones
are opposite each other. An outer portico precedes the great
gateway, leading internally to a very splendid porch, which,
before reaching the gateway of the inner enclosure, branches
off on the right to the intermediate gateway, and on the left
to the great hall of 1000 columns — 63 ft. in width by about
520 ft. in depth.

The inner enclosure is not concentric with the outer, and, as
usual, has only one gateway. The temple itself consists of a
cubical cell, surmounted by a vimana or spire, preceded by two
mandapas, and surrounded by triple colonnades. In other
parts of the enclosure are smaller temples, tanks of water,
gardens, colonnades, etc., but neither so numerous nor so
various as are generally found in Indian temples of this class.
The inscriptions go back to the first half of the I3th century.

The great looo-pillared portico in the temple is one of the
least poetic of its class in India. It consists of a regiment of
pillars 10 deep and extending to 100 in length, without any
break or any open space or arrangement. Such a forest of
pillars does, no doubt, produce a certain effect ; but half that
number, if arranged as in some of the Chalukyan or Jaina
temples, would produce a far nobler impression. The aim of
the Dravidians seems to have been to force admiration by the
mere exhibition of inordinate patient toil.

KUMBAKONAM.

If the traditions of the natives could be trusted, Kumbakonam
— one of the old capitals of the Chola dynasty — is one of the
places where we might hope to find something very ancient.
There are fragments of older temples, indeed, to be found
everywhere, but none in situ. All the older buildings seem to
have been at some time ruined and rebuilt, probably on the
same site, but with that total disregard to antiquity which is
characteristic of the Hindus in all ages as it is of our modern
"restorers." One portico, in a temple dedicated to Sri- Rama,
is very like that leading from the second to the third gopuram
in the temple of Jumbuke.rwara, described above, but, if any-
thing, it is slightly more modern. There is also one fine

CHAP. IV.

KUMBAKONAM

395

231. Gopuram at Kumbakonam. (From a Photograph. )

396 DRAVIDIAN STYLE. BOOK III.

gopuram in the town, represented in the woodcut (No. 231).
It is small, however, in comparison with those we have just
been describing, being only 84 ft. across and about 130 ft. in
height. Those of 5rirangam and Madura have, or were
intended to have, at least double these dimensions.

It is, however, a richly-ornamented example of its class, and
the preceding woodcut conveys a fair impression of the effect of
these buildings generally. It is not old enough to be quite of
the best age, but it is still not so modern as to have lost all
the character and expression of the earlier examples.

VELLOR AND PERUR.

Although the temples at Vellor and at Perur, near Koimbatur,
can only rank among the second class as regards size, they
possess porticos of extreme interest to architectural history,
and are consequently worthy of more attention than has been
bestowed upon them. That at Vellor, however, is unfortunately
situated in the fort occupied by the British, and has consequently
been utilised as a store. Walls have been built between its
piers, and whitewash and fittings have reduced it to that con-
dition which we think appropriate for the noblest works of
art in India. Enough, however, still remains to enable us to
see that the Kalyana mantapam here, though not one of the
most elegant, is one of the oldest porches in the south. As
will be seen from the woodcut (No. 232), the Vyalis and rearing
horsemen are clearly and sharply cut, and far from being so
extravagant as they sometimes are. The great cornice, too,
with its double flexures and its little trellis-work of supports,
is not only very elegant in form, but one of those marvels of
patient industry, such as are to be found hardly anywhere
else. There are many such cornices, however, in the south :
one at Avadaiyar-kovil, in Tanjor district, is deeper and more
elaborate than even this one. The outer facing there is said
to be only about an inch in thickness, and its network of
supports is more elaborate and more delicate than those at
Vellor, though it is difficult to understand how either was ever
executed in so hard a material. The traditions of the place
assign the erection of the Vellor mantapam variously to the
year 1350, and according to other accounts to about 1485.
The bracket shafts (Woodcut No. 233) are similar but even
more elegant than those in Parvati's porch at Chidambaram ;
but they are — some of them at least — attached to the pier
by very elegant open-work, such as is found in Pratapa Rudra's
temple at Worangal (Woodcut No. 252), or in the windows at
Halebid. As both these examples are earlier than 1300, it

CHAP. IV.

VELLOR AND PERUR.

397

might seem that this one was so also, but it is difficult to
feel certain when comparing buildings so distant in locality,
and belonging to different styles of art.

ortico of Temple at Vellor.

The date of the porch at Perur (3 miles south-west of
Koimbatur) is ascertained within narrow limits by the figure
of a sepoy loading a musket being carved on the base of one

398 DRAVIDIAN STYLE. BOOK III.

of its pillars, and his costume and the shape of his arm are
exactly those we find in contemporary pictures of the wars
of Aurangzib, or the early Marathas in the beginning of the
1 8th century. As shown in Woodcut No. 234, the bracket
shafts are there attached to the piers as in Tirumalai Nayyak's
buildings, and though the general character of the architecture
is the same, there is a coarseness in the details, and a marked
inferiority in the figure-sculpture, that betrays the distance
of date between these two examples.

Slight as the difference may appear to the unpractised
eye, it is within the four centuries that include the dates of
these two buildings (1350 to 1750) that practically the whole
history of the later Dravidian temple architecture is included.
For it is safe to assert that nine-tenths, at least, or more, of
those which are now found south of the Tungabhadra, were
erected or largely extended and rebuilt between these dates.

The earlier works of the seven centuries that elapsed
between the carving of the rocks at Mamallapuram and the
erection of the Vellor pagoda have almost totally perished.
But during that period, a style was elaborated and so fixed
that it should endure for five centuries afterwards, with so
little change, and with only that degradation in detail, which
is the fatal characteristic of art in India.

It seems impossible that the horsemen, the Vyalis, and
above all, the great cornice of double curvature, shown in
the woodcut (No. 232), could have been brought to these
fixed forms without long experience, and the difficulty is to
understand how they could ever have been elaborated in stone
at all, as they are so unlike lithic forms found anywhere else ;
yet they are not wooden, nor is there any trace in them of
any of their details being derived from wooden architecture,
as is so evidently the case with the Buddhist architecture of
the north. One suggestion that occurs to me is that they
might be derived from terra-cotta forms. Frequently, at the
present day, figures of men on horseback larger than life, or
of giants on foot, are seen near the village temples made of
pottery, their hollow forms of burnt clay, and so burnt as to
form a perfect terra - cotta substance. Most of the figures
also on the gopurams are not in plaster as is generally said,
but are also formed of clay burnt. The art has certainly
been long practised in the south, and if we adopt the theory
that it was used for many ornamental purposes along with wood
or stone, it will account for much that is otherwise unintelligible
in the arts of the south. But we may further suppose that
the broad sloping slabs of the earlier temples having no level
bed to rest on the wall head, and being apt to slide down,

CHAP. IV.

VELLOR AND PERUR.

399

233. Compound Pillar at Vellor.
(From a Photograph.)

234. Compound Pillar at PSrur.
(From a Photograph.)

400 DRAVIDIAN STYLE. BOOK III.

the curved form was devised to secure a flat rest on the wall
and at the same time that the wall or roof above might
have a flat plate on which to rest : and besides the outward
curve afforded a better opportunity for supporting it upon
brackets.

The loss of the earlier structures that would have made
all this clear is largely due to the Muhammadan raids in the
beginning of the I4th century, when these iconoclasts spared
no shrine they found on their marches. What escaped them
was either " restored " beyond recognition by zealous Hindu
princes and temple builders, or is now in the innermost
enclosures, inaccessible to any European capable of judging
of their style and age. The few old shrines at Conjivaram,
noticed above, and some scattered and neglected ruins do
remain, and are very instructive ; the desideratum is that
adequate surveys of them are so slow in being published.

VlJAYANAGAR.

The dates above quoted will no doubt sound strange and
prosaic to those who are accustomed to listen to the childish
exaggerations of the Brahmans in speaking of the age of their
temples. There is, however, luckily a test besides the evidence
above quoted, which, if it could be perfectly applied, would
settle the question at once.

When in the beginning of the I4th century the Muham-
madans from Delhi first made their power seriously felt in the
south, they struck down the kingdom of the Hoysala Ballalas
in 1310, and destroyed their capital of Halebid ; and in 1322
Orangal or Worangal, which had been previously attacked,
was finally destroyed, and it is said they then carried their
victorious arms as far as Ramnad. The Muhammadans did not,
however, at that time make any permanent settlement in the
south, and the consequence was, that as soon as the Hindus
were able to recover from the panic, Bukka and Harihara
princes it is said of the deposed house of Orangal, gathered
around them the remnants of the destroyed states, and founded
a new state in the town of Vijayanagar on the Tungabhadra.
An earlier city it is said had been founded there about the
beginning of the I2th century, but only as a dependency of
the Mysore Raj, and there is consequently no reason for
supposing that any of the buildings in the city (unless it be
some of the small Jaina temples), belong to that period, nor
indeed till the new dynasty founded by Bukka had consolidated
its power, which was certainly not before the middle of the
1 4th century.

CHAP. IV. VIJAYANAGAR. 401

The city was finally destroyed by the Muhammadans in
1565, but during the two previous centuries it maintained a
gallant struggle against the Bahmani and Adil Shahi dynasties
of Kulbarga and Bijapur, and was in fact the barrier that
prevented the Moslims from taking possession of the whole
country as far as Cape Comorin.

Its time of greatest prosperity was between the accession of
Krishna Deva, 1508, and the death of Achyuta Raya, 1542, and
it is to their reigns that the finest monuments in the city must
be ascribed. There is, perhaps, no other city in all India in
which ruins exist in such profusion or in such variety as in
Vijayanagar, and as they are all certainly comprised within
the century and a half, or — at the utmost — the two centuries
that preceded the destruction of the city, their analogies afford
us dates that hardly admit of dispute.

Among those in the city the most remarkable is that
dedicated to Vithoba, or Vitthalaswamin, a local manifestation
of Vishnu. It was apparently begun by Krishna Deva, at
least as early as A.D. 1513, and continued by Achyuta Raya,
1529-1542, and never was finished ; and as inscriptions in it are
dated in 1561 and 1564, recording grants of Sada^iva, we might
fairly infer that the works were stopped by the fall of the
kingdom in 1565. The temple stands in a rectangular enclosure
538 ft. by 310, with gopurams on the east, north, and south
sides. Outside the east entrance stands a lofty Dipdan pillar,
and there are two pavilions of architectural merit in the court,
besides other buildings. Round nearly the whole court ran a
deep verandah or corridor with three rows of piers ; but most
of it is now ruined. The principal part of the temple consists
of a porch or mantapam, represented in the woodcut, No. 236,
page 403. It is wholly in granite, and carved with a boldness
and expression of power nowhere surpassed in the buildings of
its class. As will be observed, it has all the characteristic
peculiarities of the Dravidian style : the bold cornice of double
flexure, the detached shafts, the Vyalis, the richly - carved
stylobate, etc. But what interests us most here is that it forms
an exact half-way house in style between such porches as those
at Vellor and Chidambaram, and that of Tirumalai Nayyak at
Madura. The bracket shafts are detached here, it is true, but
they are mere ornaments, and have lost their meaning. The
cornice is as bold as any, but has lost its characteristic supports,
and other changes have been made which would inevitably
have led in a short time to the new style of the Nayyak dynasty.

The little building on the right is the car of the god, the
base and principal storey being formed of a single block of
granite, with movable wheels, but they are the only parts that
VOL. I. 2 C

402

DRAVIDIAN STYLE.

BOOK III.

235. Plan of Vitthalaswamin Temple at Vijayanagar. (From a plan by Mr A. Rea.)
Scale 100 ft. to i in.

PLATE XI.

STONE CAR, AT THE TEMPLE OF VITTHALA, VIJAYAXAGAR, l88l
(NOW DESTROYED).

[To face page 403, Vol. I.

CHAP. IV.

VIJAYANAGAR.

403

move.1 It is shown in Plate XI. as it appeared before
its brick tower was taken down. There are, besides, either

236. View of Porch of Temple of Vitthalasvamin at Vijayanagar. (From a Photograph

by Mr Neill. )

one or two pavilions, smaller, but similar in design to that
represented in the woodcut, a gopuram, and other adjuncts,
which would be interesting, if we had the means of comparing
and describing them.2

Although the temple of Vitthala is certainly one of the
most remarkable ruins in India, and there are other temples of
great beauty and extent in the capital, it is not quite clear that
it is there the chefs-d'oeuvre of this dynasty are to be found, but
rather at a place called Tadpatri about one hundred miles a
little south of east from the capital. There are two temples
there : the one now in use — the Chintala Tiruvenkataswamin
temple — dedicated to Vishnu, is slightly the later of the two,
having been built in the time of Timma Nayudu, the local
governor, and son (or grandson) of Ramalinga Nayudu, governor
in the early part of Krishna Deva's reign, who began the
Ramcj-vara temple here.3 The wonders of the place are two

1 The upper portion was of brick, but
about twenty years ago, after an elaborate
repair under the direction of Major H.
Cole, the collector had it pulled down,
lest it should crush the base, which had
been cracked by fire. Similar stone cars
exist at Tadpatri and other temples in
southern India.

2 As Dr G. le Bon remarks, Vijayanagar
is well worthy of a complete monograph
on its architecture, as the culmination of
the style. ' Les Monuments de 1'Inde,'
pp. 161-162.

3 Inscriptions belonging to this Rame.r-
vara temple are dated from 1507 to

404

DRAVIDIAN STYLE.

BOOK TIL

gopurams belonging to the second, which is now a deserted
temple on the banks of the river, about a quarter of a mile
from the other. One of these was apparently quite finished,
the other on the north never carried higher than the perpen-
dicular part. In almost all the gopurams of India this part is
comparatively plain, all the figure-sculpture and ornament
being reserved for the upper or pyramidal part. In this
instance, however, the whole of the perpendicular part is
covered with the most elaborate sculpture, cut with exquisite
sharpness and precision, in a fine close-grained hornblende (?)
stone, and produces an effect richer, and on the whole perhaps
in better taste, than anything else in this style (Woodcuts
Nos. 237, 238). It is difficult, of course, to institute a comparison
between these gopurams and such works as Tirumalai Nayyak's
chaultri, or the corridors at Ramejvaram ; they are so different
that there is no common basis of comparison but the vulgar
one of cost ; but if compared with Halebid or Belur, these
Tadpatri gopurams stand that test better than any other works
of the Vijayanagar Rajas. They are inferior, but not so much
so as one would expect from the two centuries of decadence
that elapsed between them, and they certainly show a marked
superiority over the great unfinished gopuram of Tirumalai
Nayyak, which was commenced, as nearly as may be, one
century afterwards.

About fifty miles still further east, at a place called Diguva
Ahobalam, in Karnal district, there is a large unfinished
mantapam in plan and design very like that of the temple
of Vithoba at Vijayanagar, but its style and details are
much more like those of the Nayyaks, though local tradition
assigns it to Pratapa Rudra about 1300. Traditions, however,
usually refer to the original shrine, and if we are guided by
style, it could hardly have been erected before the destruction of
that capital in A.D. 1565. The dynasty, however, continued
to exist for one or two centuries after that time, till the country
was finally conquered by Tipu Sultan. The inscriptions have
not yet been examined, but seem mostly to belong to the
latter part of the i6th century.1 Whoever may have built it,
it is a fine bold specimen of architecture, and if the history
of the art in the south of India is ever seriously taken up, it
will worthily take a place in the series as one of the best
specimens of its age, wanting the delicacy and elegance of
the earlier examples, but full of character and merit.2

1 Among the Mackenzie MSS. at
Madras there are copies of the inscrip-
tions and other notices of the Ahobalam
temples.

8 For long the temple of Vishnu on the
hill of Tripetty or Tirupati, in North
Arkat district, was reputed to be the
richest, the most magnificent, as it was

CHAP. IV.

TADPATRI.

405

237. Entrance through Gopuram at Tadpatri, (From a Photograph.)

406

DRAVIDIAN STYLE.

BOOK III.

Portion of Gopuram at Tadpatri.

(From a Photograph.)

CHAP. IV.

CONCLUSION.

407

CONCLUSION.

The buildings mentioned, and more or less perfectly
described, in the preceding pages are in number rather less
than one-third of the great Dravidian temples known to exist
in the province. In importance and extent they certainly
are, however, more than one - half. Of the remainder, none
have vimanas like that of Tanjor, nor corridors like those of
Ramej-varam ; but several have gopurams quite equal to or
exceeding those mentioned above, and many have mantapams of
great beauty and extent. Several — such as Avadaiyar-Kdvil,
Virinjipuram near Vellor, Taramangalam in Salem district,
Kurugodu in Bellari, and others — possess features unsurpassed
by any in the south, especially the first named, which may,
perhaps, be considered as one of the most elegant of its class,
as well as one of the oldest. It would, however, be only tedious
to attempt to describe them without plans to refer to, or more
extensive illustrations than are compatible with a work of this
class. They are, however, worthy of more attention than has
yet been paid to them, and of more complete illustration than
has hitherto been bestowed upon them. Taken altogether,
they certainly do form as extensive, and in some respects
as remarkable, a group of buildings as are to be found in
provinces of similar extent in any part of the world — Egypt,
perhaps alone excepted ; but they equal even the Egyptian in
extent, and though at first sight so different, in some respects
present similarities which are startling. Without attempting to
enumerate the whole, it may be mentioned that the gopurams
both in form and purpose, resemble the pylons of the Egyptian
temples. The courts with pillars and cloisters are common to
both, and very similar in arrangement and extent. The great
mantapams and halls of 1000 columns reproduce the hypostyle
halls, both in purpose and effect, with almost minute accuracy.

certainly the most sacred of all those in
the Presidency. So sacred, indeed, was
it declared, that no unbelieving foreigner
had ever been allowed to climb the holy
hill (2, 500 ft. high) or profane its sacred
precincts. In 1870, a party of police
forced their way in, in pursuit of a
murderer who had taken refuge there,
and Mr J. D. B. Gribble, who accom-
panied them, published an account of
what they saw in the ' Calcutta Review '
in 1875 (vol. Ixi. pp. 142-156). As
he exclaims, "Another of the illusions
of my youth destroyed." The temple
is neither remarkable for its size nor its

magnificence. In these respects it is
inferior to Conjivaram, .Srirangam, and
many others ; and whatever may be done
with its immense revenues, they certainly
are not applied to its adornment. It is a
fair specimen of a Dravidian temple of the
second class, but in a sad state of dilapida-
tion and disrepair. It was originally a
5aiva temple, but was converted to the
worship of Vishnu, by Ramanujacharya,
in the I2th century. For views of the
Tirupati temples in the village below, see
Dr G. Le Bon's ' Les Monuments de
1'Inde,' figs. 199-204.

408

DRAVIDIAN STYLE.

BOOK III.

The absence of any central tower or vimana over the sanctuary
is universal in Egypt, and only conspicuously violated in one
instance in India. Their mode of aggregation, and the amount
of labour bestowed upon them for labour's sake, is only too
characteristic of both styles. There are, besides, many simi-
larities that will occur to any one familiar with both styles.

Is all this accidental ? It seems strange that so many
coincidences should be fortuitous, but, so far as history affords
us any information, or as any direct communication can be
traced, we must for the present answer that it is so. The
interval of time is so great, and the mode in which we fancy
we can trace the native growth of most of the features in India
seem to negative the idea of an importation ; but there certainly
was intercourse between Egypt and India in remote ages, and
seed may then have been sown which possibly had fructified
long afterwards.

A digression may be made in conclusion with reference to
the famous monastery referred to (p. 171), as spoken of from
hearsay, both by Fah Hian and Hiuen Tsiang. Its situation has
long been a puzzle. The second pilgrim says it was built by
one of the Andhra kings as a monastery for Nagarjuna.1 It
had lofty halls in five tiers, each with four courts and temples
containing golden images of Buddha. But after a time the
Brahmans had ousted the Buddhists and, he adds, "the way
of access to it was no longer known." The Tibetan works state
that Nagarjuna died at the great monastery of aTal-gyi-ri, a
translation of .Sri-Sailam or 5ri-Parvata, both names of a very
old place of pilgrimage on a rocky hill overlooking a gorge of
the Krishna river, and which is one of the twelve Jyotir-lingas
or great Saiva shrines of the Hindus. The place is difficult of
access, but was visited by Colonel Colin Mackenzie in I794,2
and perhaps by four or five Europeans before 1886, when the
editor made a hurried excursion to it. Mackenzie had mentioned
the animals carved on the surrounding walls in a way that
seemed to follow the arrangement described by Fah - Hian.
Though beyond the limits of the Kosala kingdom, with which
Hiuen Tsiang seemed to connect it, it was most probably
within the early Andhra dominions.

The reports made to the pilgrims were evidently exaggerated
or vague and misunderstood ; and if vSYi-^Sailam were the site of

1 The Chinese syllabus by which Hiuen
Tsiar.g represents its name may be trans-
literated as Bhr&mara-giri — "black bee
mountain." He says it meant "black
peak," which is equivalent to "Nalla-

mallai," the name of the hills on the
south of the Krishna river, to the west and
south of .Sri-iSailam in Karnal district.

2 ' Asiatic Researches,' vol. v. pp. 303'
314.

CHAP. IV.

CONCLUSION.

409

this wonderful monastery, there are no traces of Buddhism there
now. The earliest structures were possibly of wood, and the
place has been often destroyed. It was near Chandraguptapur,
the capital of an early kingdom of Sri-Sailam, occasionally
mentioned in inscriptions.

The present temples stand within a rectangular enclosure
about 630 ft. from north to south by 510 ft. from east to west,

w

mpq

]E°ra

H nff^

:n:L

Mndl^^

: ^5

i — i

!*

I — I

239-

Plan of the Temples at Sri-Sailam,
Scale 150 ft. to i in.

with a slight extension on the west side and gopurams on the
other three, all in the style of the i6th and I7th centuries, having
the lower storeys in stone and the upper in brick — though stone
is more easily procurable in such a locality. The temples are
now dedicated to Mallikarjuna, a form of Siva, and to Bhra-
marambha or Madhava and Parvati. The sculptures on the
walls of the court are irregularly disposed in blocks in the upper
four courses on the outer face. They are almost a counterpart of
those on the walls of the Hazara-Rama temple at Vijayanagar,
and belong to the same period. Indeed, one is led to suppose

410

DRAVIDIAN STYLE.

BOOK III.

that they were carved for the construction of a shrine rather
than for the enclosure wall, into the upper courses of which they
are now built in an irregular way.

From the plan (Woodcut No. 239), it may be inferred that
when this wall was erected the court was considerably enlarged
to the north and west. The temples inside are all mean and
insignificant, and, though older than the outer enclosure, they are
not of earlier date than the i6th century, and probably represent
the shrines hurriedly restored after some raid of iconoclastic
Muhammadans.1 The court is partly occupied by rows of
corridors for the accommodation of pilgrims ; but most of the
buildings are in a state of decay and ruin.

1 So late as 1855 a band of Rohillas
crossed from the Haidarabad territory
and robbed the temple of jewels valued

at Rs. 20,000, dug up the floors of the
shrines, and destroyed the ancient
images.

CHAP. V. CIVIL ARCHITECTURE. 411

CHAPTER V.

CIVIL ARCHITECTURE.

CONTENTS.

Palaces at Madura and Tanjor — Garden Pavilion at Vijayanagar — Palace

at Chandragiri.

ALTHOUGH the Dravidians were extensive and enthusiastic
builders, it is somewhat singular that till they came in contact
with the Muhammadans all their efforts in this direction should
have been devoted to the service of religion. No trace of any
civil or municipal building is to be found anywhere, though from
the stage of civilisation that they had attained it might be
expected that such must have existed. What is, however, even
more remarkable is, that kingdoms always at war with one
another, and contending for supremacy within a limited area,
might have been expected to develop some sort of military
architecture. So far, however, as is now known, no castle or
fortification of any sort dates from the Pandya, Chera, or Chola
days. What is still more singular is that they have no tombs.
They seem always to have burnt their dead, and never to have
collected their ashes or raised any mounds or memorials to their
departed friends or great men. There are, it is true, numberless
"Rude stone monuments" all over the south of India, but, till
they are more thoroughly investigated, it is impossible to say
whether they belong to the Dravidians when in a lower stage of
civilisation than when they became temple builders, or whether
they belong to other underlying races who still exist, in scattered
fragments, all over the south of India.1 Whoever these Dolmens
or stone circles may have belonged to, we know, at least, that
they never were developed into architectural objects, such as
would bring them within the scope of this work. No Dravidian
tomb or cenotaph is known to exist anywhere.

When, however, the Dravidians came in contact with the
Musalmans this state of affairs was entirely altered, in so far,
at least, as civil buildings were concerned. The palaces, the
kacherts, the elephant - stables, and the dependencies of the

1 What I know on this subject I have already said in my work on ' Rude Stone
Monuments,' pp. 455 et seqq. Conf. ' Archaeological Survey of Western India,' vol. i.
p. 8, and plate ; ' Indian Antiquary,' vol, iii. pp. 53-54, and 306-308 ; vol. v. pp,
I59> 255; and vol. vi. p. 230.

4i2 DRAVIDIAN STYLE. BOOK III.

abodes of the rajas at Vijayanagar and Madura, rival in extent
and in splendour the temples themselves, and are not surpassed
in magnificence by the Muhammadan palaces of Bijapur or Bidar.

One of the most interesting peculiarities of these civil
buildings is, that they are all in a new and different style of
architecture from that employed in the temples, and the distinc-
tion between the civil and religious art is kept up to the present
day. The civil buildings are all in what we would call a
pointed-arched Moorish style — picturesque in effect, if not
always in the best taste, and using the arch everywhere and
for every purpose. In the temples the arch is never used as
an architectural feature. In some places, in modern times, when
they wanted a larger internal space than could be obtained by
bracketing without great expense, a brick vault was introduced
— it may be said surreptitiously — for it is always concealed.
Even now, in building gopurams, they employ wooden beams,
supported by pillars, as lintels, to cover the central openings
in the upper pyramidal part, and these having decayed, many
of the most modern exhibit symptoms of decay which are not
observable in the older examples, where a stone lintel always
was employed. But it is not only in construction that the
Dravidians adhere to their old forms in temples. There are,
especially, some gopurams erected within the limits of last
century, which it requires a practised eye to distinguish from
older examples ; but with the civil buildings the case is quite
different. It is not, indeed, clear how a convenient palace could
be erected in the trabeate style of the temples, unless, indeed,
wood was very extensively employed, both in the supports and
the roofs. My conviction is that this really was the case, and
its being so, to a great extent at least, accounts for their dis-
appearance.

The principal apartments in what is called the palace at
Madura are situated round a courtyard which measures about
1 60 ft. east and west by 100 ft. north and south, surrounded on all
sides by arcades of very great beauty. The pillars which support
the arches are of stone, 40 ft. in height, and are joined by foliated
brick arcades of great elegance of design, carrying a cornice and
entablature rising to upwards of 60 ft. in height. The whole of
the ornamentation is worked out in the exquisitely fine stucco
called " chunnam " or shell lime, which is a characteristic of the
Madras Presidency.1 On the west side of the court stands the

1 For a fuller account of the palace, see
a paper by Mr R. F. Chisholm in the
' Transactions of the Royal Institute of
British Architects,' vol. xxvi. (1875-1876),

which the ground plan (fig. 240) is
reduced, and the measurements have
been taken from it. Details of the orna-
mentation are given by Dr G. Le Bor,

pp. I59ff. with plan and sections, from ' ' Les Monuments del'Inde,' figs. 233-235.

CHAP. V.

CIVIL ARCHITECTURE.

Swarga Vilasam or Celestial Pavilion, now used as the High
Court of Justice. It measures 235 ft. from north to south by

240.

Plan of Tirumalai Nayyak's Palace at Madura.

(From a Plan by Mr R. F. Chisholm.)

Scale 100 ft. to i in.

105 ft. across, and is arranged very much on the plan of a great
mosque with three domes — in fact, the whole structure, if not first
erected as a splendid mosque, is marvellously like one. The large
central dome is supported by twelve columns enclosing a square
64 ft. across. These columns are first linked together by massive
Saracenic arches. Four similar arches are then thrown across

414

DRAVIDIAN STYLE.

BOOK III.

the corners, and the octagonal drum rises from these, pierced by
a clerestory. Above this, at the cornice, 45^ ft. up, the octagon
is changed to a circle and the dome rises, in the centre, to 75 ft.
from the floor. At the north-west corner of the main building
is placed the splendid hall shown in the annexed woodcut
(No. 241 ),* the two corresponding with the Diwan-i-Khass and
Diwan-i-'Amm of Muhammadan palaces. This one, in its glory,

241.

Hall in Palace, Madura. (From Daniell's ' Views in Hindostan').

must have been as fine as any, barring the material : most of
the connected buildings, however, have long since disappeared.
This hall itself is 125 ft. long by 69 ft. wide, and its height to the
centre of the roof is 56 ft. ; but, what is more important than
its dimensions, it possesses all the structural propriety and
character of a Gothic building. It is evident that if the Hindus
had persevered a little longer in this direction they might have

1 In this view "a more decidedly
Saracenic character is given to the work
than it actually possesses." Mr Chisholm's

paper, ut sup. p. 161. The dimensions
appear much exaggerated if we take the two
small human figures as supplying a scale.

CHAP. V. CIVIL ARCHITECTURE. 415

accomplished something that would have surpassed the works of

416 DRAVIDIAN STYLE. BOOK III.

their masters in this form of art. In the meanwhile it is curious
to observe that the same king who built the chaultris (Wood-
cuts Nos. 226, 227, and 228), built also this hall. The style of
the one is as different from that of the other as Classic Italian
from Mediaeval Gothic ; the one as much over ornamented as the
other is too plain for the purposes of a palace, but both among
the best things of their class which have been built in the
country where they are found.

The last dynasty of Tanjor was founded by Ekoji or
Venkaji, a half brother of Sivaji, the great Maratha chief,
during the decline of the Madura dynasty in 1674-1675. The
palace was probably commenced shortly afterwards, but the
greater part of its buildings belong to the i8th century, and
some extend even into the iQth.

It is not unlike the Madura palace in arrangement — is,
indeed, evidently copied from it — nor very different in style ;
but the ornamentation is coarser and in more vulgar taste, as
might be expected from our knowledge of the people who
erected it (Woodcut No. 242). In some of the apartments
this is carried so far as to become almost offensive. One of
the most striking peculiarities of the palace is the roof of the
great hall externally. As you approach Tanjor, you see two
great vimanas, not unlike each other in dimensions or outline,
and at a distance can hardly distinguish which belongs to the
great temple. On closer inspection, however, that of the palace
turns out to be made up of dumpy pilasters and fat balusters,
and ill-designed mouldings of Italian architecture, mixed up
with a few details of Indian art ! A more curious and tasteless
jumble can hardly be found in Calcutta or Lucknow.

The palace buildings at Vijayanagar are much more
detached and scattered than those either at Tanjor or Madura,
but they are older, and probably represent only some of the
detached and less important buildings of what existed previous
to the sack of the city in 1565, when the Musalmans rased the
chief buildings to the ground.1 What still remains reproduces
more nearly the style of a Hindu prince's civil buildings, before
they fell completely under the sway of Moslim influence.
The remains of the palace consists of a number of detached
pavilions, baths, harems, and other buildings, that certainly
were situated in gardens, and may, consequently, have had a
unity we miss in their present state of desolation. One of
these pavilions is represented in the next woodcut (No. 243).
It is a fair specimen of that picturesque mixed style which
arose from the mixture of the Saracenic and Hindu styles.

1 Briggs's translation of Ferishta's ' Mahomedan Power in India,' vol. iii. p. 131.

CHAP. V.

CIVIL ARCHITECTURE.

Examples of Indian civil architecture are so few in the
south, that some notice may be included here of the old palace
at Chandragiri, about 30 miles N.N.E. from Chittur in North

243-

Garden Pavilion at Vijayanagar. (From a Photograph.

Arkot district It was from Chandragiri in 1639 that Sri
Rangaraya, the last representative of the Vijayanagar dynasty,
granted the British permission to erect the fort at Madras : six
years later he was overthrown by Jamshid Qutb Shah of
Golkonda.1

The principal building now left, as shown in Woodcuts
Nos. 244 and 245, presents a well-balanced facade of three
storeys surmounted by turrets which pleasingly break the sky-
line. Each floor consists mainly of a pillared hall — the piers
arched across both ways, corbelled at the angles and closed
with flat domes. The floors have projections of 6 or 7 ft.
beyond the face of the outer pillars and rest on stone corbels.
On the north side the walls pierced by the arches are built of
brick ; but the vaults, especially in the lower storey, are worked

1 After the battle of Talikot in 1565,
the representatives of the dynasty made
Pennakonda in Anantapur district their

VOL. T.

capital till 1592, when Venkatapati-
Raya removed to Chandragiri, where
the family resided till 1645.

2 D

DRAVIDIAN STYLE.

BOOK III.

in stone from stone corbels. No ornament now appears above
the basement, and how the exterior of the building was first

244.

South Elevation of Chandragiri Palace.

245. Plan of the ground floor of Chandragiri Palace.

(From Drawings by R. F. Chisholm, F.R.I.B.A.) Scale 50 ft. to i in.

finished it would be hard to say : we may be certain that it
was originally much finer than it now exists.1

Even the mixed style above mentioned, however, died out
wherever the Europeans settled, or their influence extended.
The modern palaces of the Nawabs of the Karnatik, of the Rajas
of Ramnad and Travankor, are all in the bastard Italian style,
adopted by the Nawabs of Lucknow and the Babus of Calcutta.
Sometimes, it must be confessed, the buildings are imposing
from their mass, and picturesque from their variety of outline,
but the details are always detestable ; first, from being bad
copies of a style that was not understood or appreciated, but
also, generally, from their being unsuited for the use to which

1 Fuller details will be found in a
paper illustrated by plans, elevations
and sections, by Mr R. F. Chisholm,

F.R.I.B.A., in the 'Indian Antiquary,'
vol xii. (1883), pp. 295-296.

CHAP. V. CIVIL ARCHITECTURE. 419

they were applied. To these defects, it must be added, that
the whole style is generally characterised by a vulgarity it is
difficult to understand in a people who have generally shown
themselves capable of so much refinement in former times.

In some parts of the north of India matters have not sunk
so low as in the Madras Presidency, but in the south native
civil architecture as a fine art is quite extinct, and though
sacred architecture still survives in a certain queer, quaint form
of temple-building, it is of so low a type that, as exemplified
especially in the temples which the Nathukottai Chettis are
engaged in renovating or reconstructing, it would be no matter
of regret if it, too, ceased to exist, and the curtain dropped
over the graves of both, as they are arts that practically have
become extinct.

( 420 )

BOOK IV.
CHALUKYAN STYLE.

CHAPTER I.

INTRODUCTORY.

CONTENTS.

Chalukyan Architecture — Dharwar temples — Ittagi — Gadag — Kuruvatti
— Dambal — Hanamkonda — Kirtti Stambhas at Worangal — Mysore —
Temples at Somnathpur and Belur— Temples at Halebid.

OF the three styles into which Hindu architecture naturally
divides itself, the Chalukyan is neither the least extensive
nor the least beautiful, but till about sixty years ago, it certainly
was the least known. The very name of the people was
hardly recognised by early writers on Indian subjects, and
the first clear ideas regarding them were put forward, in 1836,
in a paper by Sir Walter Elliot, in the fourth volume of the
'Journal of the Royal Asiatic Society.' To this he added
another paper, in the twentieth volume of the ' Madras
Journal ' : and since then numerous inscriptions of this dynasty
and of its allied families have been found and translated,
largely by Dr. J. F. Fleet, in the 'Indian Antiquary' and
' Journal of the Bombay Asiatic Society.'

From all this we gather that early in the 6th century
of our era, this family rose into importance at Badami, about
65 miles south of Bijapur in the Bombay Presidency1 — and
spread eastwards as far as the shores of the Bay of Bengal
between the mouths of the Krishna and Godavan, establishing
the capital of an eastern kingdom at Vengi early in the 7th

1 It is sometimes incorrectly stated
that Kalyan, in the Nizam's territory,
was their early capital ; but it was not

so before the middle of the loth century,
and under the later Chalukyas. — 'Indian
Antiquary,' vol. v. p. 318.

CHAP. I. DHARWAR TEMPLES. 421

century. They extended, in fact, from shore to shore, right
across the peninsula, and occupied a considerable portion of
the country now known as Mysore, and northward extended
as far, at least, as Nasik and Daulatabad.

The first prince of whom we know more than the name
is Pulike^in I., who began to rule at Badami about A.D. 550,
under whose two sons in succession the kingdom was largely
extended, towards the west at the expense of the Kadambas
of Banawasi, and — pushing northwards into Gujarat, where
they were checked by Siladitya I. of Valabhi. Early in
the next century Pulikej-in II. (609-642) further extended his
dominions, held at bay the forces of the great Harshavardhana
of Kanauj, invaded the Chola and Pallava territories, and
placed his brother Kubj a- Vishnu vardhana over the country
about the deltas of the Godavari and Krishna rivers, who —
about 617 — established the eastern Chalukya dynasty at Vengi.
About 626 Pulike^in II. seems to have received an embassy
from Khosru II. of Persia;1 and in 640 or 641, Hiuen Tsiang
traversed his kingdom and apparently visited his temporary
capital — probably at Nasik.2 But very soon after, towards
the close of his reign, his rule was upset by the Pallava King
Narasimhavarman, who took and plundered Badami or Vatapi,
as it was then called. By 655 Vikramaditya I. had recovered
the Chalukya dominions and even entered Kanchi.3

The family religion of the early kings was Vaishnava,
but they seem to have been very tolerant, if not eclectic, and
made grants liberally to 5aiva and Jaina temples as well as
to Vaishnava ones.

Like all the dynasties of central and northern India, the
Chalukyas suffered eclipse in the 8th century. They were
overthrown by the Rashtrakutas of Malkhed about 756, and
it was not till 973 that a descendant recovered the kingdom
and made his capital at Kalyani in the Dekhan. The temples
at Aihole and Pattadakal described above (Woodcuts Nos. 181
and 205) belong to their age, and we know they were erected
by early kings of this race ; but they do not belong to their
style. Their jikharas, or towers, either show the curvilinear
outline of the northern style, or the storeyed pyramids of the

1 'Journal of the Royal Asiatic Society,' | more than 150 miles from Bharoch, Dr.

N.S., vol. xi. p 165.

2 Hiuen Tsiang does not appear to have
visited Badami. From Kanchipuram he
went north-west (according to the ' Life ')
to Kong-kin-na-pu-lo, the position of
which is not known. From that he
went to the chief city of Maharashtra,
which he does not name, but being not

J. F. Fleet has pointed out that Nasik
best suits the conditions. — ' Indian
Antiquary,' vol. xxii. pp. 1136".

3 'Indian Antiquary,' vol. xix. pp. 1 5 if. ;
or for a detailed outline of the historical
materials, by Dr. J. F. Fleet, see
' Bombay Gazetteer,' vol. i. pt. ii. pp.
33S-38I.

42a CHALUKYAN STYLE. BOOK IV.

Dravidian. It is as if this intrusive race adopted hesitatingly
the earlier styles of the country, but that it was not till they
had consolidated their power, and developed peculiar insti-
tutions of their own, that they expressed them in the style
to which their name has been affixed.

The materials doubtless exist for settling these and most
other questions connected with this style; but, unfortunately,
much of them are locked up in the offices of the Archaeological
Survey ; and probably more are to be found in the Nizam's
territory, which is still almost a terra incognita to us in so
far as architecture is concerned. No extended survey has
yet been made of such remains as may exist there by any
one having a knowledge of the art or of the interest attaching
to the forms and age of the buildings.

The Muhammadan invaders from the beginning of the
1 4th century spared no temples that came in their way on any
of their raids, and doubtless the largest were the greater
sufferers. But after the final conquest and the rise of the
Musalman dynasties — the line of their capitals — Bijapur,
Kulbarga, Bidar, and Haidarabad — which have long occupied
the native country of the Chalukyas, is painfully suggestive of
the destruction of Hindu temples ; and the ruins and broken
sculptures that lie all round the neighbourhood of Kalyani — at
N^rayanpur, Sitapur, Tiprad, etc., bear abundant testimony to
the iconoclastic zeal of the conquerors.1 But still the wealth
of remains that exists in Dharwar and Mysore on the south and
west, and in the Berars on the north of the Nizam's territories,
is so great that all certainly cannot have perished, and many
will probably yet be found to solve the historical enigmas,
though they may not be sufficient to restore the style in its
integrity.

The Chalukyan style was naturally evolved from the
Dravidian, and the earliest temples within its area are not
always clearly marked off from that type : it was only by degrees
that it acquired its distinctive character. Unfortunately, most
of the earlier and the finer examples perished during the early
Moslim invasions and under the later rule of the various
Muhammadan dynasties of the Dekhan. The area over which
the style extended includes Mysore and all the Kanarese
country — its birthplace — in the west ; eastwards its southern
limit was the Tungabhadra and Krishna rivers ; and on the
north it perhaps extended to a line drawn from the south end
of the Chilka lake towards Nagpur, and thence westwards and
south-westwards to the coast. But we know too little as yet of

1 ' Archaeological Survey of Western India,' vol. iii. pp. 20, 23, 38-40.

CHAP. I.

DHARWAR TEMPLES.

423

the remains in the Haidarabad districts to define the northern
boundary with any certainty.

From the remains that have escaped entire destruction in
this area, we gather that the earlier Chalukyan temples pre-
served on the whole the general plan of the Dravidian shrines,
but the corners were made more prominent by flat increments
placed on them, whilst the projections on the walls were but
slight, the central one on each face of the shrine being made
much broader and important. The .nkhara and roof soon lost
the distinctively southern storeyed form and became stepped,
forming pyramids of different heights, with breaks corresponding
to those of the walls, and with broad bands up the sides of the
j-ikhara answering to the larger face in the middle of each side
of the shrine. Later, the plan often became star-shaped, the pro-
jecting angles lying in circles whose centres were in the middle of
the shrine and mandap respectively.1 The broader faces on the
sides, however, were retained for the principal images of the cult.
The pillars supporting the roof of the halls or mandaps were
arranged in squares ; the device of placing twelve pillars so
disposed in a square that eight of them could be connected by
lintels to support a roof or dome of larger dimensions was almost
unknown to the style.2

A favourite arrangement in the later temples was the group-
ing of three shrines round a central mandapa or hall. The
pillars are markedly different from the Dravidian type ; they are
massive, often circular, richly carved and highly polished. They
are usually in pairs or fours of the same pattern, the whole effect
being singularly elegant. Their capitals are wide with numerous
thin mouldings immediately below the abacus ; and under these
is a square block, whilst the middle of the shaft is carved with
circular mouldings. Frequently the capitals and shafts have
been actually turned in a sort of lathe in which the shaft was
held vertically.

In Dravidian temples at Badami, Pattadakal, Elura, and
elsewhere, pierced stone windows are not unfrequent, but the
most richly carved examples of these belong specially to the
Chalukyan style. Generally the temples stand on a terrace,
sometimes 10 to 15 ft. wide, quite surrounding them, and from

1 ' Archaeological Survey of Western
India,' vol. iii. p. 21 and plate 18. The
star-shaped plan is obtained by the over-
lapping of equal squares having a common
centre with their corners all equidistant ;
and as projecting angles must always
correspond to the corners of the cella and
mandapa, the number of angles between
the wider faces on two adjacent sides are

usually three or five. The re-entrant
angles will then always be larger than
right angles. With unequal projections
the corners, in a plan similar to that in
the Northern style, may also lie in a circle.
2 In the outer mandapa of the great
temple at Hangal, however, it was intro-
duced.

424

CHALUKYAN STYLE. BOOK IV.

3 to 6 feet in height — a feature which adds considerably to
the architectural effect. The structures were erected without
mortar, and the joints very carefully fitted. The whole outer
surface was covered with great variety of sculpture, of floral and
geometric patterns intermixed with mythological figures ; and
generally the mouldings of the base were carved with the
succession of animal patterns prescribed in the ' Silpa Sastras '
or architectural treatises.

The Dharwar district may be regarded as the cradle of the
style, and it may help to make its features better understood, if
before describing the remains farther east and south, we first
•notice some of the larger temples near and in that district.

At Ittagi, a small village in the Haidarabad districts, lying
some 21 miles east-north-east from Gadag, is a large Saiva
temple surrounded by the ruins of smaller shrines, etc., belonging
at latest to the early half of the nth century, which, though
deserted and partly ruinous, must be regarded as one of the
most highly finished and architecturally perfect of the Chalukyan
shrines that have come down to us. In the opinion of the late
Meadows Taylor, the principal temple is perhaps superior in
decorative art even to the Gadag temples. In it "the carving
on some of the pillars and of the lintels and architraves of the
doors is quite beyond description. No chased work in silver or
gold could possibly be finer. ... By what tools this very hard,
tough stone could have been wrought and polished as it is, is
not at all intelligible at the present day ; nor indeed from
whence the large blocks of greenstone rock were brought."1
A plan of the group is given on Woodcut No. 246, and it may
be noted that the plan of the shrine is not star-shaped, but
follows the old Dravidian form. The outer walls of the small
shrines have been stripped for building stone by the villagers.

The temple, shown in Plate No. XII., consists of an open
mandap and a closed hall with antechamber and linga shrine.
The square dome that once crowned the jikhara and the super-
structures on the roofs of the mandapas, with most of the screen
wall round the outer one, are now lost, as well as much of the
projecting cornice round the latter and over the entrance porches.

The form of these cornices, it may be remarked, indicates to
some extent the age of individual structures, as it is probable
the flat sloping form preceded the more ornamental one with
double flexure. The inner hall is 27 ft. square, and besides the
entrance from the front mandap, has also doorways on the north
and south sides, with pillared porches. The jambs and lintels
of all the entrances, as is usual in temples from the earliest

i ' Architecture of Dharwar and Mysore,' pp. 47-48. The stone is not so " very
hard," however, as Meadows Taylor had supposed.

CHAP. I.

ITTAGI.

425

period and in the caves, are very elaborately carved ; and their
porches have ribbed roofs of beautiful and ingenious designs.

246. Plan of the Ittagi Temples. (From a Drawing by Mr. H. Cousens.)
Scale too ft. to i in,

The central niches on the outsides of the shrine walls here
are larger and more prominent than in most other temples, and
have a half jikhara minutely carved over each. The niches
themselves, once occupied by characteristic Saiva images, are
now quite empty. Above each of these is a succession of highly
ornamental carved frontons, rising one above another, in decreas-
ing sizes to the summit. The pillars in the open mandap are of
varied patterns, richly carved, the central four being star-shaped
in plan, as are also the two outer pillars of each side porch, — a
form that was a favourite one, as it occurs in various temples of
this period and later, as in the Dambal temple of Dodda
Basavanna. The central panel in the roof, 8 ft. 5 in. square, is
also richly sculptured, with a figure of Nate^a in the middle
and, in the corners, are artistic arabesques in high relief round
kirttimukh faces.1

1 For a drawing and section of this
panel, which is in five stages, fully 4 ft.
in total depth, the uppermost being 3 ft.
1 1 in. square, with descriptive note, see

' Technical Art Series of Illustrations of
Indian Architectural Decorative Work,'
1888, plate 6,

426

CHALUKYAN STYLE.

BOOK IV.

At Kukkanur, within 4 miles northwards from Ittagi is a
notable group of temples, indicating that at an early period the
place must have been of considerable note. Most of them stand
together in a walled enclosure inside the village, about 82 yds. in
length from north to south, and 75 yds. from east to west. On
the east side is a massive gateway of perhaps the i6th century,
and the temples are arranged as in the plan (Woodcut No. 247).
The larger temple opposite the entrance is in use, and is

247. Plan of Temples at Kukkanur. Scale 100 ft. to i in.

perhaps, in its present form, of about the age of the enclosure.
But on the west side of it is an interesting congeries of nine
shrines, now containing lingas, of an earlier date — possibly
going back to the pth century. The .dkharas of these are more
Dravidian in contour than Chalukyan ; but the pillars inside
with round shafts and spreading capitals, and the tendency to
reduce and render less emphatic the storeys of the spires, mark
them as indicating an early advance towards the latter style.
The outer surfaces of the walls also are overlaid with carving
after the Chalukyan fashion ; but the red sandstone of which
they are built has decayed considerably from exposure and
obliterated much of the detail.

There are other temples here of which some are probably
of about the same age as these. To the south-west of the
village is one, in tolerably good preservation, which is quite
Dravidian in outline with indications of a tendency to the more

CHAP. I.

GADAG.

427

elegant proportions of the Chalukyan forms, and with gablets
on the middle of each tier of the .dkhara, indicative of the
beginning of the continuous band up each face that is so
prominent a feature in the later examples.

Among the five more notable temples at Gadag, that of
Somejvar, now appropriated as a school, though small, is worth
notice. Except the upper por-
tion of the .rikhara and the outer
open mandap that stood on the
east, it is still fairly entire. From
the plan (Woodcut No. 248) and
the photograph (Plate XIII.),
it appears to be a typical ex-
ample of the style. The mandap
is 21 ft. 10 in. square inside with
four round pillars, 9 ft. 4 in.
between centres, supporting the
roof; the antarala or ante-
chamber to the shrine is here
cut off from the mandap by a
screen with a door of its own,
and is 9 ft. square with a narrow
window on each side. The

Shrine itself is O ft. 3 in. square, 248- Plan of Someuvar Temple, Gadag.

. J _. . n. Scale 25 ft. to i in.

and, as IS common in Chalukyan (From a Plan by Mr. H. Cousens.)

temples, it has no pradakshina.

The doorways and all the outside walls are very richly decorated
with carving. The whole temple, however, seems to be silted up
to the level of the top of the raised basement on which these
temples usually stand, and which gave the rich wall sculptures
that elevation at which they could be seen to advantage. As in
other instances, the materials of the basement had probably been,
at least partly, removed before the ground level was raised to its
present height.1

Of quite as much interest is the temple of Trikuteyvar in the
fort of Gadag.2 As the name may imply, it is a triple-shrined
temple with two mandapas in line with the east and west shrines,
and the third cella built off at a later date from between the
mandapas on the north side. The carving has been exceedingly
good, but the figures have been much defaced by violence, and
the whole overlaid by successive coats of whitewash. Round the

1 Close to this temple is another known
as that of Rame^vara, of the same plan and
style, but much plainer, having scarcely
any sculpture on its walls, and is more
dilapidated.

2 Trikutejvara — ' lord of the three-
crested mountain.' Trikuta is the
mountain in Ceylon on which Ravana's
capital stood, also a range on the south
of the fabled Mount Meru.

428

CHALUKYAN STYLE.

BOOK IV.

mandapas the perforated screen work is well worth attention.
The original jikhara had been ruined, and is replaced by a
modern brick tower, altogether unworthy of the rest of the
building, which otherwise may be compared with the Belur
temple mentioned below. An inscription records the restora-
tion, or perhaps reconstruction, of this temple in the loth century,
to which period it possibly belongs ; but there were later repairs
at least, as under some of the images on the exterior are tablet
inscriptions of A.D. 1192, 1199, and 1213.

Close to this, on its south side, is a small temple dedicated
to Sarasvati, the goddess of learning and consort of Brahma,
and is one of the most remarkable for elaborate sculpture
even among those of this style. It is sadly dilapidated, but
the pillars are of the most varied patterns, and carved in the
minutest detail, as are also the screen walls, the roof, and
doorway of the shrine.1

As already noted, the doorways of Indian temples have always
been objects of sculptural decoration. Panels in the roofs, as in
the instance cited at Ittagi, have also been treated with like care,
a favourite device being the division into nine compartments in
each of which was represented the mythological guardians,
regents, or Dikpalas of the eight points of the compass, the
central panel being appropriated to the presiding divinity of
the temple.2 The rich sculptures on the Muhammadan Mihrabs
may be, and probably are largely due to the taste of Hindu
workmen, who applied to them the decorative style they had been
accustomed to employ on the doorways of their own shrines.

The subject cannot be adequately illustrated in a work of
this compass, but a single example (Plate No. XIV.),3 may
help to convey some idea of the character of the doorways
of Chalukyan temples. This is from one of the many very
interesting old temples at Lakkundi, a village 7 miles east-
south-east from Gadag. These temples are mostly fallen to
ruin, having suffered severely from the Chola invasion in the
nth century, when those at Lakshme^var were also destroyed.
And, at a much later date, in the feuds between the Brahmans
and Lingayats, they further suffered. The finest and one of
the largest of these temples is that of Ka^ivijve^var, which is a
double one, a western temple consisting of a shrine, hall and
porch, and a smaller eastern shrine with antarala, connected
with the porch to the west by a raised platform.

The doorways on the south and east of the hall are beautiful

1 For a fuller description of the temples
at Gadag by the editor, see 'Bombay
Gazetteer,' vol. xxii. (Dharwar)pp. 7i3ff.

8 A. good example of this, from

Ganginkatti in Dharwar, is given in
' Technical Art Series,' 1888, plate I.

3 From 'Technical Art Series,' 1887,
plate 3.

PLATE XIV.

DOORWAY OF THE SHRINE IX KASIVISVESVAR TEMPLE AT LAKKUXDI.

Scale i -25th.

[To face page 428, Vol. I.

CHAP. I.

KURUVATTI.

429

examples of intricate and delicate chiselling ; but the shrine
doorway, here represented, rivals the others in design and
workmanship. This doorway, though small — the entrance being
only 2 ft. 9 in. wide by 5 ft. 8 in. high — is of good proportions.
The variety of ornamental detail on the three fascias within the
pilasters may be studied on the illustration. It is so delicately
chiselled and in parts so undercut as to be almost detached from
the stone. The pilasters support a cornice over the door frame,
and from its centre depends a shield presenting Gaja-Lakshmt
or Sri, the goddess of success, bathed by elephants. On the
upper side of the cornice are elephants righting, soldiers and
other figures, spiritedly executed, but now much damaged.
Above is a frieze divided into five panels by carved uprights,
and containing figures of .Siva and Parvati with attendants in
the central one, Brahma and Vishnu in those to left and right,
and devotees in the end panels. This is surmounted by a
projecting moulding carved with a leaf ornament that appears
in all ages of Hindu art. In the recesses between the mouldings
of the doorway and the pilasters, supporting the roof, a single
figure is inserted and a rod or stalk with leaf tracery branching
off but stopping below the capitals of the door pilasters. The
three fascias of the architrave are also very richly and beautifully
sculptured.

At Chaudadampur on the Tungabhadra, about 12 miles
north from the railway at Ranibennur, is a fine bold temple
of black stone belonging to the nth century, with all its
details more completely finished than in some others. As will
be noticed in the photograph (Plate XV.), its defects in design
are the form of its dome and the insignificance of its crowning
member or finial, which latter, however, is probably not the
original Kalas.

KURUVATTI.

At Kuruvatti on the right bank of the Tungabhadra, 17 miles
west from Harpanahalli, and about 3 miles from Chaudadampur,
is a temple now dedicated to Malikarjuna (Woodcut No. 249).
It is on the same general plan as that of Some^var at Gadag,
but with a porch and doorway on the north as well as the
south side of the mandap. In the shrine are three recesses in
the walls, either for so many important images, where the
temple had only one shrine, or for the vessels used in the worship.1

1 In the Ke.rava temple at Huvinaha-
dagalli, in the old Chalukya temple at
Nagai, 25 miles south-east from Kulbarga,
and in others, are similar deep recesses.

— Rea's 'Chalukyan Architecture,' p. 21
and plate 92; H. Stone, 'The Nizam's
State Railway,' plan at p. 198.

430

CHALUKYAN STYLE.

BOOK IV.

The square antechamber has two narrow
windows and a richly-carved doorstep —
the " moonstone " of Ceylon — in front of
the cell door. As may be seen from the
plan, the mandap is small — only about
19 ft. square, with the usual four pillars
supporting the roof. They have square
bases, highly carved on each face, and
round shafts broken into numerous
members. As will be seen from the
south elevation (Woodcut No. 25O),1 the
roof of the mandap has been destroyed,
but the jikhar is entire — except the finial
which is a modern restoration. The pro-
jecting porches to the north and south
entrances have three pillars on each side
— square in plan with the corners slightly
recessed. The panels in the walls,

249. Plan oflSSTvatti Temple, formed by very attenuated pilasters, are

Scale 50 ft. to i in. carved above, some with imitations of

.nkharas and others with a sort of

festoon issuing from the mouths of makaras on each side. In

250. Kuruvatti Temple, south elevation, Scale i-2ooth.

the south door stands a loose slab over 6 ft. high by 4 ft.

1 From Rea's 'Chalukyan Architecture.'

CHAP. I.

DAMBAL.

broad, carved with a male figure with two arms and attended
by four females. All the details are sharp and the carving so
good that even at Halebid it would be difficult to point out
any individual piece showing more complete mastery over the
material than the brackets representing female figures with
encircling wreaths on the fronts and inner sides of the capitals
at the east entrance.1 The temple is probably of somewhat
later date than the preceding.

Dambal, some 13 miles south-east from Gadag, and 16 miles
south-west from Ittagi, must have been, in early days, a seat of
Buddhism, for we find that in A.D. 1095 a Buddhist inscription
there makes mention of a vihara built by sixteen Settis, and of
another vihara of Taradevi at Lakkundi.2 It has still three old
Saiva temples — all much injured. That of Dodda Basappa or
Basavanna, outside the town to the north-east, differs in plan
from any of the known temples in Dharwar districts (Woodcut
No. 251). It presents us with what appears to be a late form of
the Chalukyan jikhara, without the broad faces on the north,
west and south sides. The plan is star-shaped on the outside,
being formed of numerous rectangular points, which represent
the corners of six squares whose diagonals vary round a
common centre by 1 5 degrees each. The plan of the mandap
is similarly formed with eight squares at equal angles. The
angles are carried up the walls and roofs of shrine and hall.
The smaller string-courses of the roof being left in block, may
indicate that the work was not entirely
finished, though the effect is as sparkling
as if they had been completed to the
extent originally intended. But even as
it stands it would not be easy to point
to a more graceful form of roof for the
shrine. At first sight it may appear
somewhat strange and outre, but its form
gains with familiarity on the judgment
of the architectural critic.

The hall measures 23^ ft. square
inside, and, like the Some^var temple at
Gadag and others of the same class, it
has an entrance from the south as well
as from the east. A long porch has
been roughly built at some late date,
projecting from the front to cover a Scale 50 ft. to i in.
gigantic Nandi or bull of Siva. The
two pillars of the south porch and the doorway have been

1 Rea's ' Chalukyan Architecture,' pp.
21-24 and plates 56-68.

2 'Indian Antiquary,' vol. x. pp. 185-
190.

432 CHALUKYAN STYLE. BOOK IV.

elaborately carved ; and the dome of the mandap is supported
by four pillars also equally richly chiselled. Spanning the two
slender pillars in front of the antechamber or antarala is a
frieze, 8 ft. in length and between 3 and 4 ft. in height, richly
carved with unrivalled care though now damaged. And at the
entrance of the shrine is a doorstep — perhaps the most beautiful
in design in any temple in Western India. This temple was
perhaps one of the latest, designed late in the I2th century,
before the Muhammadan raids put an end to temple building.

Lastly, about 9 miles north from Chaudadampur, at the
sacred junction of the Varada with the Tungabhadra, in the
small village of Galaganath is another Chalukya temple, dedicated
to Galagej-vara (Plate XVI.). It is built of black granite, and its
appearance is striking owing to the base of the shrine or vimana
being entirely surrounded by a peculiar stepped abutment that
looks somewhat like an afterthought, and is quite out of place as
an architectural feature. Had it been below the shrine walls it
might have been contrived to add dignity to the tower ; but as
it is it gives the whole spire a much more pyramidal form than
in other temples, and it is not elegant. It may be, however, that,
to prevent the sinking of the foundations in deep sandy soil, the
base was extended, at a later date, to great thickness to support
the superincumbent weight of the .rikhara. Otherwise, though
of no great dimensions — about 80 ft. by 40 over all — this temple
is a good example of the period when it was erected, or about the
first half of the 1 1 th century.

HANAMKONDA AND WORANGAL.

When the Haidarabad or Nizam's territory has been
examined and completely surveyed, we shall probably be
able to trace all the steps by which earlier examples developed
into the metropolitan temple of Anamkond or Hanamkonda,
the old capital, 4 miles north-west of Orangal or Worangal
fort. According to an inscription on it, this temple was erected
in A.D. II62,1 by Pratapa Rudra, who, though not a Chalukya in
blood, but a Ganapatiya or Kakatfya, had succeeded to their
possessions and their style.

The temple itself is triple, having three shrines of very
considerable dimensions, dedicated to Siva, Vishnu, and Surya,
arranged round a central hall. In front of this temple is a great
mandapa or portico, supported on pillars, of which one hundred
and thirty-two are free standing, disposed in a varied pattern,
but without any sign of the octagonal arrangement for a dome.
Between this portico and the temple was the pavilion for the

1 ' Indian Antiquary,' vol. xi. pp. gff.

CHAP. I.

TEMPLE AT HANAMKONDA.

433

Nandi which has fallen within the last twenty-five years, and the
huge granite bull still remains among the fallen pillars and lintels'
This Nandi pavilion and the great pillared hall were all of granite

252. Doorway of Great Temple at Hanamkonda. (From a Photograph.)

and comparatively plain and devoid of ornament. The temple
itself measures outside 102 ft. from east to west, by 83 ft. 8 in.
from north to south, and stands on a raised basement or podium
10 ft. wide all round it.1 Like others of these late temples, this
one was never finished. It was too extensive for one king's reign,

1 Views of the hall or rnantapam and the temple are given in Workman's
'Through Town and Jungle," pp. 127 and 129.

VOL. I.

2 E

434

CHALUKYAN STYLE.

BOOK IV

even for one so powerful as he was who undertook it, and before
it was heartily taken up again the Muhammadans were upon
them (in A.D. 1310), and there was an end of Hindu greatness
and of Hindu art.

Some of its details, however, are of great beauty, especially
the entrances to the shrines, which are objects on which the
architects, as usual, lavished their utmost skill. The preceding
woodcut (No. 252) will explain the form of those of the great
temple, as well as the general ordinances of the pillars of the

253-

Kirtti Stambha at Orangal. (From a Photograph.)

great mandapa. Nothing in Hindu art is more pleasing than
the pierced slabs which the Chalukyas used for screens and
windows. They are not, so far as I recollect, used extensively
in other styles, but as used by them are highly ornamental
and appropriate, both externally and internally.

CHAP. I.

TEMPLE AT HANAMKONDA.

435

The pillars, too, are rich, without being overdone ; l and as
the central four in the main hall are of the same design and
beautifully carved, the effect of the whole is singularly varied,
but at the same time pleasing and elegant. The roof also is
covered by great slabs richly sculptured.

In Orangal or Worangal fort there are four Kirtti Stambhas,
as they are called (Woodcut No. 253), of one pattern, facing
one another, which have formed the gateways to a large temple
that once occupied the centre of the area, but of which only
a group of pillars and lintels belonging to its south-west corner
now remains. The distance between the north and south gate-
ways is 480 ft., and between those on the east and west 433 ft.,
leaving ample space for a temple of unusual size and splendour.2
It cannot be said they are particularly elegant specimens
of art. Their main interest lies in their being the lineal
descendants of the four gateways at Sanchi (Woodcut No. 38),
and they are curious as exemplifying how, in the course of a
thousand years or thereabouts, a wooden style of building
may lose all traces of its origin as clearly as they do ; for it
seems most unlikely that any such form could have been
invented by any one using stone constructions, and that only.

There are also in the Orangal fort a great number of smaller
temples and shrines, in the same style as the great temple, and,
like it, apparently mostly dedicated to Siva, from the presence
of his bull almost everywhere. Most are ruined ; and, judging
from appearances, I am inclined to believe this is owing to
Moslim violence. The mode of building is without mortar,
and the joints are by no means well fitted. The style is also
remarkably free from figure-sculpture, which is generally the
thing that most easily excites the iconoclastic feelings of the
followers of the Prophet.

Lastly a simple example of the style in a village temple
has been cited at a place called Buchhanapalli, not far from
Haidarabad 3 (Woodcut No. 254). There are four principal
faces on the walls of the shrine, larger than the others : three

1 The most elaborately chased pillars
of this style are to be seen in the temple
of Ramappa near Palampet, about 30
miles north-east from Hanamkonda.

- Cousens' ' Lists of Antiquarian Re-
mains in the Nizam's Territory,' pp. 4&f.
To Mr Cousens is due the materials
for the revision of this account of the
Worangal and Hanamkonda remains.

* ' Buchropully ' was placed on Mr
Fergusson's map 50 miles west from
Haidarabad, and in that position (Lat.
17° 31' N., Long. 77° 48' E.) is the

village of Buchhanapalli, 14 miles north
from Dhariir railway station. Mr Cousens
( ' Lists of Antiquarian Remains in the
Nizam's Territories,' p. 2) conjectured that
it might be ' Bachropully,' 12 miles north-
west of Secundarabad. In ' Glimpses of
the Nizam's Dominions' by A. Claude
Campbell' (1898), pp. 452 and 459, are
two views of the shrine of a Hindu
temple at 'Bichpalli.' This place is
10 miles south-east from Indur ; and the
temple, which is of unusual plan and
considerable merit, stands on a rocky

436

CHALUKYAN STYLE.

BOOK IV.

occupied by niches, the fourth by the entrance. The roof is
in steps, and with a flat band on each face in continuation
of the larger face below. The summit ornament is a vase,
in this instance apparently incomplete. The porch is simple,
consisting only of sixteen pillars, disposed equidistantly, without
any attempt at the octagonal dome of the Jains or the varied

254-

Temple at Buchhanapalli. (From a Photograph.)

arrangements subsequently attempted. The .rikhara is a straight-
lined cone, and its decorations in steps is as unlike the Dravidian
spire in storeys as it is to the curvilinear outline of the Jaina and
northern temples. The porch too, is open, and consists of
columns spaced equidistantly over its floor, without either the
bracketing arrangements of the southern or the domical forms of
the northern styles. Situated as it was locally, half-way between
the Dravidian and northern styles, the Chalukyan retained or
borrowed occasionally a feature or form from one or from the
other, but not to such an extent as to obliterate its individuality,
or to prevent its being recognised as a separate and distinct style
of architecture.

knoll near the village. It consists of inserted short circular pillars surmounted

shrine, hall, and porch, but the .rikhara by rampant Vyilis. — Cousens, 'Lists of

and roof are ruined. Round the shrine Antiquarian Remains : Nizam's Terri-

js an open pradakshina, in which are tory,' p. 63.

CHAP. I.

MYSORE.

437

MYSORE.

It is in the province of Mysore, however, that the Chalukyan
style attained its fullest development and highest degree of
perfection during the three centuries — A.D. 1000 to 1300 —
in which the Hoysala Ballalas had supreme sway in that
country. Several temples, or rather groups of temples, were
erected by them — one at a place called Somnathpur, a small
village on the left bank of the Kaveri, south of Mysore,
built by Soma, the general of Narasimha Ballala III., and was
completed in 1270;* another at Belur, in the centre of the
province, owed its origin apparently to Vishnuvardhana, in or
about A.D. 1117; the last and greatest at a place they called
Dorsamudra — now known as Halebid, 10 miles east by north
from the last-named, from which the capital was removed by
Vishnuvardhana about 1135. It continued to be the metropolis
of the kingdom, till it was destroyed, and the building of
the great temple stopped by the
Muhammadan invasion in A.D.
1310-1311.

Like the great temple at
Hanamkonda, the Kej-ava temple
at Somnathpur is triple, the cells,
with their jikharas, being attached
to a square pillared hall, to the
fourth side of which a portico is
attached, in this instance of very
moderate dimensions (Woodcut
No. 255). The whole stands in a
square cloistered court, measuring
210 ft. by 172 ft. over all, and
has the usual accompaniments of
entrance-porch, stambha, etc.

The following woodcut (No.
256) will give an idea — an im-
perfect one, it must be confessed
— of the elegance of outline and
marvellous elaboration of detail

that characterises these shrines. Its height seems to be only
about 30 ft., which, if it stood in the open, would be almost too
small for architectural effect ; but in the centre of an enclosed
court, and where there are no larger objects to contrast with
it, it is sufficient, when judiciously treated, to produce a

Plan of the Kesava Temple

at Somnathpur.
Scale loo ft. to i in.2

1 Rice's ' Mysore Gazetteer,' vol. i. p. 514.

* From a lithographed plan in Rice's ' Epigraphia Carnatica,' vol. iii. pt. i.

438

CHALUKYAN STYLE.

BOOK IV.

considerable impression of grandeur, and apparently does so in
this instance.

Temple at Somnathpftr. (From a Photograph.)

The temple at Somnathpur is a single but complete whole ;
that at Belur, on the other hand, consists of one principal

CHAP. I.

MYSORE.

439

temple, surrounded by four or five others and numerous
subordinate buildings, enclosed in a court by a high wall
measuring 380 ft. by 425 ft, and having two very fine gate-
ways or gopurams in its eastern front.1 As will be seen from
the plan (Woodcut No. 257), the great temple consists of a
very solid vimana, with an antarala, or vestibule ; and in front of
this a porch or mahamantapam of the usual star-like form,
measuring 90 ft. across. The entire length of the temple, from
the east door to the back of the cell,
is 115 ft., and the whole stands on
a terrace about 3 ft. high, and from
10 ft. to 15 ft. wide. This is one
of the characteristic features of
Chalukyan design, and adds very
considerably to the effect of their
temples.

The arrangements of the pillars
have much of that pleasing subor-
dination and variety of spacing
which is found in those of the
Jains, but we miss here the octa-
gonal dome, which gives such
poetry and meaning to the arrange-
ments they adopted. Instead of
that, we have only an exaggerated
compartment in the centre, which
fits nothing, and, though it does

give dignity to the Centre, it does 2S7i Plan of Chenna Kejava Temple at
it SO Clumsily as to be almost ?elur. Scaleiooft.toijn. (From

offensive in an architectural sense.

It is not, however, either to
its dimensions, or the disposition of its plan, that this temple
owes its pre-eminence among others of its class, but to the
marvellous elaboration and beauty of its details. The effect of
these, it is true, has been, in modern times, considerably marred
by the repeated coats of whitewash which the present low order
of priests consider the most appropriate way of adding to the
beauty of the most delicate sculptures. Notwithstanding this,
however, their outline can always be traced, and where the
whitewash has not been applied, or . has been worn off, their
beauty comes out with wonderful sharpness.

The following woodcut (No. 258) will convey some idea
of the richness and variety of pattern displayed in the windows

Rice's ' Epigraphia Carnatica,'
vol. v.)

1 By the plan in Mr Rice's ' Epigraphia Carnatica,' vol. v. pt. i., the court is about
404 ft. long on the north side, and 426 on the south.

440

CHALURYAN STYLE.

BOOK IV.

of the porch. These are twenty-eight in number, and all are
different. Some are pierced with merely conventional patterns,
generally star-shaped, and with foliaged bands between ; others

View of part of Porch at Belftr. (From a Photograph.)

are interspersed with figures and mythological subjects, the
nearest one, for instance, on the left, in the woodcut, represents
the Narasimha Avatara, and other different scenes connected
with the worship of Vishnu, to whom the temple is dedicated.
The pierced slabs themselves, however, are hardly so remark-
able as the richly-carved base on which they rest, and the deep
cornice which overshadows and protects them. The amount of
labour, indeed, which each facet of this porch displays is such
as, I believe, never was bestowed on any surface of equal extent
in any building in the world ; and though the design is not of
the highest order of art, it is elegant and appropriate, and never
offends against good taste.

The sculptures of the base of the vimana, which have not
been whitewashed, are as elaborate as those of the porch, in
some places more so ; and the mode in which the undersides
of the cornices have been elaborated and adorned is such as is
only to be found in temples of this class. The upper part of
the tower is anomalous. It may be that it has been white-

CHAP. I.

MYSORE.

441

washed and repaired till it has assumed its present discordant
appearance, which renders it certainly a blot on the whole
design. My own impression rather is, that, like others of its
class, it was at first left unfinished, and the upper part added
•at subsequent periods. Its original form most probably was
that of the little pavilions that adorns its portals, one of which
is represented in the following woodcut (No. 259), which has
all the peculiar features of the style — the flat band on each
face, the three star-like projections between, and the peculiar
crowning ornament of the style. The plan of the great tower,
and the presence of the pavilions
where they stand, seems to prove
almost beyond doubt that this
was the original design ; but the
design may have been altered as
it progressed, or it may, as I suspect,
have been changed afterwards.

There seems to be little or no
doubt about the date of this temple.
It was erected by Vishnuvardhana,
the fourth king of the race, to com-
memorate his conversion by the
celebrated Ramanujya from the
Jaina to the Hindu faith. He
ascended the throne, A.D. 1114, and
his conversion took place about
1117; and it is probable that the
temple was finished before his death
in 1141, but as the capital was re-
moved by the same king to Halebid,
it is just possible that the vimana of
the great temple, and the erection
of some at least of the smaller
shrines, may belong to a subsequent
period.

Mysore abounds in remains of
this style, but they have not been adequately surveyed.1 In the
north-west of the province, at Balagami or Belagavi, an ancient
site, there are some five ruined temples whose rich sculpture is
equal in taste and perfection of workmanship to any of the class.
Among these, the temple of Kedare^vara is perhaps the oldest,
and a view of it from the north-east is given in Plate No. XVII.

259. Pavilion at Belur.
Photograph. )

(From a

1 A number of lithographed ground
plans, to small scales, are published in
Rice's ' Epigraphia Carnatica,' but with-
out descriptions. In Workman's 'Through

Town and Jungle ' (1904) are some well-
selected photographic views of Somnath-
pur, Koravangula and other Mysore
temples.

442

CHALUKYAN STYLE.

BOOK IV.

In plan it bears a general resemblance to the Somnathpur
temple, having three shrines round a hall (Woodcut No. 255),
but, as in the Belur temple (No. 257), with projecting entrances
to the mandap or porch on both sides as well as on the east
front.1 The roof of the mandap is quite destroyed, probably
by Moslim violence, and a miserable covering of tiles replaces
it ; but the three vimanas or jikharas are fairly entire, with the
Syala. or leogriff emblem of the Hoysala kings seen over the
block on the roof, projecting in front of the central spire.
Beyond, to the west, is seen part of a ruined temple ; and to
the south is the old Nandi pavilion, now also covered with tiles ;
and the base of a dhwaja-stambha near the corner of it. These
Balagami temples are ascribed to Jakanacharya, the celebrated
architect and sculptor of the Hoysala kings, to whom is ascribed
also the Halebid, Somnathpur, Mulbagal and, by tradition,
many other temples in this region.

There are also at Kubbatur, near Balagami, and at
Haranhalli, Arasikere, Koravangula, Naglapur, Turuvekere, and
other places in Mysore, monuments that await and deserve,
more than almost any others, to be fully illustrated.2

HALEBiD.

The earliest temple known to exist at Halebid was a small
detached shrine dedicated to Siva as Kedare^vara,3 and was
erected by Vira Ballala and one of his queens, probably about
1219. Its general appearance, nearly forty years ago, will be
understood from the next woodcut (No. 260). It was star-
shaped in plan, with sixteen points, and had a porch, so entirely
ruined and covered up with vegetation that it was difficult to
make out its plan. Its roof was conical, and from the base-
ment to the summit it was covered with sculptures of the very
best class of Indian art, and these so arranged as not materially
to interfere with the outlines of the building, while they im-
parted to it an amount of richness only to be found among
specimens of Hindu art. If this little temple had been illus-
trated in anything like completeness, there was probably
nothing in India which would have conveyed a better idea of
what its architects were capable of accomplishing. But, alas,

1 There is a plan of this Balagami
temple and some details in ' Epigraphia
Carnatica,' vol. vii.

2 Dr. G. Le Bon remarks : ' Si jamais
il m'arrive de retourner dans 1'Inde, je
ferai certainement une etude speciale du
Mysore. Cette region, presque vierge,

offrira surement une abondante moisson
au explorateurs. ' — ' Les Monuments de
1'Inde,' p. 173.

3 This has been erroneously called
Kaite^var and Kaitabhejvara by some
writers. — 'Mysore Gazetteer,' vol. i.
p.

CHAP. I. HALEBID. 443

this cannot be : this gem of* Indian architecture is no more ;

260.

Kedaresvara, Halebid. (From a Photograph by Capt. Tripe.)

vegetation did its relentless work unchecked, and the pile is
long ago a shapeless mound.1

The Kedare^vara temple was, however, surpassed in size

1 In 1876 Mr. Fergusson wrote : — " In entirely destroyed by the trees, which
a very few years this building will be have fastened their roots in the joints

444

CHALUKYAN STYLE.

BOOK IV.

and magnificence by its neighbour, the Hoysale^vara temple at
Halebid, which, had it been completed, is one of the buildings
on which the advocate of Hindu architecture would desire to
take his stand. Unfortunately, it never was finished, the works
having probably been stopped by the Muhammadan conquest
in 1311 A.D.1

The general arrangements of the building are given on the
annexed plan (Woodcut No. 261), from which it will be perceived

that it is a double temple. If
it were cut into halves, each
part would be complete with a
pillared mantapam of the same
type as that at Belur above
referred to, an antarala or
intermediate vestibule, and a
sanctuary containing a lingam,
the emblem of Siva. Besides
this, each half would have in
front of it a detached, pillared
porch as a shrine for the Bull
Nandi, which, of course,
. would not be required in a
Vaishnava temple. Such
double temples are quite
common in India, but the
two sanctuaries usually face
each other, and have the porch
between them. Its dimensions may roughly be stated as 200 ft.
square over all, including all the detached pavilions. The

261. Plan of Hoysalervara Temple at Halebid
Scale 100 ft. to i in.

of the stones. In a drawing in the
Mackenzie collection in the India Office,
made in the early part of this century,
the building is shown entire. Twenty
years ago it was as shown at p. 443. A
subsequent photograph shows it almost
hidden ; a few years more, if some steps
are not taken to save it, it will have
perished entirely. A very small sum
would save it ; and, as the country is in
our charge, it is hoped that the expendi-
ture will not be grudged." But no
attention was paid to this warning, and
as Mr. L. Rice says : — " With shame be it
written — Mr. Fergusson'sgloomy anticipa-
tions have been completely fulfilled. . . .
Some of the most perfect figures have
been conveyed to Bangalor, and set up
in the Museum, but divorced from their
artistic setting they have lost their
meaning. A proposal has been made,

I believe, to convey the ruins to Mysore,
and erect the restored temple there as
a memorial to the late Maharaja."—
'Mysore Gazetteer' (1897), vol. i. p. 515.
Mr. Rice has preserved for us two photo-
graphs of the temple in 1866, and 1886
in ' Epigraphia Carnatica,' vol. v. pt. i.
Later, in 1907, the Mysore Government
tried to restore the temple, but the result
is reported as not very successful, as
empty spaces had to be filled in with
plain slabs. But these are much better
than the crude attempts made in other
cases to imitate the old work.

1 The date of its foundation is not
known, but as Halebid or Dorasamudra
became the capital only in the middle of
the 1 2th century, it was probably begun
somewhat later, and possibly well into
the next century.

CHAP. I.

HALEBID.

445

temple itself is 160 ft. north and south by 122 ft. east and west.
Its height, as it now remains, to the cornice is about 25 ft. from
the terrace on which it stands. It cannot, therefore, be considered
by any means as a large building, though large enough for effect.
This, however, can hardly be judged of as it now stands, for
there is no doubt but that it was intended to raise two great
pyramidal spires over the sanctuaries, four lower ones in front
of these, and two more, as roofs — one over each of the two
central pavilions. Thus completed, the temple would have

262.

Restored view of Temple at Halebid.

assumed something like the outline shown in the woodcut
(No. 262),1 and if carried out with the richness of detail
exhibited in the Kedare^vara (Woodcut No. 260), would
have made up a whole which it would be difficult to rival
anywhere.

The material out of which this temple is erected is an
indurated potstone of volcanic origin, found in the neighbour-
hood. This stone is said to be soft when first quarried, and
easily cut in that state, though hardening on exposure to the
atmosphere. Even this, however, will not diminish our admira-
tion of the amount of labour bestowed on the temple, for, from
the number of parts still unfinished, it is evident that, like most
others of its class, it was built in block, and carved after the
stone had become hard. As we now see it, the stone is of a

1 The finials on the two principal
jikharas, by some mistake of the engraver,
have been wrongly presented. A vase

supported on a small dome is to be
understood. — G. Le Bon, ' Les Monu-
ments de I'lnde,' fig. 265.

446 CHALUKYAN STYLE. BOOK IV.

pleasing creamy colour, and so close-grained as to take a polish
like marble. The pillars of the great Nan'di pavilion, which look
as if they had been turned in a lathe,1 are so polished as to
exhibit what the natives call a double reflection — in other words,
to reflect light from each other. The enduring qualities of the
stone seem to be unrivalled, for, though neglected and exposed
to all the vicissitudes of a tropical climate for seven centuries,
the minutest details are as clear and sharp as the day they were
finished. Except from the splitting of the stone arising from
bad masonry, the building is as perfect as when its erection was
stopped by the Muhammadan conquest.

It is, of course, impossible here to illustrate completely so
complicated and so varied a design ; but the following woodcut
(No. 263) will suffice to explain the general ordonnance of its
elevation. The building stands on a terrace ranging from 5 ft.
to 6 ft. in height, and paved with large slabs. On this stands a
frieze of elephants, following all the sinuosities of the plan and
extending to some 710 ft. in length, and containing not less
than 2000 elephants, most of them with riders and trappings,
sculptured as only an Oriental can represent the wisest of brutes.
Above these is a frieze of " jardulas," vyalas, or conventional
lions or tigers — the emblems of the Hoysala Ballalas who built
the temple. Then comes a scroll of infinite beauty and variety
of design; over this a frieze of horsemen and another scroll ; over
which is a bas-relief of scenes from the Ramayana, represent-
ing the conquest of Ceylon and all the varied incidents of that
epic. This, like the other, is about 700 ft. long. (The frieze of
the Parthenon is less than 550 ft.) Then come celestial or
conventional beasts and birds, and all along the east front a
frieze of groups from human life, and then a cornice, with a rail,
divided into panels, each containing two figures. Over this are
windows of pierced slabs, like those of Belur, though not so rich
or varied. These windows will be observed on the right and
left of the woodcut. In the centre, in place of the windows, is
first a scroll, and then a frieze of gods and heavenly apsarasas —
dancing girls and other objects of Hindu mythology. This
frieze, which is about 5 ft. 6 in. in height, is continued all round
the western front of the building, and extends to some 400 ft. in
length. Siva, with his consort Parvati seated on his knee, is
repeated at least fourteen times ; Vishnu in his various Avataras
even oftener. Brahma occurs several times, and every great
god of the Hindu Pantheon finds his place. Some of these are

1 They were, in fact, set vertically in a
sort of pit and turned, probably in water,
giving them a very smooth surface and

chasing out the very fine mouldings with
an accuracy and uniformity that could
hardly have been otherwise attained.

CHAP. I.

HALEBID.

447

263. Central Pavilion, Halebid, East Front. (From a Photograph.)

448 CHALUKYAN STYLE. BOOK IV.

carved with a minute elaboration of detail which can only be
reproduced by photography, and may probably be considered as
one of the most marvellous exhibitions of human labour to be
found even in the patient East.

It must not, however, be considered that it is only for
patient industry that this building is remarkable. The mode in
which the eastern face is broken up by the larger masses, so as
to give height and play of light and shade, is a better way of
accomplishing what the Gothic architects attempted by their
transepts and projections. This, however, is surpassed by the
western front, where the variety of outline, and the arrangement
and subordination of the various facets in which it is disposed,
must be considered as a masterpiece of design in its class. If
the frieze of gods were spread along a plain surface it would lose
more than half its effect, while the vertical angles, without
interfering with the continuity of the frieze, give height and
strength to the whole composition. The disposition of the
horizontal lines of the lower friezes is equally effective. Here
again the artistic combination of horizontal with vertical lines,
and the play of outline and of light and shade, far surpass
anything in Gothic art. The effects are just what the mediaeval
architects were often aiming at, but which they never attained
so perfectly as was done at Halebid.

Before leaving Halebid, it may be well again to call atten-
tion to the order of superposition of the different animal friezes,
alluded to already, when speaking of the rock-cut monastery
described by the Chinese Pilgrims (ante, p. 171). There, as here,
the lowest were the elephants ; then the lions ; above these
came the horses ; then the oxen ; and the fifth storey was made
with shapes of pigeons. The oxen here is replaced by a conven-
tional animal, and the pigeon also by a bird of a species that
would puzzle a naturalist The succession, however, is the same,
and, as mentioned above, the same five genera ol living things
form the ornaments of the " moonstone " thresholds of the various
monuments in Ceylon. Sometimes in modern Hindu temples
only two or three animal friezes are found, but the succession is
always the same, the elephants being the lowest, next above
them are the lions, and then the horses, etc. When we know
the cause of it, it seems as if this curious selection and succession
might lead to some very suggestive conclusions. At present we
can only call attention to it in hopes that further investigation
may afford the means of solving the mystery.

If it were possible to illustrate the Halebid temple to such
an extent as to render its peculiarities familiar, there would be
few things more interesting or more instructive than to institute
a comparison between it and the Parthenon at Athens. Not

CHAP. I. HALEBID. 449

that the two buildings are at all like one another ; on the
contrary, they form the two opposite poles — the alpha and
omega of architectural design ; but they are the best examples
of their class, and between these two extremes lies the whole
range of the art. The Parthenon is the best example we know
of pure refined intellectual power applied to the production of
an architectural design. Every part and every defect is calculated
with mathematical exactness, and executed with a mechanical
precision that never was equalled. All the curves are hyperbolas,
parabolas, or other developments of the highest mathematical
forms — every optical defect is foreseen and provided for, and
every part has a relation to every other part in so recondite a
proportion that we feel inclined to call it fanciful, because we
can hardly rise to its appreciation. The sculpture is exquisitely
designed to aid the perfection of the masonry — severe and
godlike, but with no condescension to the lower feelings of
humanity.

The Halebid temple is the opposite of all this. It is regular,
but with a studied variety of outline in plan, and even greater
variety in detail. All the pillars of the Parthenon are identical,
while no two facets of the Indian temple are the same; every
convolution of every scroll is different. No two canopies in
the whole building are alike, and every part exhibits a joyous
exuberance of fancy scorning every mechanical restraint. All
that is wild in human faith or warm in human feeling is found
portrayed on these walls ; but of pure intellect there is little —
less than there is of human feeling in the Parthenon.

It would be possible to arrange all the buildings of the
world between these two extremes, as they tended toward the
severe intellectual purity of the one, or to the playful exuberant
fancy of the other ; but perfection, if it existed, would be some-
where neaV the mean. My own impression is, that if the
so-called Gothic architects had been able to maintain for two
or three hundred years more the rate of progress they achieved
between the nth and the I4th century, they might have hit upon
that happy mean between severe constructive propriety and
. playful decorative imaginings which would have combined into
something more perfect than the world has yet seen. The
system, however, as I have endeavoured to point out elsewhere,
broke down before it had acquired the requisite degree of
refinement, and that hope was blighted never to be revived. If
architecture ever again assumes an onward path, it will not be
by leaning too strongly towards either of the extremes just
named, but by grasping somewhere the happy mean between
the two.

For our present purpose, the great value of the study of

yOL. I. 2 F

45o CHALUKYAN STYLE. BOOK IV.

these Indian examples is that it widens so immensely our basis
for architectural criticism. It is only by becoming familiar with
forms so utterly dissimilar from those we have hitherto been
conversant with, that we perceive how narrow is the purview
that is content with one form or one passing fashion. By rising
to this wider range we shall perceive that architecture is as
many-sided as human nature itself, and learn how few feelings
and how few aspirations of the human heart and brain there are
that cannot be expressed by its means. On the other hand,
it is only by taking this wide survey that we appreciate how
worthless any product of architectural art becomes which does
not honestly represent the thoughts and feelings of those who
built it, or the height of their loftiest aspirations.

.

174

PRINTED AT THE EDINBURGH PRESS,
9 AND II YOUNG STREET.

PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET

UNIVERSITY OF TORONTO LIBRARY