AN OUTLINE OF

INDIAN TEMPLE ARCHITECTURE

 BULLETIN
OF THE
MADRAS GOVERNMENT MUSEUM

AN OUTLINE OF INDIAN TEMPLE

ARCHITECTURE

By
F.H.GRA VEL Y, D.SC., F.R.A.S.B
Government Museum, Madras

Published by
The Director of Museums
Government Museum
Chennai-600 008.
2006

I
/
" 20Q-5-1
First Edition: 1936
Reprinted: 1950, 1960, 1976, 1992, 1999,2006

© Director of Museums
Government Museum, Chennai - 8.

Price: Rs. 45/

 M.A.SIDDIQUE, I.A.S Government Museum,
Director of Museums Chennai-600 008.

PREFACE

India is revered for her religious tolerance. Her temples present a

panoramic variety of aesthetic excellences with religious favour. Thiru F. H.

Gravely the famous indologist and the former Superintendent of this Museum, in a

brief outs ketch has explained the different styles of Indian Temple Architecture

developed at different zones of the sub-continent. This work, an authoritative

presentation on temple architecture was out of print for a long time. For the benefit

of the serious readers of Indian Architecture, now we reprint this work.

Chennai-8 M.A.SIDDIQUE, I.A.S.,

24.02.2006 Director of Museums.

206-5--1 a
 AN OULINE OF INDIAN TEMPLE ARCHITECTURE

By F.H Gravely, D.Se., F.R.A.S.B., Superintendent, Government Museum,

Madras

Since the foundation of the modern study of Indian architecture were laid
over half a century ago by the publication of Ferguson's "History of Indian ~nd .
Eastern Architecture" a vast amount of fresh information has become available; and
the time now seems ripe for a revision and simplification of his treatment of the
st~ :es adopted in the construction of temples. I llave already had occasion to make
~lIggestions in this direction in a paper by Mr. T.N. Ramachandran and myself on
"The Three Main Styles of Temples Architecture recognized by the Silpa-Saslra"
published as part I of Volume III of the Bulletin of the Madras Government Museum
(1934). The objects of the present paper are to develop this theme in such a way
was l\) provide. if possible, a unified framework as a basis for further study. that will
o;atisfactor1.1y interpret the facts so far as they are kno\\n; and to suggest a
terminology reasonably free from controversial implications.

It has been recognized that Fergussion'o; separation from one another of

Buddhist, lain and Hindu architecture is unsound. Buddhism and .lainism as we
known them to-day sprang out of Hinduism, and there is every reason to believe that
their temples have had a similar history. In any case it is clear that they are closely
related and cannot be fully understood independently of one another. Among
surviving monuments those relating to Buddhism naturally belong mainly to the
early times when Buddhism was most widely prevalent in India; and as Hindu
buildings were evidently then still being made exclusively (or almost so) of
materials of comparatively low durability, surviving Hindu monuments mostly
belong to subsequent times. It has therefore been found convenient to refer to the
fonner as the Buddhist architectural period. even though when taken literally this
name may be considered misleading. since Hinduism preceded Buddhism in India
and continued contemporaneously with it. as well as surviving it to the present day.
The difference between surviving Buddhist and Hindu monuments it thus mainly ..:
chronological. and no .essential architectural different seems to exist between them
and Jain monuments.

Buildings in different architectural styles may differ from one another either
in external fornl. or in decorative detail, or in both. Careful study of the forms and
evolution of decorative detail
2 Bulletin. Madras Government }\1useul11

in the temples of the Tamil country has already let Jouveau-Dubreuil to regard these temples
as belonging to a style distinct from that of the superficially similar temples of Kanarese
country I with which they are usually identified, a conclusion since confirmed by
Ramachandran and myself from a study of extemal from 2 • But as practically nothing seems
to have been worked out regarding the characteristics and evolution of the fonns of
decorative details found in temples outside the Tamil country, decorative detail cannot yet
be utilized in formulating any general system of classification~ for whieh therefore, for the
present, extemal form alone remains. The production of workable hYI=0theses that would
render intelligible the various forms of pillars, corbels, niches, etc., in temples outside the
Tamil country is indeed, long overdue and .would probably do more than could anything
else, both to facilitate the correct dating of building without inscriptional record of t~eir
history, and to make possible an intelligent popular interest in the archaeology of Indian
temples. .

The Vimana or shrine is the part of a temple the extemal form of which has, form as
long ago as the time of complication of the early Silpa-Sastras. been found to be of most
help in distinguishing between different styles. As a rule, moreover, it is the most
ccnspicuous part, though in the finest temples of the Tamil country build sl:bsequent to abut
1100 A.D. it is usually small, often quite insignificant and dominated by immense gopura
towers over the gateways. A gopura is always much broader that it is deep, but a vimana is
usually square or more or less rounded in plan. In the two chief fonns of Indian temple,
characteristic respectively of the north and south of the Peninsula, the towers consist of a
vertical-side body containing the cell a of the l'imana or the gateway of the gopura,
surmounted by the tapering portion which may conveniently be designated the spire. The
terminates, above a somewhat narrow neck, in the bulbous structure bearing the finial.
Fergusson' calls the bi.Jlbous structure the "crowning member", and as its correct technical
designation is not yet finally settled, I purpose to refer to it here as the crown.-l

Temples having different types of extemal form seem to have

originated independently in different parts of India, either direct from earlier structures built
of more or less perishable materials or from small flat-roofed temples of which a few
survive from the Gupta period. 5 • Over a vast area extending from t~le base of the
Himalayas southwards to the Krishna River and its tributary the ,\1alrahha, l'imanas
are characterized by the predominance of their vertical over their horizontal lines,

I .Iouneau-Dubreuil. "Archeologie du Sud de rInde. I Architecture" Ann Mus Guimet XXVI (paris. (1914)
appentidx. pages 172-182. especially pagc 179. .
2 Gravely and Ramachandran. "The Threc Main Styles of Temples An:hitecture Recognized by the Silpa-
Sastras."Bull. Madras Govt. Mus (N.S.Geni Sec.)1II part I (Madras. 1934). page 23.
3 See. for instanee.:the quotation of page 15 below.
4 In my joint paper with Mr. Ramachandran(loc.Cit). we considered this to be the sikhara. As. however. Prof
Pisharoti and Mr. Balakrishan Nayar have since pointed out to me that this involves difficulties in the
interpretation of (:crtain texts relating to \ imanas it seems best to avoid it at present. In any case the term
sikhara relate. to a structure forming part of the tower (Vimana) in all the three styles oftemple-nagara. vesara
and dravida recogrnzed by the ~ilpa-Sastras. and cannot rightly be used as a distinetive designation for the tower
characteristic of the northern torm. as it has been by some authors.
: for reference see Gravely and Ramachandran. loc. cit. page 10. footnote 3.
 All Outline ~lI11llian Temple Architecture 3

and by the resemblance of their crown to an amalaka' fruit the star-gooseberry

or myrobalan (genus Philanthus), a motif that is often repeated at regular intervals
all the way up, especially on the comer elements (fig. I ). South of the Krishna several
distinct forms of temple are found, but one of them which overlaps the area of the
Northem from of temple in the basin of the Malprabha-is so much more widely distributed
than any other that it has come to be regarded as the form typical of the

Fig.l.
Temple of Northern form {from Gravely and Ramachandran. after Cousens; final restored.}

Southem part of the Peninsula 2• In the form of the temple the vima1la consists of a series of
sUl'sessively smaller and smaller tiers of n.iniature pavilions. horizontal lines thus
predominating over,

I Concerning Amalaka see Pisharoti'. Calcutta Oriental Journal. I. pages 189-196

2 On page 13 of our paper already mentioned. Mr. Ramachandran and I call attention to a temple, the
ruin') of which are recorded by Fuhrer as found at Ramnagar in the Bareli district of the United
Provinces, which suggested to us that temples of stores diminishing in size from below upwards. each
surrounded by cells. were at onetime in use in northern India-and that much earlier than those known
in the south. Having subsequently learned of the unreliability of many of Fuehrer recorded. we wrote
to M . K N. Dikshit of the Archaeological Survey for further particulars. He tells us that "None of
the photographs and drawings relating to this temple have been published, not has any other
archa~')logist visited the place during the last 43 years .... The diminishing stories which you infer
from the description seem to me quite impossible, inasmuch as the total height of the mound was not
much and it is very likely that the so-called stories are but buildings connected with successive
periods of occupation. "There is therefore no evidence that temples of Southern form ever occurred in
the north, and in the absence of such evidence there is no reason to suppose it.
4 Bulletin. Madras Government Afuseum

Vertical, capped by a cupola-like crown (figs 2 and 3). S\.ich appear to be most easily
defined characteristic of the two principle forms of Indian temple. which, from their
geographical distribution, may conveniently be termed the Northern, and Southern forms
respectively.

Northern form of temple has undergone considerable modification in different

districts. and no doubt also at different times. and may not improbably prove to be
characteristic of a number of distinct styles. all presumably (though not necessarily) with a
common origin. Unfortunately my knowledge of such temples is nQt sufficiently intimate to
permit of the expression of any opinion either on the vexed question of the origin of this
form or on the apparently almost untouched question of its subsequent evaluation '. It was
called Northern or Indo-Aryan by Ferguson and of these terms the former is geographically
corred and. unlike the latter, is free from further implications. Because this fonn of temple
is northern some authors with the Nagara style of the Silpa-Sastras have also identified it
but the correctness of this identification is very doubtful ~see Gravely and Ramachandran.
loc. cit.).

The southern form of temple has been regarded by Fergusson and most
subsequent authors as associated with a single style. to which the name Dravidian has
generally been .lpplied. Actually, however. two different styles have in this way
been confused. styles which were already differentiated diverged still more
. widely, following different lines of evolution as regards both external form and
decorative detail (see Gravely and Ramachandran. loc. cit.. pages 23-25, etc). One of these
style (fig.3, of the Virupaksha Temple at Patadakal) was developed side by side with
temples of Northern form the Kanarese country, to which it seems to be confined: the
other (fig.2 of the shore Temple at Mahabalipuram) belongs with equal definiteness of the
Tamil country. though from about the time of the Vijayanagar Empire onwards it
spread over a much wider area. The latter has been studied in detail by Jouvenau-
Dubreuil in this "Archeologie du Sud de I'inde" (Paris. 1914) and in his much
shorter "Dravidian Architecture" (Madras. 1917), where he so restricts the meaning.

I Nirmal Kumar Bose in :.is "Canons of Oris san Architecture" Calcutta. 1932) has ,tudied lh.: lemples of Orissa
in the light of local manw.rripts. written comparatively reccntl~. of which he considers th;" .. It is certain Ihal the
tradition recorded in them has been handed do\\ n from very ancient times" (page 5). Four main types are
recognized in these manuscripts. the Rekha. Bhadra. Khakhara and GOlld(l'a (pages 78-80 and pI. opposite pages
20 and 40) and each type is subdivided into four classes (page 81). The (joudiya typc. \\hich as its name implies
is identical with the Bengal leaf-hut /orm of temple (see bellm. pagc II. is a recent introduction of which only
two examples are known (page 78). the Rekha type is vimana tower of Northern /orm. The Bhadra typc and the
Rekha" are joined to each othcr in a very intimate manner... the format tJ\Vcr ofl\iorthern from the Bluu/ra typc
and the Rekha" are joined to each other in a very intimate manner:' the limns /i)rnling the .\1uhAsa/a (page 154)
or vestibule to the lane which is the typical form of vim alia ill Orissa. In the famous temple at Konarak.
however. the l'imana appears to be of the Bhadra type. which bearers an obvious resemblance to the \'imana
}·:adamba temples (see below. page 9). The main l'haracteristic of the Khakhara type seem~ to be its rectangular
instead of square plan. In forn). though not in use. it thus bears much the same sort of relation to the t:pical
Orissan "imana as the Tamilian gopura does to the Tamilian \·imana. It s end walls are sometimes straight (pI.
opposite page 33) as in the Tamilian gopura, sometimes curved (pI. opposite page 37). When the northern form
of temple has been similarly studied in other areas it is not unlikely that it may revcal a similar diversity amount
its buildings elsewhere.
Several early temples of Northern form are considered in eh. VIII "Temple' and Architecture" of R.D.
Benerji's "Eastern Indian School of Mediaeval Sculpture" (Arch. Sun, Ind., Nc\\ Imp. Ser. XL VIII. Delhi.
1933).
 An Out/ine (?f Indian Temple Architecture 5

of the term Dravidian-as applied to architecture-as to confine it to this latter style alone. As
the \'ima110S of this style seems to fit the Silpa-Sastra definition of the Drm'ida style some
form of this word may perhaps be retained for it-in which case. in view of the linguistic and
cultural significance of the word Dravidian. misunderstanding will probably be minimized if
Drarida is used in connection with architecture. As. however. there is still much difference
of opinion as to the correct interpretation of the Silpa - Saslra definitions of architectural
styks. and as it may further perhaps be argued that the term Dravida. if it does apply should
continue to be taken to include both styles of the Southern form of temple, I propose in this
paper to use the term Tamilian instead.

Fig.2
Temple of Southern form. Tamilian type (from Gravely and Ramachandran).

The ramilan style differs from that of the Kanarese type of Southern temple chiefly in its
decorative details (see Jouveaue -Dubreuil. "Archelogie de Shoud de I'inde" I, pages 171-
182): but there are distinctive differences between the two in external forms as well. For
instance, though the crown of the viman(l is octagonal in the earliest surviving temples of
both, it has a window ornament on each face in the Tamilan style but one on alternate faces
in the Kanarse (Gravely and Ramachandran, lac cit., pI. It fig. 1): and whereas it mostly
retains this original octagonal shape ,~ later Tamilian

206-5-2
6 Bulletin. A1adras Government Museum

temples it is square in all later Kanarese one (Ioc. cit.. pUI, figs2-4) except the
Mallikat:iuna at Pattadakal in which it is round. In the latter style, moreover. a decorative
gable is added to the front of the spire at the comparatively early date (Ioc.cit..pl.ll,fits.3-4).
a feature not found in Tamilan temples. I And there are other differences also, conceming
which see Gravely and Ramachandran. loc. it.. page 16. As these Kanarese temples were
erected by the Chalukyas. 1 of which dynasty they are the earliest surviving monuments. they
may conveniently be designated Early Chalukyan. TJ1is term is not only historically correct
for the earliest Chalukya buildings that survive. but will further help to avoid their confusion
with later temples derived from the (see below. pages 15-18) to which the general term
chalukyan was applied-in a geographical and not a dynastic sense" -by Fergusson. though
the most celebrated of them are now known to have been the work of a different dynasty.
the Hoysalas. Whether Early Chalukyan architecture should be so defined as to include
temples of

I A somewhat similar structure is found in the Earl! Chola temples at Tanjore and Gangaikondacolapuram. but

quite 10\\ in proportion of the immense height of the vimanas. The Principal vimana of the temple at Sri
Rangam seems at first sight to be another exception. But the resemblance is clearly superficial. I know of no
means by which ib date can be fixed. and the site is of course a very ancient one. Its window ornaments prove.
howe\t:r. tllat in its present condition the building must be much less ancient. Its unusual fonn has apparentl~
heen arrived at by the \\ idening. to pennit of it containing a ~ufliciently large image of the recliningg
Ranganatha. of the inner end of the apsidal fonn of shrine mentioned on page 12 bellm.

: On page 62-63 of his "Chalukyan Architecture of the Kanarese districts" (Arch. Surv. Ind. New Imp. Scr. XIII.
Calcutta 1926) Cousens points to the remarkable similarity existing between the Virupaksha temple of one of the
queens of the Chalukyan King Vikramditya II at Pattadakal and the Kailasa temple of the conqueror of the
Chalukyas (or his son). Krishna I of the Rashtrakuta d~ nast). at Elura (see Fleet's "Dynasties of the Kanerese
districts ... '-· in "Gazt:tteer of the Bombay Presidency". 1896. page 39). Though the decorative details of the
E1ura templs. like that of the Pattadakal ones. has been shown by loveau-Dubreuil ("Arch logic do Sud de ]"]nde"
I. pagt:s 171-182. pl.ivil-Ixiv) to in very different 'style from that of the temples of the Tamil country the [Iura
temples difler from typical Chalukyan ones and resemble those of the Tamil Country in their use of the
octagonal form of crown long aftt:r it had been abandoned by the Chalukyas in favour of the square form. and in
the presence of small gopuras-very likt: thost: of the Kailasantha temple at Conjeeveram-as well as in "the
number of repetitions of the lion" commented on by Cousins. In my joint paper with Mr. Ramachandran alrt:ady
from the .) dmil country we were surprised to be unable to find an)thing distinctively Tamilan about it. I now see
that in this temple and in the Mallikar:iuna Irom the same reign. alone so far as I know among Chalukyan
temples, there are small gopuras. These are. however situated. as they are also in the Kailasa tempk at Elura
over the temple doorways behind the porches (see Burgees The Ancient Momuments. Temples and Sculptures of
India II. Mediaeval Monuments". pI. 277 and Hurlimann "Picturesque India". pI. 100) instead of ovcr the
gateways of the one or more outer courts as in Tamilan temples.

Mr. Ramachandran has called my attention to the fact that direct Tamilian influence in the Elura
temples is quite possible in view of the recorded alliance. at about the time when they were made. between the
Rashtrakutas and certain members of the Pallava royal family against their common foe the Chalukyas. (See
Anant Sadashiv Altekar. "The Rashtraakutas and Their Times". Poona Oriental Series. No.36-38: also
Ramachandran. loum. & niv. Bombay. History. Economics and Sociology. 1 page 240-41). I have to thank
Mr. Ramachandran for his ready help in all matters about which I had occa.,ion to consult him in connection with
this paper.

'See page 41 of his "Architectural Memoir" in "Architecture in Dharwar and Mysore" photographed
by Dr Pigou. A.C.B. Neill and Col Meadows Taylor. with an Historical and Descriptive Memoir by Col.
Meadows Taylor and Architectural Notes by .lames Ferguson (London, \866).
 An Outline olIl1dial1 Temple Architecture

Northem form built by the same king. or whether these latter temples resembled other
Northem temples in their decorative detail as well as in their outward form. and thus are not
to be regarded as belonging to any distinctively Chalukyan style. are unfortunately questions
that cannot be answered until the characteristics and development of decorative detail have
been more fully worked out for both forms of temple in the Chalukyan country. and for the
Nonhern forin generally. Nor is it possible to say whether Chalukyan and Hoysala can most
conveniently be regarded as distinct styles or as t\\-o periods of a single style.

The Northern form of temple apparently dominates the whole of the extensive
area of which it is characteristic; but the South em form seems to have been confined, till the
rise of the Vijayanagar Empire, to the Chalukyan and Tamil Kingdom. Except for a fev,'
cave temples and the early apsidal temple at Chezarla. partially nothing seems yet to be
known of the archaeology of the Telugu country subsequent to the decline of the Buddhist
centers in the lower pan of the Kistna basin 1: and

Fig.3.
Temple of southern form. Early Chalukyan style (after Cousens). The gable in front of the spire is
absent in the earliest examples.

such scanty evidence as I have seen points to a distinctive style existing there. History
suggests the likelihood of its being derived from Early Chalukyan temples and this seems to
be supported by sculpture. but I know of no published evidence one way or the other. In the
west. ho\vever. two styles are found which appear to be independent in origin from any of
those described above. one characteristic of the al1l;ient Kadambd ktngdom of the Kanarese
country. the other of the Malabar Coast.

1 Concerning the small square shrines of the nonhern part of this country see below. page 12.

206-5-2a
 8 Bulletin..4fadras Government Museum

In a Kadamba temple (fig.4) the vimana is square in plan. with its tower pyramidal
and composed of a series of horizontal step-like stages. each of ,,,hich is usually decorated
marginally with a uniform series of quadrangular vertical projections-a kind of roof. it
must be noted. which is also found covering vestibules attached to vimanas of Northem
foml in Orissa. at Bhunaveswar and elsewhere.' The stages are more numerous and less
elevated than in temples of Southern form. and are devoid of pavilion omamentation. The
cr?wn. when present is square.

In a West Coast or Malabar temple (fig.5) the walls resemble a wooden railing in
structure and are as a rule ~ti" made of wood to-day. though stone copies from about the
time of the Vijayanagar Empire also exist (see Cousens "Chalukyan Architecture:' pages
134-137. PI. cxliii. cxliv and cxlviii).

FigA
Kadamba type of temple (from Gravely and Ramachandran). An early example before the
addition of the gable in front of the spire. Crown also absent.

Such temples may have either a simple pitched roof of overlapping slabs. or they may have
a series of pitched roofs one above another (fig 5). which bear an obvious resemblance to
the multiple, pitched roofs of Chinese and Napalese temples (fig.6). In the first edition of
his "History of Indian and Eastern Architecture" (1899, page 308--omitted in the
1910 edition revised by Burgess) Fergusson refers to the well known anthropological
similarity between the Nayars of Malabar and the Newars of Nepal. especially as regards
their custom of polyandry. If the similarity between these

I Concerning small square vimanas with similar spires in Ganjam see below. page 12 see also abov~,
page 4 footnote concerning the large vimana of this type at Konark.
 An Outline qlindian Temple Architecture 9

two nm.... widely separated peoples should prove to be due to their both representing an
ancient stock, formerly of ,more general distribution. which has been split into two
and driven to shelter in the mountains of the north and behind those of the south-west.
these multiple roofed temples would'presumably represent an older style of architecture than
the Northern and Southem and now occupying the wide area behveen them. But it
must not be forgotten. on the other hand. that the West Coast has from very early
times been in direct contact with China by sea. a fact to which the Chinese fishing nets of
the Cochin and Travancore backwaters still bear testimony. Temples that seem to show a
more or less definite affinity to those of Malabar do. however, exist here and there in other
part of the Peninsula also. The Nataraja shrine of the Chindambaram temple, for instance,
has wooden wall of peculiar construction somewhat reminiscent of those of a Mal£bar
temple. though the shape of the building as a whole, with its golden roof,
is clearly connected, through the Durga temple (the so-called Draupadi. Ratha) at
Mahabalipuram. \..-ith that of square thatched huts illustrated in Buddhist sculptures
from Amaravati. The temple in true Malabar style at

Fig. 5.
Malabar type of temple (after Fergusson).

Harpanahalle, BellaI)' district, shown in PI. Ix.xxiv and ixxxv of Meadows Taylor
and Fergusson's "Architecture in Dharwar and Mysore" is probably due to some direct
connection between those who built it and the Kanarese West Cost; but I know of no
confirmatory evidence of this. nor whether any other such temples occur in the Deccan.

In the Kashmir valley of. the Westem Architecture in Dharwar and Mysore" is
probably due to some direct connection between those who built it and the Kanarese West
Coast; but I know of no· confirmatory evidence of this. nor whether any other such temples
occur in the Deccan.

Himalayas multiple roofs are found in yet another style of temple (fig.7). According
to Furgusson these temples bear two or three roofs "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 veranda. separated from the
10 Bulletin..Madras Gowrnment Museum

centre either by walls or IT!erely by a range of pillars. In the wooden example the interval
between the two roofs seems of have been left open for light and air; in the stone buildings it
is closed with omaments. Besides this, however. all these roofs are relieved by dormer
windows. of a pattem very similar to those found in mediaeval buildings in Europe: and the
same steep, sloping lines are used also to cover doorways and porches, these being virtuall)
a section of the main roof itself, and evidently a copy of the same wooden construction: the
pillars which support the porticoes ... are by far the most striking peculiarity of this style.
their shafts being so distinctly like those of the Grecian Doric. and unlike anything of the
class found in other part of India. I

Fig. 6.
Nepal type of temple (after Fergusson).

Lastly, mention bust be made of the form of temple illustrated in fig.8 this. as has been
pointed out by Gangoly on pages 23-24 of his "Indian Architecture" (Little Books on
Asiatic Art, VoI.III), is "~vidently borrowed from leaf-huts very common in Bengal;' the
region to which it belongs. In this fonns of temple with curved eaves we also find the same
tendency of a multiplication of roofs one above another.

I See also chapter iii "Architectural Styles" in Ram Chandra Kak's "Ancient Monuments of Kashmir"
(India Society, London, \933).
 An Outline of Indian Temple Architecture Il

Multiplication of roofs is thus a feature of the different forms of temple typical respectively
of Malabar. Bengal. and the Eastern and Western Himalayas. Should any or all of these
four styles be grouped together as belonging to single from? The answer to this must depend'
on whether or not they can be shown to have a common origin. independent of the Northern
and Southern form of temple characteristic of the rest of the country. At present this cannot
be donc. and to attempt any such grouping '.,ould be to prejudge the issue. But the different
kinds of Indian templc knO\vn to me-there may of course be other that have escaped my
notice-all either belong or seems to have been derived from the above mentionea
fundamt:ntal types. which are as follows; temples of Northern.

Fig.7. Fig.7. Kashmir type of temple (after Fergusson) Fig.8

Fig. 8. Bengal leaf-hut type of temple (after Gangoly).

from. whether comprising one or several styles it is not yet possible to say: temples of
Southern form in two distinct styles. the Early Chalukyan and the Tamilian: and temples in
five apparently independent styles of more restricted range. namely those of the ancient
Kadamba kingdom. Malabar. Bengal. Nepal and Kashmir. To this list of prototypes must be
added apsidal temples. based on the Buddhist chaitya hall. and square kmples. based on a
square thatched hut with simple roof. and straight caves. which occur sporadically in the
style of their locality as regard detail. in various place such as Mahabalipuram (the so-called
Sahadeva and Draupadi Rathas respecstively). Chezarla. Aihole (the so-called Dl'rga) and
Chidambaram (the Nataraja shrine).

Small square plain shrines base on a square but with simple roof and straight caves occur
in many places and should perhaps be regarded as an independent type. though when decoration is
12 Bulletin. Madras G(wernme111 Museum

added it naturally follows the local style. These temples are, however. so insignificant that
little or attention seems yet to have been paid to them. When traveling recently by train from
Calcutta to Madras. I noticed that such temples seemed to be the principal. if not the only.
form of temple in Ganjam and the northern part of the Telugu country. In Gal\iam the
pyramidal roof is marked by a series of horizontal courses which. in at least one example a
little to the south of the Chilka Lake. have the form of low step-like stages decorated with
quadrangular vertical pr~iection as in the vestibules of some of the temples of the adjoining
province or Orissa and in the shrines of Kadamba temples. The crown. however. is unlike
that of either being. so far as I could observe. round but neither clearly separated from the
pyramidal spire by a distinct neck nor ribbed like an amalaka. In some. crown and finial
seem to be merged into one another to form an octagonal column. Further south all trace of .
the crown disappears and the horizontal courses becomes less marked and less numerous till
near Vizagapatnam they are often entirely absent. Almost immediately to the south of
Vizagapatnam, however, horizontal courses reappear in many such shrines in the form of
deep steps. suggestive of terraced spire of the Southern form of temple. though without its
decoration. In the Madura district also, round about Kodaikanal Road station. there are
many small squares shrines. built the differ from those just described in the their spire is
curvilinear and is much smaller at its base than is the body of the shrine from which it rises.
It may be either square or circular in plan, in the latter case being more or less hemispherical
l'lstead of pyramidal: and sometimes ornamentation of Tamilian type is added. Evidently
these little shrines differ considerably in form one part of the country to another and deserve
more attention than their simplicity would seem to indicate

It is unlikely that any of the above mentioned styles. except perhaps the last if it is
to be regarded as distinct, have remained unchanged throughout the ages during which they
have been in existence: and the more complex of the temples of Bengal. such as the
Dakshinesvara temple. near Calcultta. figured on pI. 126 of Hurlimann's "Picturesque
India". have evident Iv arisen fornl a combination of the Northern form of ·ov.er with the
Bengal leaf-hut fonn- of temple,1 two prototype which belong to contiguous areas. More
investigation seems. however, to be needed before the evolution of any of the styles found
north of the Krishna can be made clear: and this seems to be the position with regard to the
Malabar style also.

The evolution ofthe Tamilian style has been worked out by Jouveau-Duuuil. whose
results are embodied in the two books mentioned on pages 5-6 abO\e. He
divided the time from the making of its earliest surviving examples by the Pallava 'king
Mahendravarman l. early in the seventh century A.D., to the commencement of its
Modern period about 1600 A.D., into four periods of approximately 250 years each.
The first is the period of Palava temples (fig. 2, P.5) ending with the

I Concerning other forms taken by temples in Bengal see N.R. Bhattasali "lconography of Buddhist
and Brahmanical sculptures in the Dacca Museum" (Dacca 1929), pp.xii-xviLpl.1 xxxi-Ixxxxii. ""8 An
idea of the temples that were erected over these images" and S.K. Saravati "The Begunia Group of
temples of Bengal" Journ .. Inc. Doc. Oriental Art. II (Calcutta. 1934). pages 130-140) pl.xlv.3 text-
figs, See also "Bengali Temples and their General Characteristics" by Manmohan Chakarvarti Journ.
Asiatic Soc. Bengal, N.S.V. 1909, pp.141-162. 13 figs.
 An Olltline of Indian Temple Archite(·ture 13

Final su~iuction of the Pallava dynasty of Conjeevaram by the Cholas fonn further south in
about 850 A.D. It seems likely that in their original kingdom the Chola already practiced an
allied-but in certain respects distinct-style of the own. presumably in perishable material:
for the with their supremacy marked changes in certain details make their appearance
somewhat abruptly in surviving buildings. The supremacy of the Cholas lasted for about
500 years: and throughout that time a gradual evolution took place, resulting in sllch great
ditTerences between the buildings of its beginning and end that it is convenient to divide in
into to Early and a later Chola period of 250 years each. This gradual evolution
continued though the 250-year period of the Vijayanagar Empire by which the Later Chola
period was followed. and so on into the Modern period.

The Pallavas made cave. freestanding monolithic ane! structural temples. never of
very large size. and either with gopuras or with these smaller than the vimana. Their largest
surviving temple is the Kailasanatha at Conjeevaram, but most of their finest sculpture is at
Mahabalipuram. The Early Cholas built largest temples in which the Vimana still
dominated the gopuras. being is some instances indeed, as at Tanjore, of immense size.' The
Later Cholas seem probably to have

Fig. 9.
Evolution of carbel. Tamilian style (slightly modified from Jouveau-Dubreuil).
A. Pallava B. Early Chola. C. Later Chola. D. ViJayanagar. E. Modem. Transitional forms develop in
the later parts of all these periods except the first.

enlarged existing temples, the shrines of which were considered too sacred to touch, in
preference to constructing new ones: for their vimanas are as a rule insignificant and their
gopuras immense as, for instance of Tiruvannamalai and Chidambaram. And the tradition
thus established has remained in fashion ever since. The Vijayanagar Kings, though they
continued the building of immense gopuras, concentrated attention on the mandapas or
pillared halls, marvelously developing the carving of their many rows of monolithic pillars,
as may be well seen at Conjeevaram and Vellore. Onl)' with the coming of the Modern
period in about 1600 A.D. does the custom seem to have been commenced of faking
mandapas in the more convenient from the spacious halls or corridors, such as are found at
Madura and Rameswaram, as well as in many earl·ier temples where they form addition for
reconstructions. to enable large numbers of people to obtain an unobstructed view of any
ceremonies performed. The successive periods thus differentiated are equally
well characterized by successive developments in decorative motifs, the pillars, corbels (see

I I For an account of the Early Chola temples of Koranganathan at Srinivasanur. Tiruchirappalli

district and at Gangaikondacolapuram. see Percy Brown "Two Cola Temples" Journ. Ind. Soc.
Oriental Art. II (Calcutta. 1934) pages 2-6. pl.ii-vi I text-fig.

206-5-3
14 Bullelill. Madras GOl'ernmenl Aluseum

fig.9). niches, etc. all showing a gradually increasing degree of elaboration. though many of
the earlier and simple fonns, being comparatively cheap and easy of execution. have
persisted and are still in use today. Details will be found in .Iouveau-Dubreuirs tWl) books
already mentioned.

The course of evolution. especially of decorative motif. in the Kadamba and Early
Chalukyan styles is as yet much less clearly understood: but in the former style it has been
studied in ch. viii "Architecture" of pt. viii "Internal History" of Moraes' "The Kadama
Kula. a History of Ancient and Mediaeval Karnataka" (Bombay. 1931) and the broad
outlines ofthe obviously parallel development in external form the these two geographically
contiguous style have been indicated in the paper by Mr. Ramachandran and myself already
referred to. The Kadamba type of vimana. (fig 4.p.8) seems to have developed from a flat-
roofed temples by the addition of a series of succt!ssively smaller and smaller horizontal.
roof like stages. a square cupola-like crown being often. but in early times at least not
invariably, inserted between the pyramid thus formed and the finial. The builders of the
Early Chalukyan temples. as already mentioned above. used both the Northern and the
Southern (fig3p.6) form of tower. For over a century both forms of temple seem to have
been built in one and same locality. each without apparent effect upon the other. But
whether they differed from one another in decorative details as well as in form. and if so
whether and to v.'hat extent they influenced one another in respect of it. never seems to have
been investigated and can only be decided by further field study of the buildings. none of
which unfortunately I have ever seen.

There can be little doubt that the Southern f<.mn of ,imana has been derived from
pyramidal Buddhist monstery of vihara of successively smaller and smaller stories. each
consisting of a central mandapa surrounded by monks' cells: for a Fergussoon has pointed
out {loc.cit., 1899 edition. pages 135-136. fig. 66-67: 19 10 edition. I, pages 171 -173. figs.
89-90) it appears to be practically a reproduction in miniature of such a \ihara. This use of
tiers of pilaster and miniature cells as a mode of decoration had. moreover. probably begun
to applied to buildings of various shapes-elongate. whether rectangular or apsidal. as well as
square-at some period antecedent to that of the earliest surviving examplt!s. among which all
these shapes are already included at Mahabalipuram.

The origin of the Northern of vimana is much more obscure. Ferugusson (Ioc.cit ..
1899 edition. pages 438-439: 1910 edition. It page 119) with reference to the tower of the
Huchchimalligudi temple. at Aihole says "It will also be observed in this tov.er that eve~
third course has on the angle a form which has been described as an alii a/aka in speaking of
the crowing members of northern temples. Here it looks as if the two intermediate courses
simulated roofs, or a roof in two stories. and then this crowing member was introdm:ed. and
the same thing repeated over and over again till the requisite heights \,vas obtained:' This
does not, hO\vever. account for another characteristic of the Northern form of tower. namely.
the setting forward of the middle portion of each face. which seems to be an equally marked
feature particularly of simple and presumably early examples. Other investigators have.
therefore, suggested its evolution from a slupa raised. as was often the case. on a series of
terraces, each of them set forward in the middle of each side to accommodate flights of steps
fonn the one below. the whole being surmounted by a series of umbrella one above another.
of which the lowest has become enlarged and developed into the Gma/aka-a-mode of
evolution that receives.
 An Outline o(lndian Temp/e Architecture 15

strong from the series of votive stupas of various forms that have been set up round the
Mahabodhi temple at Buddh Gaya and elsevv'here.

Whatever is origin. the Northern form of tower. found side by side with Early
Chalukyan examples of the Southern form. is taller than the latter in proportion to its
breadth and. though its spire is composed ofa series of horizontal courses. the lines of these
are subordinate to the stronger vertical lines resulting from the setting fOf\\/ard of the middle
pOltion of each face: while the Southern and Kadamba form of tower both have pyramidal a
spire consisting of a series of strongly marked horizontal courses. which in the former are
deeper and less numerous than in the latter and decorated with miniature pavilion.

In the t".,o or three earliest surviving Early Chalukyan temples of Southern from
(Gravely and Ramachandran. loc. cit.. pl.ii. fig. I ) the crown is octagonal as in the Tamilian
style from which. however. as I have already explained elsewhere (Ioc.cit.. pagel6) these
temples differ in other features \vhich place them at the base of the Early Chalukyan series
the crown is square (Ioc.cit.. pl.ii fig.2). except in the MillikaJ;.iuna at Pattadakal. which
seems to be unique among temples. w'ith cupola-like crown in hanging it round. In
Kadamba temples the crown is not always present at any rate among those of simple and
presumably early construction: but when there is one it is always square. In both the Early
Chdukyan and Kadamba series the next development is the addition of a projecting gable of
the tront of the spire (Ioc.cit.pl.ii.fig.3) a form of ornament never found in the Tamilian
style.! In Chalukyan temples it is first seen in those of MillikaJ;.juna and Virupaksha (see fig
3.page 6 above). created by the queens of Vikramaditya II (733-746 A.D.) and it presumably
appeared in Kadamba temples at about the same time.

The first attempt to combine into a single Chalukya building elements from both the
Northern and Southern form of temple probably occurs in the Papanatha temple at
PattadakaL which is believed to be if an)'thing a little earlier than the Virupaksha. But this is
essentially a temple of Northern from to the body which miniature pavilions from the
southern from have been added. and on these lines no further developments seem to have
followed.

At Kukanur we find. however. in the Navalinga and Kallesvara temples, (PL.I.A) a

type of combination that proved to have much greater possibilities. In these temples the
tower is essentially of Southern form. but it has the middle portion of each face set forward
as in the Northern fornl. the strong vertical lines of which thus come to be superimposed
upon the strong horizontal lines of the Southern fonn: The Navalinga temple departs
comparatively little fonn the typical Southern form. but the Kallesvara shows a more
extensive incorporation of Northern elements, and its miniature pavilions are already
subordinated to the lines of the courses from which they spring.

Neither of these temples seems to bear any inscription giving a clue to its date,
nor do dates seem as yet to be determinable for the stages by which the composite style
that they initiated developed into the style of the Hoysala dynasty. What
those stage probably were may. however, be briefly indicated by reference to two or three

! Concerning the curious shrine of the Srirangam temple see above. pate 7. and footnote I.

206-5-4
16 Bulletin. ,\1adras Government Museum

temples selected for the purpose. Whether such transitional temples should be called Later
Chalukyan or Chalukya-Hoysala Transitional. or whether -as I am inclined to suspect-they
belong to two successive periods, of which the first characterized by comparative restraint
and the second by a profusion of external decoration, is not yet clear. The large Jain temple·
at Lakkundi figured by Cousens in pl.lxi of his ., Chalukyan Architecture'" affords a good
example a little more advanced than the kallesvara temple at Kukkanur but with the
decoration under restraint, while the temples of Tarakesvara at Hangal (Cousens,
10c.cit.,p!.1 xxxvi) and of Santesvara at Tilivalli (Cousens, 10c,ciLp!.c.)may be taken as
representatives of the more unrestrained transitional type of building.

In the two temples last mentioned the shrines have begun to lose the square plan of
an Early Chalukyan shrine and to assume the star-shaped plan of a typical Hoyysala one.
The crown is reported by Cousens to be modern in the Tarakesvara temple, but that of the
Santesvara shown unmistakable evidence of its derivation from a square crown of southern
cupola-like shape: while up the middle of each side of the spire of both there is a series of
plaques suggestive of the Northern form of tower. Cousens (Ioc.cit.page92) says, moreover,
of the former temple "The little sikkharas on the shrine walls are of the Chalukyan type
while those round the low wall of the open hall are of the northern,"

The tower of the Mahadeva temple at Ittagi (cousens,)6c.cit.,p!.ci) and a sculpurual

miniature of a very similar tower (pUI) form over a niche in a temple of the Bellary district
and now preserved in the Madras Museum, show a slightly different combination of
elements from the Northern and Southern forms of temple, a combination in which the ends
of pavilions from the latter are clearly recognizable projecting on either side from behind the
central series of decorative plaques from the former.

As in the Virupaksha (fig.3, p.6) and Mallikar:iuna at Pattadakal, among Early

chalukyan temples, the hall in front of the shrine of the Tarakesvara temple at Hangal (pU
B) has an entrance on either side, each with a large porch, in addition to the median entrance
at the eastern end. This median entrance is one corresponding to the single entrance which
alone is found in Early Chalukyan temples other than the two just mentioned; but in the
Tarakesvara temple instead of opening direct to the exterior, it is connected by a covered
way with a large star-shaped Mandapa, and this covered way has each side a small flight of
steps forming additional entrances. The plan of the Mahadeva temple at Ittagi is on similar
lines. Cousen gives no plan of the Santesvara temple at Tilivalli, but mentiones that it has
three porches.

In the Kesava temple built in 1268 A.D. at Somnathpur (Narasimhachar ''The

Kesava Temple at Somnathpur" Mysore Arch. Ser., Architecture and Sculpture in Mysore.
I, ;I.i) which may be taken as an example of the fully developed style of the Hoysala period,
development already seen in the temples just described have been carried further, the
porches to Northern and South of the hall are replaced by vimanas of the same size and
form as the median one, I all of them being definitely star-shaped(p 1.1.0) and the whole
temple is raised on a platform.

iVimanas grouped round a common halis (or series of halis) are found in the Navalinga
temple at Kukkanur, which forms the commencement of the transitional series (see above, page 15).
Cousens says, however (Ioc. cit.page 74), "It needs but a glance at the plan to see that this collection
of shrines and halls is not the result of a premeditated design or it would have been more
symmetrical'".
 All Outline of Indian Temple Architecture ]7

The Hoysala sty Ie must thus be regarded as the final outcome of the un ion of the
two forms of temple adopted by the Chalukyan dynasty: but it seems to have absorbed at
Kandamba style also. As has already been pointed out above, the evolution of the Kadamba
and Early Chalukyan temples followed parallel courses. As the predominating lines of the
Kadamba spire were horizontal. the superimposing of the vertical lines of the Northern form
of tower upon it would have much the same effect as it had in the cas~ of the spire of the
Southern form of the vimana. And in the Lakshmidevi temple at Dodda Gaddavalli, built-in
1113A.D (Narasimhachar "The Lakshmidevi Temple at Dodda Gaddavalli" Mysore Arch.
Sel".. Architecture and Sculpture in Mysore III) there are unmistakable signs of the use of the
Kadamba as well as of Chalukyan models.

This temple was built about a century and a half before the Kesava temple at
Somnathpur and the Shrines are still fundamentally square in plan (p I.I.C .j. nor is the
temple raised on a platform. It has, however. four vimanas. the fourth replacing the large
mandapa found at the eastern end of the Tarakesvara temple at Hangal. the only entrances
being. consequently, the pair immediately in front of this fourth shrine. Of these four
shrines that of Lakshmidevi (Narasimhachar. loc.cit.. pl. front of this fourth shrine. faces
east and has been selected for special ornamentation in a manner related to that of the
composite transitional vimanas of the Tarakesvara and Santesvara temples described above.
though its crmvn is of scarcely more advanced type than of the Kallesvara temple at
Kukkanur at the base of the transitional series. and less so than that of the Varaha-
Narasimha temple at Halsi in Kadamba style (Moraes. "The Kadamba Kula."fig ..32,
opposite page 289). The other Shrines of the main temple, and also apparently all the five
additional shrines in the temple enclosure have towers of the Kadamba type (Ioc.cit.p l.i and
ix), Though the Lakshmidevi shrine faces east. it is situated not at the end but on one side of
the end on one side of the central hall. the main axis of the temple extending north and south
instead of east west.

This completes the summary of the various types of building from \\hich temples of
different parts of India have been derived. so far as these are known to me. and for the
merging of there of these styles to produce the composite the Hoysala type of temple.
Underlying all of them it will be seen that there is at least one common feature-the
production of towers by a process of vertical repletion of roofs. in others a repetition of
terraces. But vertical repetition of some kind is present in all forms. though in the Northern
the resulting horizontal lines are dominated by stronger vertical ones. May not this deeply
rooted tendency to vertical repetition indicate some sort of fundamental unity underlying all
the various styles of Indian temple architecture?

206-5-4a
 PLATE I.

".,,'-',
-
..•
~ ..'

.....
.....

''''; \

•
J
. __~f

Plans illustrating the development of the Hoysala type of temple from the Early Chalukyan type--
A. Kallesvara Temple. Kukkanur (after Cousens).
B. Tarakesvara Temple. Hangul (after Cousens).
C. Lakshmidevi Temple. D~J(ida Goddavalli (after Narasimbacher)
D. Kesava Temple, Somnathpur (after Narasimhachar)

i'
BULL, MADRAS GOVT. MUS. (N.S.) G.S.III (2) PLATE II

DECORATIVE MINIATURE OF A VIMANA SHOWING COMBINATION OF NORTHERN AND SOURTERN

ELEMENTS, BELLARY DISTRICT
 INDEX
PAGE PAGE PAGE

temples ... 11-14.16 Lakshmidevi temple

A
EarlyChola period 13 Doddagodda\alli 17
Aihole 11.14
Ellura 6 Later chalukyan temples 15
Amalaka 3.12.14
9 F Latt"r chola puiod 13
Amaravati
Aspidal temples II Finial 2.12 Leaf - hut temples 10.12
B G \1
Begunia 12 Gable (decorative) 6 'Iadura 13
Brllar) District 7.H) Gangaikonda Madura District 12
Bengal temples 11-12 Colapuram 6.13 Mahabalipuram 4.9.11.
Bhadra temples 4 Ganjam 12 13.14
Bhuvanesvar 4 Gopura 2.4 Mahabodhi temple.
Body (ofa vimana) 2 Goudi)'a Temples 4 Buddhga)a 15
Buddh Gaya 15 Grecian Doric Pillars 10 \Iahade\a temple. Ittargi 16
Buddhist architecture I Gupta Period 2 '\Iahendra\arman I 12
Buddhist monaster) 14 B '\Ialabar temples ... 8·9.11-12
Buddhist sculptures 9 Balsi 16 ,\Iallikarjuna temple.
C Bangal 16.17 Pattadakal 6.15-16.
Cella 2 Barpanahalle 9 ,\Ialprabha rher 2
Chaityahall II Himalay'an Temples 9-12 ,\Iandapa 14.16
, Chaluk)a-Boysala Bindu architecture Modun period 13-14
transiritional Temples 15 Hoysala temples ... 6.7.16.17 \Ionastery( Buddhist) 14
Chalukyan Temples 6-7. Huchchimalligudi temple. '\I ukhsala 4

Chezarla
11-14.16
7.12
Aihole 14 'I)'robalan
, 3

Childambaram 13 Indo-Aryan temples 4 'agara temples 2.4

Chilka Lake 12 Ittargi 16 'ataraja shrine chidambaram ... 9.11
Chinese fishing nets 9 .J 'a\alinga temple. h:ukkanur ... 1:=;.17
Chinese temples 8 .Jain architecture 'ayars of \Ialabar 8
Chola temples 6.13 h: 'epalese Temples 8.10.12
Conjee\aram 6.13 h:adamba tempes 4.8.11. 'ewars of 'epal II
Crown (of a temple tower) 2~i-.12, 15.17 'orthern form of temple 4.7-9.
15.16 h:ailasa temple. [llura 6 11-15.17
Crowning member 2.14 h:ailasanatha temple. 0
D Conjeenram 6.13 Orissa 4.8.12
Dakshines\'ara h:alles\'ara temple. P
temple. Bengal 12 h:ukkanur 15-17 Palla\'a period 12
Deccan 9 h:anarese temples 5-9 Pallnas 6
Decorative motifs. h:ashmir temples 10.11 Papanatyha temple.
denlopment of 14 h:esava temple. Pattadakal 15
O')dua Godda\'alli 16 Somnathpur 16.17 Pattadakal 6.15.16
II 1'1(. 10 h:hakhara temples 4 Philanthus 3
II : ',' oadi Ratha. h:odaikanal road 12 R
1\1 h::balipuram 10.12 h:onarak 4 Ramesvaram 13
DI \' da temples 2.5 h:oranganatha temple. Ramnagar 3
DIJ g t, Aihole II Srinivasa nalur. 13 Ranganatha 6
Dravidian temples 4.5 Krishna I 6 Rashtrakutas 6
Durga Temple. h:rishna riwr 2 Rekha temples 4
Mahabalipuram 9 h:ukkanur 15-17 S
E 1. Sahadevaratha,
Early Chaluk)'an 6.7 Lakkundi 16 Mahabalipuram 11
 P.-\GE PAGE PAGE

Sanpesvara temple. T Vihara 14

Tilivalli 16 Tamilian temples 4-7.11- Vija~anagar period ... 4.7-8.12-
Shore temple. 12.15 13
'\lahabalipuram 5 Tanjore 6.13 Vikramaditya II 6.15
Shrines. small square II Tarllkesvara temple. Hangal ... 16-17 Vimana 2-4.6.8.
Silpa-Sastras 2.5 Telugu country 7.17 13-14
Somnathpur 16-17 TiIi\'ali 16 Virupaksha temple.
Southern form of temples ... 4.5.9.11. Tiru"annamalai 13 Pattadakal ... 5.6.15-16
13-15.17 Vizagapatam 12
Spire 2.12.15 V
Srini\'asanalur 13 Varaha-:\arasimha temple.
Srirangam 6 Halsi 17
Sta r-goosberr~ 3 Veil ore 13
Stupa 14 \iesara temples 2
Vestibule 4.8
 N.T.P. SUPPLIED BY THE DEPARTMENT
PRINTED AT GOVERNMENT CENTRAL PRESS, CHENNAI-600 079.

.~"~
c_ j
);-