The most characteristic feature of a class of temple architecture in India is its spire or shikhara which forms the termination of the upper portion of the body, or vimana, of the temple. In the temples of the so-called Aryan or Āryāvarta type the spire is more or less a curvilinear one with a corrugated stone at the top—known as the amalaka—surmounted in its turn by the kalasa or the water jar finial. In South Indian architecture the vimana is, however, a many-storied one and, instead of the kalasa or amphora, it ends in octagonal or domical structures. The spire lies just above the cella or garbha griha and, except in the case of Dravidian temples with their tall gopurams or gateways, it is the most elevated part of the temple and attracts attention from a distance. Dr. Coomaraswamy has truly remarked that in Āryāvarta style ‘the bulging spire with carved ribs rising above the shrine’ is ‘often repeated upon itself, as an architectural ornament.’

The origin of the shikhara is still shrouded in obscurity and we propose in this paper to consider the various theories promulgated by scholars, in regard to the shikhara and its significance, and to attempt a genetical account—so far as it’s possible with the materials available—with reference to the distribution and relative antiquity of the still existing archaeological remains.

Leaving aside the Buddhist cave temples where architecture or architectonics, properly speaking, can find no place—the oldest example of the shikhara is to be met at Bodh-Gaya—in the spire of the Mahabodhi temple (Fig. 1). According to the generdity of opinion, the original structure is said to date from the 1st century before the Christian era. Mr. H. Longhurst, however, believes that the Bodh-Gaya style cannot be dated earlier than the 11th or 12th century A.D. He says that the Burmese introduced this peculiar style of architecture into Bengal (sic) but it seems there’s little doubt that they originally borrowed it from Southern India.1 Although the Bodh-Gaya spire has had to undergo reconstruction at a later period, we can, from the representation of it as given in the Kumrahar plaque (Fig. 2), discovered by Dr. D. B. Spooner, clearly make out that it originally had no curvilinear sides and that the prominent features of Aryan or Āryāvarta class of spires were altogether wanting. Fergusson2 has observed with reference to the spire of the Bodh-Gaya temple that “the tower took a straight-lined course like the doorway at Missolonghi and the ‘Gates of Lions’ at Mycenae, while the Hindus took the more graceful curvilinear shape, which certainly was more common in remote classical antiquity and, as it is found at Persia, may have reached India at a remote period.” The Mid-Victorian archaeologists seem to have had a special penchant for theories relating to the alleged importation of architectural forms into India and, so far as the curvilinear shikhara is concerned, it seems to have been in a manner countenanced even by so keen-sighted an archaeologist as Raja Rajendra Lai Mitra.3 We shall discuss in its proper place a modern recrudescence of this theory but in order to be able to form a correct estimate of such views it is necessary that the earliest types of Indian spires should at first be taken into consideration.

Of the old Indian structural temples—those of the Gupta age may be mentioned as next in sequence to the Mahabodhi shrine, though there are few points of similarity between the two classes of structures. There is a clear gap of 3 or 4 intervening centuries which still remain unabridged. The Gupta temples, which are met with at Eran, Bilsar,4 Sanchi, Udaygiri, Tigowa, Deogarh and Nachna-Kuthara are built of stone and in the oldest specimens ‘the flat roof, the square form, and the stern simplicity all point to the rock-hewn cave as its prototype.’ As Cunningham observes in regard to the Gupta temples at Sanchi, there are in the neighboring hill of Udaygiri actual rock-hewn examples of this type.5 At Udaygiri moreover, there is a false cave temple on one of the sides of which has been built up—the roof being a natural ledge of the rock6. At Nachna-Kuthara (near Jaso in Central India) the Parvati temple which has a curious conventional imitation of rock carvings on all the outer faces of its walls—apparently in the fashion of  ‘old temples on the rock’—bears strong testimony to the persistence of cave features'.7 Mr. R. D. Banerjee, Superintendent of the Archaeological Survey, Western Circle, who visited the place at the request of the Editor of "Rupam", ascribes it to the early Gupta period i.e., to the 4th or 5th century A.D.8 There is however another spired Gupta temple at Nachna, that of the Chaturmukha Mahadeva which more prominently deserves our attention. At the time when General Cunningham visited the site, it had a tall spire with slightly curved sides nearly 40 ft. in height. Mr. Banerjee, who gives a representation of this temple in PI. XVII of his report, describes it as an earlier Gupta temple and remarks that the large four-faced lingam of the Mahadeva is certainly earlier than the temple itself.9 Cunningham, who tried to be more definite in his estimate of the age of the Chaturmukha temple, says that it must be considerably later than the other (the Parvati temple) and is probably not older than 600 to 700 A.D.10 This temple and the spired Gupta temple of Deogarh in Jhansi District appear to have been built on the same model and both stand on raised platforms.

Of the Deogarh temple, however, the spire—to quote from Cunningham’s Report— ‘is long in ruins though several specimens of amalaka fruit, which forms the special ornament of a Hindu spire, are lying about.’ Cunningham ascribes the same age to the Deogarh temple as that given to the spired temple at Nachna-Kuthara. His main reason appears to have been that in these temples—his fifth characteristic of Gupta temples viz. the flat roof— has given place to the spire which he regarded as one of the latest characteristics of the Gupta style. He observes in his connection, ‘…as some of the flat-roofed Gupta temples are certainly as late as A.D. 400, and others probably a century later, I think the Deogarh temple cannot well be placed earlier than A.D. 600 or later than A.D. 700.”11 The discovery however by Mr. Y. R. Gupte of an epigraphical record in characters of about the end of the 5th century, noting a gift on the part of one Govinda, the lord of Kesavapura12 serves substantially to disprove Cunningham’s views on this point. Another interesting find, a clay seal (Fig. 3.), among certain seals and old relics discovered at Parbati13 especially attracted the attention of General Cunningham as it bore the representation of a shikhara temple with a pennon floating from the top. While this particular seal bore no date, the other seals which were discovered at the spot were found to contain inscriptions in Gupta characters. It may therefore be surmised that the shikhara temple depicted on the seal must have been a fairly familiar type in the Gupta period and existed not later than the 5th century. Among the archaeological remains of Nagari, the ancient Madhyamika, which must have been in a flourishing condition from the 3rd century B.C. to 7th century A.D.14 is a quarry of the Gupta period, ‘exploited for the purpose of sculpture’. Dr. D. R. Bhandarkar discovered two unfinished amalaka pieces15 which certainly presume the existence of shikhara temples here, in about the 5th century A.D.—an epigraph of this period having been discovered in a different part of the locality.16 There are ruins of several Gupta temples also at Tigowa in Central Provinces, among which a flat-roofed one, built in a style similar to the cave-temples at Udaygiri and the structural temples at Eran, has been considered as the oldest by Cunningham, but there are certain other temples possessing ‘spire roofs covered with the usual pinnacle of amalaka fruit,’ which were “undoubtedly all Brahminical as not a single fragment of Buddhist or Jaina sculpture has been found among the ruins.’17 These temples are believed to have been built not later than the 5th century A.D. and are probably as old as the 3rd18, and if the earliest limit herein mentioned be taken as applicable to some of these shikhara temples also, there would still be the period of three centuries to account for, in order to prove the Āryāvarta temple as a type parallel and contemporaneous to the Mahabodhi. This, however, is a presumption quite unsafe to make in view of the trend of archaeological evidence. The ruins of the two brick temples at Riunpur, District Bareli, now identified with ancient Ahichhatra, are believed to be very old and one is supposed to date from the 1st century, as a number of coins of the so-called Mitra kings were discovered in one of them. The spires of these temples were not apparently of the shikhara form. Dr. Führer believed that ‘the highest mound’, ‘a lingam temple’ ‘rose up in tiers’ and the other a large two-storied Shaiva temple of carved brick ‘had its 1st terrace surrounded by 9 cells and the 2nd by 7 cells’19, thus hardly allowing any room for a superimposed shikhara. The old brick temple at Bhitargaon in Cawnpore District, which is not of such great antiquity, has lost its upper portion (Plate XV, Cunningham’s A. S. R., Vol XI and it is impossible to say whether the spire which crowned the cella bore any resemblance to the Āryāvarta type. Cunningham believed that ‘in its general outline and in the arrangement of bands of ornaments and sculpture’ it made a close approach to the brick temple at Bodh-Gaya. It appears that Cunningham was not so certain as Dr. Vogel of the very early origin of this temple and he only indicated the downward limit by suggesting that it could not be placed later than the 7th or 8th century.20 Dr. Vogel from a similarity of carved brick wall, (ornamented pilaster alternating with terra cotta panels) with those of Nirvana temple at Kasia is inclined to presume that this shrine would go back to the Gupta—if not to the Kushana period.21century. ‘The Arts and Crafts of India and Ceylon’, pp. 121-122 But even if it is believed to be as old as he suggests, it does not disprove the testimony of available facts as to the evolution of the Āryāvarta shikhara in India between the 3rd and the 5th centuries of the Christian era. Among the Sanchi remains curvilinear shikhara temples of the Northern types have also been found, but Sir John Marshall, in his excellent Guide to Sanchi assigns them to the mediaeval period, i.e., about 10th century of the Christian era.22 We have now to turn to the Western Presidency to find examples of spired temples which are as old, if not older, as these specimens of Gupta architecture. In his paper on the Origin and Mutation in Indian and Eastern architecture published in the Transactions of the Royal Institute of British Architects, Mr. Simpson refers to the temples at Aiwulli (Aihole) and Pittadkul (Pattadakal) in Bijapur District, Bombay Presidency, as further evidence of the fact that the Brahmins constructed temples more or less in imitation of chaitya halls. He states that the shikharas constructed over the cellas in these shrines were ‘Dravidian in style’ and ‘quite distinct from the character of the northern shikhara23 Mr. Simpson seems to have been misinformed about the temple types to be found at Aihole—the ‘Aryyapura’ of ancient inscriptions—and its neighboring village of Pattadakal, for we find Mr. Cousens remarking, in connection with the Northern type of shikhara surmounting a shrine at Hutchhimalligudi, that ‘similar shikhara is to be met on old temples in Aihole and Pattadakal.’24 These temples, though not earlier than the 5th century, were not erected under Gupta influence, and they seem to indicate that the introduction of the spire-form here wasn’t due to any foreign or, as Mr. Cousens remarks, any dynastic influence. The Durga temple25 at Aihole and few others of its class (Figs. 4) have elaborately decorated spires of the Northern Orissan type26 and at Pattadakal (Fig. 5), we have, as ‘contrasted more strikingly than at Aihole,’ the Northern and Southern types of shikharas erected long before the evolution of the hybrid Chalukyan style.27 While opinions may differ as to the previous existence of an independent style closely akin to the Chalukya, Mr. Cousens may be taken as correct in his opinion that the temples at Aihole ‘show an unbroken sequence in the styles from the 5th to 14th centuries, from the early cave to the latest mediaeval temple.’28 At Aihole there are flat-roofed massive temples showing ‘the first step from cave Vihara’ like the one known as Ladkhan, which Mr. Cousens assigns to the middle of the 5th century. While in Kontguddi temple, which is of a different class, it is to be noticed that the Chalukyan shikhara in embryo—showing as it does the square base of a tower about 5 feet high—erected over ‘the central bay of the roof.’29 The Durga temple follows ‘the lines of the apsidal Chaitya cave’ and ‘clearly marks the transition from early cave-work to constructional types.’30 Pattadakal, which is only six miles from Aihole, seems likewise, ‘to have been a point upon the dividing line between the northern and southern styles of shikharas where they both overlapped.’31 Not the Dravida alone which flourishes in the south, but the other two styles of Indian Architecture (Nagara and Vesara) were, as mentioned by Dr. P. K. Acharyya, ‘geographical’ in the same sense as ‘Graeco-Roman orders’, and the observation applies mutatis mutandis to the shikhara types. The peculiar spire of the Durga temple brings us, in search of a parallel, to Bhubaneshwar (Figs. 6, 7 & 8) in Orissa, where the closest resemblance is to be met in shikharas of the Parasurameswar and of the Rajnani temples. Mr. Havell, it is true, in his work on the Ancient and Mediaeval Architecture in India, speaks about miniature monolithic shikhara discovered at Sarnath (Figs. 9 & 11) but it appears that the stupas containing representations of shikharas are mostly of the mediaeval period and as such can hardly establish the existence of the spire-form in pre-Christian architecture in this country.32 With the representations of this class of temples on miniature stupa at Sarnath should also be compared the clay seals from Nalanda, two of which we are able to illustrate here by the courtesy of the Editor and Mr. P. C. Nahar, the owner of these specimens (Figs. 12 & 14). They apparently belong to the 11th century. It must not however be supposed that the similarity in these cases is only of a casual character because in the Orissan temples a peculiar line of development is noticeable in the enlarged finial and amalaka coping-stone. In one of the temples lying to the SW of the village of Aihole the shikhara takes ‘more of the bolder curve of Bhubaneshwar—that of Hutchhimulliguddi being considerably straighter in outline.’ Mr. Cousens fixed the age of Parasurameswar at the middle of the 5th century and in this he goes even beyond Messrs. M. A. Arnott33 and M. Ganguly34. The epigraphical evidence, however, seems to support the opinion of Dr. Vincent Smith, who would rather place the Bhubaneshwar temples in the 9th or the 10th century. An unpublished inscription found on the porch of the Parasurameswar temple, recording the endowment of two ‘Naivedya-athakam’ by one Rasheswar with the object of benefitting thereby an ascetic Brahmana, has been deciphered at my request by Dr. R. C. Mazmndar and the epigraphical details examined. Dr. Mazumdar is of opinion that the characters belong to the class of alphabets used in the records of the Somavamsi kings of Kataka and have a striking resemblance to those of the Vakratentuli charter of Mahabhava Gupta I, published by Mr. B. C. Mazumdar.35 The Parasurameswar inscription is a little earlier than the last and, as Somavamsi kings of Kataka have been referred by Dr. Fleet to the 11th century A.D.36, the epigraph may be, for some reason, assigned to the 10th century. In Parasurameswar temple there are sculptures of the most elaborate character and there is nothing to denote any tentative effort which is naturally to be expected from early Orissan workers in stone material. The temple walls have certainly a later appearance and the structure can hardly be ascribed to an age as early as that of the Gupta temples. Even if it be taken for granted that the inscription was put in at a later date, it would hardly be wise to go beyond the limit fixed by Dr. Vincent Smith, viz., the 8th or 9th century of the Christian era.37 Probably contemporaneous with the Parasurameswar temple, are the monolithic shikhara temples of Masrur and the Dhamnar Hindu temples, and one of the early Jaina temples at Osia. The last mentioned temple is held to have been erected during the reign of Vatsaraja between A.D. 770 and 800, and may unhesitatingly be assigned to the 8th century, if not to an earlier date38. The Masrur temples which are situated in Kangra District possess no inscription. From the fact that these monolithic shrines exhibit all the features of Indo-Aryan style, ‘perfected as if they had been practised for centuries,’ there being nothing archaic and no copying of wooden models, Mr. H. Hargreaves is unwilling to believe that they date from an epoch earlier than the 8th century.39 The rock-hewn Hindu temples of Dhamnar in Rajputana40 were made out of a huge pit-like excavation in the midst of which masses of living rock have been hewn into the shape of shikhara temples. The age of these curious shrines can be ascertained from an epigraph at the pedestal of an image of Kalki avatara which are in characters of 8th or 9th century.41 From the 9th to the 13th century, the Āryāvarta style of spire-temples flourished in all its glory. In Orissa, it culminated in the famous Black Pagoda of Konarak which, more than the Muktesvara temple, deserves to be styled as ‘the Jewel of Orissan Art’. Of this temple the spire is said to have been left incomplete but specialists themselves are not agreed on this point. All attempts at reconstructing, on the basis of the present remains, indicate, however, that it was built on the orthodox Orissan model. In this connection, mention should be also made of the temple of Narasinha-natha, situated in a different part of the Utkal tract. This temple is ascribed by Dr. D. R. Bhandarkar to the 9th century, or even to an earlier date42, and may be taken as an example of the transition type between the Pattadakal and the Orissan models. Mr. Simpson has held that the age of the Orissan temples may be determined from the bend or slope of the spire. In temples of a later date, the bend begins mostly on the top, whereas in the tower of the Lingaraj temple (Fig. 8) —which, though, later than Parasurameswar, is undoubtedly one of the early examples of Orissan spire-construction—‘the line rises quite straight from some distance above the cell and only bends a little at the summit.’43 In the Khajuraho group of temples (Figs. 13 & 17) we find a close congener of the Orissan variety. In these shikhara temples of the correct Āryāvarta pattern, constructed mostly between 10th and 11th century, we find, as in the case of Orissan shrines, no uncertainty in any of the parts. ‘Each form is defined and settled; the plan had attained to a recognized type. The decoration is rich, elaborate and beautiful, implying a long experience’ before such results could have been achieved.44 The same remark applies to a certain Lakulishvara temple at Chhotan, Marwar State, believed to be of the 11th century A.D. which, though comparatively small in size, is certainly not unhappy indesign being related less remotely to the Khajuraho chain than to any South-Eastern model (Havell's Handbook PL. XV. A). Of the more recent examples of the spire temples preserving tradition of the Indo-Aryan shikhara may be cited the temple of Dhumeswar at Pawa, Gwalior and the Jaina temples of Ranpur in Jodhpur State. The latter shrines date from the 15th century and have regular curvilinear spires bearing a great resemblance to Orissan shikhara. It is, however, between 11th and 14th century that the shikhara reached its acme of development and it was during this period that the norm seems to have been carried beyond the confines of India. The temple in Lopbari in Siam, depicted in Döhring's Buddhistische Tempel anlagen in Siam Vol. II (Tafel 5), is undoubtedly a temple of pure shikhara type. Of the two classes of old spire pagodas extant in Burma, one is of regular curvilinear form having a close affinity to the curvilinear Indo- Aryan type (Fig. 21). As an example may be mentioned the Ananda pagoda of Pagan. Mr. Charles Duroiselle presumes the existence of an intercourse between a certain Burmese Tantric Buddhist sect and the Vajrayanist and the Sahajiyas of Nepal.45 The Sahajiya element was also strong in mediaeval Orissa and it is not improbable that the curvilinear shikhara form found its way to Burma via Orissa, if not from Northern India. The Bebe pagoda in Prome District bears a considerable resemblance in general characteristics to the old stone Shiva temple at Bajaura in Kulu Valley, Punjab. The shikhara of the former, however, is not characterized by pilasters and panels at the sides, and the amalaka which is so prominent in the Punjab temple is missing in the Burmese specimen.46 The shikhara has been observed over shrines at Triloknath and Nirmand and in view of the frequence of these spires in different parts of Northern India, their transference to Nepal (Fig. 25)47, Indian-Tibet on the Himalayan border, or ‘Western Tibet’ as it used to be formerly called, seems to be a natural sequence. These shikhara temples were believed to have been introduced at Nirmand between the 7th and 11th century and the shrine near the temple of Chandi Devi, of which we give an illustration, may be regarded as a representative example.48 Although Behar and Bengal are both close to Orissa, there is nothing surprising in the fact that in these two provinces the temple form and its spire should show an independent course of development. A distinct departure from the Indo- Aryan shikhara is noticed among the characteristics of the Bengali temples described by the late Rai Bahadur Monmohan Chakravarti49 as also in Tirhut types of temples dealt with by Dr. D. B. Spooner in his lecture before the Behar and Orissa Research Society.50 The greater part of the present province of Bengal is formed out of the deltaic silt deposit. The absence of suitable stone building materials and the nature of the soil are not at all conducive to the preservation of ancient monuments. The existing remains are post-Mussulman and not one of them ‘can be authentically put before the 17th century.’51 It may be that the multi-towers of the Bengali temples were suggested by ‘the Kiosk groups of the later Afgan tombs’ but we find on a careful scrutiny that these towers or miniature sikharas placed over the hut-roofs peculiar to the Bengal style of temples still retain, to a certain extent, the curvilinear form of the Indo-Aryan shikhara, although the amalaka stone is wanting and the tops of some of their spires are built in short horizontal tiers which give them a somewhat corrugated appearance. As an example may be mentioned the spires of the Nava-ratna temple of Raghunath at Baxa, of which an illustration has been given by Rai Bahadur M. Chakravarti, and the very well-known Kantanagar temple in Dinajpur District which has been mentioned by Mr. Fergusson and Dr. V. A. Smith.52 As specimens of temple spires of typical Āryāvarta pattern we may mention the Rajabarir Math of Dacca and Ichhai ghost’s Deul (temple) of Birbhum which have been published by Mr. N. K. Bhattasali in a recent article.53 Both of them though not pre-Mussulman are certainly old enough to be cited as examples of Bengali architecture free from extraneous influence. Dr. Spooner has been at great pains to show that the spires of Tirhut temples are free not only from the influence of Orissan models but also of the Benares type called ‘Bengal temple’ by Fergusson. Dr. Spooner admits that ‘Orissan temples are centuries older than the structures now extant in Tirhut.’54 It is true that in these shrines conical, octagonal, pyramidal and domical spires are to be seen (Figs. 18, 19 & 20) in all their varieties but this modern deviation from the outline of the curvilinear class of shikharas can hardly dispose of the strong influence of the once prevailing Indo-Aryan model, which is so much in evidence in ancient and mediaeval temple architecture in different parts of the country. In no age or clime can architects of genius continue to be slavish imitators of a single model, and if the architects of Tirhut were successful in devising novel spire- forms, it would hardly be justifiable to believe in the absence of any existing remains warranting such a presumption, that these norms can be placed at a period of remoter antiquity than spires of the Āryāvarta class, merely on the ground of a supposed simplicity in structure. As a special feature of Tirhut types, Dr. Spooner has laid stress on the constancy of the rhythm existing between the vertical panelling and the horizontal banding consisting sometimes of ‘very schematic miniatures’. The decorative miniature in horizontal paneling and reduplication of spires is, as Dr. Spooner observes, a striking peculiarity of Hindu ornament and may have been due to an aesthetic sense of the play of light and shade. It is well known that vertical arrangements of miniature spires on the body of the shikhara is also a characteristic of Orissan architecture, though horizontal bands of this sort are not prominent in temples of the Lingraj type (Fig. 8). These details, however, are rather out of place in a general discussion of shikhara structures and a more minute comparative study may well be left to archaeologists interested in the subject. In our search of the original curvilinear type among these diverse forms, we find, as Dr. Spooner has himself admitted, faint traces of a curvilinear outline ‘peculiar to the Indo-Aryan Benares example in the Ramchandra Mandir at Ahiari.’55 We can hardly agree with Dr. Spooner in his opinion that ‘the individual instances’ where the shikhara according to him has thus been influenced in a minor way are merely accidental. The persistence of the older, and we would add, the original, model certainly proves the great hold it had over the mind of the Indian architect and the form would hardly have been so widespread if it had not been the principal prototype of temple spires in India. Even in a land like Tirhut, so rich in varieties of architectural forms, these survivals of the norm of curvilinear Āryāvarta shikhara after a course of centuries should not be cast aside as mere monstrosities even though the later developments may be more striking on aesthetic grounds. Those who direct their gaze into the vistas of the dim Past would rather assert that like the upright stone and the obelisk, the pyramid and the minaret, the spires were originally emblems of the lingam and ‘were seized upon to express the theistic idea.’56 In India, where sex-worship is still widely prevalent, a theory like this does not, at first sight, seem inconsistent. That some pagodas assume an analogous shape ‘has also been remarked by Paolino57 in connection with the prevalence of phallic symbols in India. The fact that spire-temples are built not only to indicate the Shaivite cult but the Vaishnavite and the Buddhist cults as well, remains, however, to be reckoned with. Whether the Phallic idea may be regarded as having even a remote bearing on the primal significance of the shikhara or not, there is no doubt that in the construction of these tapering structures pointing heavenwards like symbolic finger-posts, the Hindu architects were inspired by a two-fold object. The desire to call prominent attention to the shrine, even of distant on-lookers, was certainly there, as the typical flag-adorned upright bamboo58 of wayside rural shrines still bears witness, but though not so apparent to the vulgar, it was no less an emphatic attempt to express in the language of architectural forms the spiritual salvation of men; and to denote the heavenward ascent which a true devotion to the deity, enshrined in the cells below, was believed to ensure to the pious and the pure-hearted. Mr. Havell, who is inclined to find in the shikhara an emblem of Vishnu, the Preserver59, sees here only an application ‘of the water-lily symbol to the roofing of the cell’ which contains ‘the deity’s image in any of its sectarian forms—Buddhist, Jain, Vaishnavite or Saivite etc.’60 He holds that ‘shikhara is Indo-Aryan, not only, because it was mostly found in Northern India or the Ancient Āryāvarta, but because it was introduced into India by the early Aryans and was peculiarly their own contribution to Indian building traditions.’61 Mr. Havell derived the origin of the shikhara from Mesopotamian sources and refers to the royal fortress palaces of analogous forms depicted in the stele of Naram-Sin, now in the Louvre (Fig. 22), and in one of the sculptures figured in Layard's Nineveh (PI. 16, 2nd series), ascribed to the age of Sennacherib. According to Craig, the translator of Dr. Winckler’s History of Babylonia and Assyria, Naram-Sin flourished some time about 3750 B.C. and Sennacherib ruled the Assyrian kingdom nearly three thousand years after him, from B.C. 705 to 68162 and reached the pinnacle of his power in B.C. 689, when he captured Babylon. The cradle of the Aryan race and civilization was, in all probability, somewhere in Eastern Asia. It might have been in Eastern Turkistan or in Asia Minor, but we tread indeed over shadowy ground when we go beyond B.C. 2000 to trace the transference or importation into India of early architectural forms and models. Dr. A. K. Coomaraswamy opines that some of the early motifs, common to ancient Greece and India, reached the latter country via Bactria across the Caucasus range or via Persia across the Persian Gulf, either from Syria or the coast of the Aegean Sea.63 Presuming that the Assyrians derived the architectural device from the common culture ground of the Aryan family, there still remains the necessity of proving that the Assyrians, the Babylonians, or any of the now extinct ancient races, whose architectural remains have been discovered in this tract, had constant and direct intercourse with India even at a later age. The Assyrian and Babylonian cultures might have been historically interdependent and it may be natural to presume from the respective geographical situation that Babylon was in early communication with India. It is even believed by some scholars that the Hindus came into contact with Babylonians at a date later than 2000 B.C.64 The only relic from ancient Babylonia, so far discovered in India, is a stone cylinder seal now in the Nagpur Museum65 and it would be, we think, considered somewhat overbold to hang on this slender thread, the story of Mesopotamian importation of the shikhara, in itself the partial resurrection of an older theory about alleged Iranian or Persian origin which had died a natural death. The fact that the ancient flat-roofed temples of India precede the shikhara structures, and in their turn, are preceded by cave-temples hollowed out of living rock, effectually discountenances any supposition of this sort.

While we are grateful to Mr. Havell for thus bringing to prominent notice, these shikhara-like forms of Assyrian origin, which are undoubtedly the oldest Asiatic types known to us, we think that the conclusions drawn by him, as regards the significance of domical forms and the shikhara-like towered structures, are hardly warranted by fact. One of Mr. Havell's examples is drawn from Plate 17 of Layard's Nineveh or to give the volume its full title, A second series of the Monuments of Nineveh including bas-reliefs from the Palace of Sennacherib and bronzes from the Nimroud (London, Murray, 1854). The plate bears this short description ‘Workmen with implements and ropes for moving a winged Bull’ (Konyunjik) 8. The attempt to associate the spire in the bas-reliefs with the Vishnu or Creator, because of the flowering tree near it and the domical form with the Destroyer (Shiva), because of the presence of yew trees depicted in the vicinity, seem to us to be too idealistic for a matter-of-fact study of architecture. The symbolism sought to be imposed in this connection is clearly out of place. The main theme of the picture is the carrying of a winged bull to the top of an artificial mound. We find in the reproduction a well-defined channel marking a flowing stream and four carts pulled by men apparently in parallel rows. The yew trees are to be found not only near the dome-like building, but also in the lower parts of the panel. Yew trees and flower-bearing trees of an exactly similar type are also depicted in Plate 15 of the same album (Fig. 23), close to a mountain ridge at the top of the picture where no architectural representations appear. This seems clearly to point to the fact that the trees are merely of a conventional kind. The earlier tombs might have been domical in form and the towers or the so-called shikhara structures were probably associated with fortresses, but it is too far-fetched to draw from them any analogy as regards the shapes of Vaishnavite and Shaiva temples.

As we have seen in the course of our enquiry, it has not been possible to trace back beyond the 5th or 6th century A.D. the authentic exellence of shikhara form in this country. It appears to be certain that as in the case of mediaeval Chalukyan shikharas, the Indo-Aryan shikharas which were of undoubtedly earlier origin formed no part of the original temples which were flat on the roof. Among the ancient monuments of Aihole, has also been found a temple in which the shikhara or tower is totally different, being ‘far more archaic looking and clumsy.’ Mr. Cousens almost admits that ‘it gives one the idea of an early stage in the evolution of the Northern style of tower’ and the sole ground of his objection to this inference is that it is ‘not likely that a crude attempt would be erected side by side with the perfected article.’66 In view of the innate conservatism of the Indian master-builders exemplified in the copy of cave types and the more primitive wooden form, this objection seems to be clearly beside the mark. Freak buildings are the rank architectural growth of the present day and if the shikhara, in its fully developed form, had really been imported from abroad, there would hardly have been left any trace of early embryonic examples among these temple-types of ancient Aryyapura. Fergusson, in his work on Indian and Eastern architecture, was unable to trace out the origin of the shikhara beyond describing it as a ‘constructional necessity’. He maintains however, that the shikhara temple is a sure indication ‘of the existence, past or present, of a people of Dasyu extraction.’ Though Mr. Havell finds fault with this theory, Dr. A. K. Coomaraswamy says with no uncertain emphasis that the later styles of architecture have clearly been shown to be the development of aboriginal and non-Aryan structures built of wood, bamboo, thatch, etc., and that architecture had not made much progress among the Aryans when they first entered India. The question is a debatable one and at present it is difficult to hold extreme views on either side. That the non-Aryan’s predecessors of the Aryan invaders made their contributions to the common wealth of culture in the Vedic or post-Vedic period is probable and it is certainly interesting to find in a thatched temple-hut of the Todas, a still surviving example of aboriginal shikhara. What is still more striking is the use of a flat stone to cover the opening at the top of the spire where certain relics are intended to be hidden away from the public gaze.67 From its position at the top of the spire, this stone may well be regarded as the fore-runner of the amalaka. Mr. Simpson who gives an illustration of this structure in his learned paper on Origin and Mutations in Indian and Eastern Architecture, referred to above, is of opinion that the circular rooms in Behar caves with dome-like roofs are the developments of these Toda huts.68 It may be argued that these aboriginal structures were more or less confined to certain localities in Southern India, and, as such, were hardly likely to influence the cave constructions in ancient Magadha. The form, however, must have been familiar to a certain section of aboriginal inhabitants, and it might have caught the fancy of the Aryan settlers, who had come into contact with them. The fact that the Aryan conquerors themselves used a combination of mat and thatch in the construction of their sacrificial sheds, as Mr. Simpson has shown from Prof. Eggeling’s description in the Satapatha Brahmana69 may be regarded as lending countenance to the possibility of imitation of such aboriginal constructions, especially in localities where the intercourse was more frequent and intimate. Mr. Havell in his new book on Art (A Handbook of Indian Art by E. B. Havell, 1920, p. 14) says, in the course of his remarks on ‘the Lotus dome’, ‘in the same way the curvilinear shikhara is the technical modification of the conical hut of Mesopotamian and Persian villages. In both cases the forms were fully developed constructively in India many centuries before Indian craftsmen were pressed into the service of Islam’. As regards this alleged Mesopotamian cum Persian origin of the shikhara, the distinguished critic apparently suggests that the primitive model was imported by the way of ancient Persia. There is some archaeological evidence to show that at the ancient Taxila there was a real point of contact between Indian and Persian cultures. Although among the ruins of that ancient university town of the North-Western border, an Aramaic inscription and the remains of a fire temple have been discovered in the course of recent archaeological excavations not one iota of evidence is yet forthcoming to prove the existence of any shikhara structure coeval with the other monuments. From its position on the border and its cosmopolitan population, Taxila certainly was a place where an imported architectural plan or design was most likely in evidence. If the evolution of the temple spire or shikhara from the conical hut be accepted as a probable hypothesis, I do not see why Toda huts of aboriginal origin, (illustrated in Mr. Simpson’s article), should not be accepted as their ultimate prototypes. The borrowing of primitive hut forms from Persia and Asia Minor has not yet been proved. On the other hand, the familiarity of Aryan settlers with the types of hut of their non-Aryan neighbors may be, I believe, safely presumed under the circumstances. At a later date, the translation of the design into structures made of wooden material may certainly have introduced the ribbed and angular features common to certain types of the shikhara. This seems to have been very aptly illustrated by Mr. Havell by the illustration of a ruined temple at Khajuraho (Handbook pi. XXVII).

While this may be considered as one of the probable theories of the origin of the shikhara, a second, and one hardly more palpable, has been favored by some other scholars of whom Prof. Macdonell may be mentioned as a prominent representative. In his paper on ‘The Buddhist and Hindu Architecture in India’ read before the Royal Society of Arts on February 25th, 1909, he expressed the opinion that ‘the Southern or Dravidian style of architecture is found only within the tropics or south of the 23rd degree of the Northern latitude and the northern or Indo-Aryan style is found only north of the Tropic of Cancer excepting only the eastern and western extremities and that the Dravidian temple has been evolved from the Buddhists monastery while the Indo-Aryan Hindu temple has been developed from one of the two other classes of Buddhist buildings namely the stupa or chaitya.’70 This view has not yet met general acceptance. As Mr. Simpson points out, the progressive stages of development from stupa to the shikhara has never been satisfactorily explained71, and the step showing the preliminary change from stupa to the shikhara form is also missing.72 What has been said of stupa in this connection applies also to the chaitya, and Prof. Macdonell seems to have been in doubt as to which of these two classes of structures the evolution and final shaping of the Indo-Aryan shikhara was due. Mr. Longhurst in a footnote hints at the combination of stupa and chaitya as the basic elements of some of the Hindu temples73, but so far as we are aware, no satisfactory solution has yet been arrived at, regarding the origin of shikhara along these lines of enquiry. It would be hazardous to premise the origin of the curvilinear shikhara from the bell-like dome of the Buddhist cave near Jellalabad74 or from the curvilinear sides of the arch-like cave roof at Bala-Mughrab75 in Afgan territory, while the architectural details pointing to a probable connection with the Euphrates valley is, as Mr. Simpson frankly confesses, ‘so early, that we have no historical knowledge of it.’76

Mr. A. H. Longhurst who lends his support to the theory of the evolution of the temple from the stupa and cites, as examples, the rock-cut temples of Kholvi, where the stupa has been hollowed out into a cell and an image of Buddha placed within, finds difficulty in explaining the curvilinear shape of the Indo-Aryan roof. He suggests ‘the development of the stupa, with perhaps a slight combination of the vihara conventionalism.’77 or 1278 century of the Christian era, when different spire forms, including the Āryāvarta variety, had already attained their full development and fixity of outline. On a careful scrutiny of the illustration given by Mr. Dikohit, we have failed to notice the least semblance to the shikhara tower with its curvilinear sides and corrugated amalaka, and it is very likely that the modeling of the miniature stupa with its tiers of umbrellas over a bell-shaped body, somewhat in the manner of a spire-surmounted sanctum, was due to the strong reaction of Hindu models. We have seen in Puri tulasi-manchas (masonry stands for sacred basil plants) fashioned in the form of temple shikharas According to this view ‘the distinct tiers of ornamental masonry, which form the tower or roof of the temples of this class’ is crowned with ‘a conventionalized stone umbrella, fluted in imitation of the sacred lotus.’ Mr. Longhurst seems to have been convinced that the ama-laka was only a transformation of the umbrellas placed in tiers over Buddhist chaityas. He considers the circular necking between the top of the tower, and this conventional coping-stone with corrugated edge, as merely indicative of the umbrella staff, and the vase-like finial is, according to him, ‘only a development of the gilt knob of the regal or state umbrella.’ That the amalaka ornament occurs not only in the Mahabodhi temple—an instance of a comparatively modern reconstruction—but also in pre-Christian Asoka pillar, is well known and it is also a fact that among the chaityas discovered in the course of excavation at Taxila the conventionalized series of umbrellas over chaitya shrines have been discovered.79 century A. D If the umbrella symbol had become conventionalized into the amalaka as early as the days of Asoka in the third century B.C., their survival, at a later age in a more or less recognizable shape underneath the tree of the chaitya, would have been hardly probable. The coping-stone of the shikhara, whether it be regarded as the ‘solar symbol’80, or as the ‘lotus symbol’ of Vishnu, must have been, as its presence on the Asoka pillar implies, merely the reproduction of a well-known architectural ornament and not a thing inherently connected with the shikhara. The resemblance of the member to the embolic myrobalan, to the udumbara (Indian fig) or to any variety of water-lily or the lotus flower—(Nymphaea cerulea or Nilumhum speciosa) is not of much importance from the point of view of our enquiry, if its raison d'etre is lost to sight. Mr. Havell finds in the amalaka only an application of the water-lily symbol to the roofing of the cell81, but the gradual enlargement of the coping-stone till it attains the generous proportions met with in mediaeval Orissan temples, implies a real ‘architectural necessity’. The object of the ‘amala sila’, the name by which it is known to local architects, must have been to keep in their places the pieces of stone forming the top of the dome. It does not only cover up the opening like the flat stone on the thatched Toda spire, but by its size and position also holds the other stones in their places. It has nothing to do with the ornamental or utilitarian umbrellas we meet in Indian architecture. Even in the example given by Mr. Longhurst of a Shiva temple in Chamba State82, we find the amalaka retaining its independent existence between two umbrellas on the Indo-Aryan shikhara temple, which bears an unmistakable kinship to those of Rajputana and Orissa. No convincing resemblance is noticeable between the ribs of the umbrella and the curvilinear shikhara with its longitudinal panellings and pilasters, intended to give full scope for the play of light and shade. The parallelism here is too far-fetched to provide any satisfactory solution. If any parallel is sought for, in curvilinear forms of the shikhara class, it must be traced to wooden and bamboo structures and especially to the latter. Mr. Simpson has tried to illustrate the bamboo origin of the shikhara83 from the picture of a car in a partly dismantled condition. It must be admitted that of all the theories tentatively put forward, his is the one most likely reasonable and supported moreover by scriptural references. Mr. Havell, who would rather trace the origin of the shikhara from the bamboo frame-works of royal cars used in battle fields84, finds fault with the writer who derives Hindu temples from rathas or cars of Aryan warriors ‘on account of the poetic imagery used in the Ramayana and Mahabharata’ and ‘the attempts made by mediaeval builders to give literary imagination concrete form by placing stone wheels on the sides of vimana’ as in the Kanarak temple.

The more conscious realistic imitations of wood and bamboo processional cars appears to have been made in the mediaeval period85 owing to the desire of the architects to supply more novel forms and to satisfy the craving for artistic innovations, but it would be a mere attempt to confuse the issues if stress is laid in this connection only on the chronological sequence. The familiarity of ancient Hindus with chariots of different classes and the undoubted influence it exercised on popular imagination on account of its secular and ritualistic uses cannot, however, be gainsaid. Processions of cars seem to have formed a part of festivals as early as the days of Asoka.86 In the Artha Sastra, the Superintendent of Chariots is directed to construct chariots of gods, festal chariots, battle chariots travelling chariots, chariots used in assailing enemies, strongholds and training chariots.87 The Buddhist car procession of the 4th century A.D. has been described by Fah-ien88 and the car-festival of the Jaina saint, Jivanta Svamin, by the poet and lexicographer, Hema Candra Suri.89 The Hindu car festivals at Ramesvaram, Tiruvadamudur, Puri, Bhubaneshwar and the car festival of Goddess Viraja at Jajpur are too well-known to need special mention. Paolino also mentions the cars of Tiravancoda and Cangi-puri.90 In Bengal, the cars of Rishra and Mahesh are familiar even to stay-at-home residents of Calcutta. These facts certainly prove that the sacerdotal associations of the rathas have been too well established not to affect the religious architecture of India. Mr. Simpson quotes a passage from the French translation of Ramayana by M. Fauche to show that the city of Ayodhya, in its other features, still appeared, from the numerous shrines for the deities, like the coach-house, in which was stationed, here below, their living or animated cars (‘il semblait encore, a ses nombreaux autels pour les dieux comme la remise ou stationnaient ici-bas leurs chars animes’). The fact that the shrines of the pre-Pauranic age depicted in the sacred epic were shikhara-like in form can very well be presumed from the description given above as also from the following verse occurring in the 2nd chapter of Ramayana in connection with the decoration of the city of Ayodhya for Rama’s coronation. ‘Sitabhra-shikharabeshu devatayatanesu ca, catus-pathesu rathyasu caityesvattalakesu ca91 (Ea.—Bangabasi, p. 149). The use of the word shikhara in the verse in connection with the abodes of gods is specially deserving of notice. There are scholars, however, who regard the passages in the Epic referring to buildings and architecture as subsequent interpolations, but according to Prof. Macdonell even the more recent passages in the Ramayana cannot be later than the 2nd century B.C.92 In view of the available archaeological evidence, however, one would be inclined rather to suggest, that these passages were, in all likelihood, inserted at even a more subsequent period. No ruins of any shikhara of the first two centuries of the Christian era have yet been unearthed, and in the absence of positive evidence it is difficult to assert that the shikhara spires, if any of that early age, were of the present make and pattern. As we have already stated, the spire depicted on the antique Kumrahar plaque, is not of the Indo-Aryan variety. The inevitable corollary to a more complaisant supposition would be that the early Indian architects, unlearned to rediscover it at the later Gupta regime.

It is hardly necessary to add that the car theory per se need not be lightly abandoned on this ground.

The bamboo origin of the shikhara was the direct outcome of the constructional peculiarity of the car. So long as this is admitted, as Mr. Havell has been constrained to do, it is beside the point to argue that the shikhara evolved not from the processional ratha but from the military types of cars, the royal chariots of the battle fields, which were said to have been provided with accommodation for archer body-guards, on raised platforms, and protected with tower-like bamboo coverings containing loopholes. The ‘luminous, awful, foe-subduing cowpencleaving chariot of the ceremonial’ described in the Rigveda (R. V. 2, 23, 3) of which a translation has been given by Muir93 was probably of this description. If read between the lines the adjective ‘heaven-reaching’ would seem also to point to the tallness of the protective superstructure. This car may have been an imitation of the ancient watch-tower as Mr. Havell would seem to suggest, but a vehicle constructed like this in wood, would certainly have been too cumbrous in warfare. It is more likely that in India, where varieties of bamboos are so easily available, the lighter and the cheaper material would supersede the use of planks and timber and by its natural peculiarity would impart a curvilinear outline to the framework. Probably in the course of anthropomorphic development, the cars of the gods would be imagined of this very shape. In the sacred books the word vimana and the celestial ratha are more or less synonymous. To use the picturesque language of Mr. S. Krishnaswami Ayengar ‘Vimana in its origin implies an old-world Zeppelin.’ Though the processional cars might have been so called from a fancied similarity in form to the flying cars of the gods these latter in their turn must have been imagined on the basis of mundane models. We agree, however, with Mr. Ayengar in his opinion that the name vimana thus came to be applied to ‘the tower of inner shrine or the sancta of temples’ and that the vimanas ‘took the place of war chariots.’94 It seems to us to be very likely that the shikharas of Āryāvarta temples came to originate from an imitation of the curvilinear vimanas, even as the mandapas of Orissan temples came to be constructed roughly on the lines of South Indian processional cars of Tiruvadamur type.95

The one main significance of vimana, in fact its principal symbolic expression, seem to have escaped the attention of the specialists in Indian architecture. The shikhara spires over the main shrines in the sancta standing as they do for the old world zeppelin, point necessarily to their imputed celestial origin, and also emphasizes the kindred points in their relations respectively to the human worshippers and the heavenly deities. The spire indicates the deity’s descent to earth by means of the flying car and it further implies that by means of the shikhara the faithful worshipper would, through his assiduous devotion at the shrine and the religious merit accruing on the construction of the temple, be able to ultimately reach the heavenly regions. That this idea must have been innate in the Hindu architect is forcibly brought out in the case of two shikhara temples which the symbolism has been of a less esoteric character.

In a 10th century temple at Tilasma in Mewar, called Talesvara, the shikhara of which is in the Guzrat style, there is just under the amala sila a figure standing against the body of the spire with a conical cap and a sword dangling at the right side represented in the very act of ascending. This figure has been supposed to be that of the royal personage who built the temple and who, by means of that meritorious act, ascended lo heaven.96 In describing another shikhara, this time of a temple at Udaipur in Rajputana, Dr. D. R. Bhandarkar marks ‘one noteworthy circumstance connected with this tower which consists of ‘the carving of a male figure immediately below amala sila at the NW corner.’ Dr. Bhandarkar adds, ‘I was not able to ascertain whose figure it was. The man appears to be intended to hold in his hand the lower end of the flag-staff, which no doubt was placed here. The superb dress and the profusion of ornaments point to his high rank and it’s possible that we have in him a sculpture of the Paramara Udayaditya who constructed the temple.’ To explain the peculiar situation of the figure near the spire-top, Dr. Bhandarkar refers to the shikhara of Talesvara and indicates, by his comparison, that in this case also the sculptor wanted to illustrate ‘the ascent to heaven as a meritorious act.’97 In the madar or large subsidiary shrine to the North-West of the main shrine in the Jaina temple at Ranpur is a sculpture of a sammeta shikhara98 said, ‘just outside the former (not the 2nd madar) is a slab, representing the sacred hill of Girnar.’99 In both of these are prominent, miniature reproductions of shikharas serving as vimanas, and canopying the shrines of trithankaras who are represented as figures in a sitting posture. The symbolic presence of the shikhara implies the heavenward ascent of these Jaina saints. The correctness of this explanation, we believe, will be readily conceded, when it is remembered that on the recent demise of a popular Indian leader, his admirers carried his photograph on a vimana to indicate his translation to heaven as a reward of his manifold merits. This significance of the shikhara must be kept in mind in order to understand its wide prevalence and its popularity, even among sectarian communities like the Jainas who crave rewards for their pious acts not less eagerly than the Hindus. It has not been possible within the scope of this single paper to deal fully with the many-sided problem we set before us at the commencement, and the writer will feel grateful if a competent scholar is induced by this humble effort to take up the matter and give a further and more satisfactory elucidation of the whole question.

  • 1. Longhurst, The influence of Umbrella in Indian architecture, Journal of Indian Art No. 122, p. 6
  • 2. Fergusson’s History of Indian and Eastern Architecture (revised edition by Dr. Burgess), p. 325
  • 3. Mitra’s Antiquities of Orissa, Vol. I, p. 31
  • 4. Of the Bilsar Gupta temples which date probably from the reign of Kumara Gupta, only certain mounds are left indicating the site. It is not therefore possible to say whether there were any spired structures among them. (Temples D. & F. p. 17, Vol. XI A. S. R.)
  • 5. A. S. R., Vol. IX, p. 62
  • 6. Cunningham, Ibid, p. 46
  • 7. A. S. R. (Cunningham), Vol. XXI, p. 96
  • 8. Progress Report Arch. Survey, W. Circle, 1919, p. 61
  • 9. Ibid
  • 10. Cunningham, A. S. R., Vol. XXI, p. 95, et seq
  • 11. Ibid, Vol. X, p. 110
  • 12. Annual Progress Report of the Supdt., Hindu and Buddhist monuments, Northern Circle, p. 16
  • 13. A. S. R., Vol. XV, p. 10
  • 14. Memoirs of the Arch. Survey of India, No. 4, pp. 124-125
  • 15. Op. cit., p. 126
  • 16. Op. cit., p. 124
  • 17. A. S. R., Vol. IX, p. 41
  • 18. Op. cit., p. 47
  • 19. Progress Rep. Epigraphical and Archaeological Branches, North-Western Provinces and Oudh, 1891-92, p. 2
  • 20. Cunningham’s A. S. R., Vol. IX, p. 43
  • 21. Annual Prog. Rep. Arch. Survey, N. Circle, 1908, p. 31 Dr. A. K. Coomaraswamy considers the Bhitargaon temple as the oldest structural example in the northern style and assigns it tentatively to the 4 ---- note here --- th
  • 22. Marshall’s Guide to Sanchi, p. 127
  • 23. Transactions, Royal Institute of British Architects, Vol. VII (N.S.)
  • 24. Cousens ‘The ancient temples of Aihole,’ Ann. Rep. Arch. Survey, 1907-8
  • 25. Progress Report, W. Circle, 1907-8, p. 203
  • 26. Progress Report, W. Circle, 1909, p. 34
  • 27. Progress Report, W. Circle, 1910, pp. 40-41.
  • 28. Loc. cit., 1909, p. 35
  • 29. Annual Report Arch. Survey, 1907-8, p. 192
  • 30. Dr. Burgess, from a comparison of the style of its interior, is of opinion that the Durga temple is allied to the cave III at Badami and he places it within a century after the cave, though Mr. Cousens would rather make the temple precede the cave by a century
  • 31. Progress Report Arch. Survey, W. Circle, pp. 40-41
  • 32. Rai Bahadur Pandit Daya Ram Sahni’s Guide to Sarnath, pp. 218, 219, 220, 222
  • 33. Report with photographs of the repairs executed to some of the principal temples at Bhubaneshwar, Preface, p. 3
  • 34. Orissa and her Remains, p. 307
  • 35. Epigraphia Indica, Vol. XI, pp. 93 ff
  • 36. Ibid., Vol. III, p. 333
  • 37. A history of Fine Art in India and Ceylon, Chap. VII, p. 191
  • 38. D. R. Bhandarkar’s Temples of Osia, Arch. Survey Ann. Rep., 1908-9
  • 39. Arch. Survey Ann. Rep., 1915-16, pp. 47-48
  • 40. The Dhamnar caves are situated near the village of Chandwas about 50 miles to the SW of Jhalpathan. For illustrations see Ind. Antiq. Sept. 1910, p. 246, PI. V & VI
  • 41. Cunningham’s A. S. R., Vol. II, p. 277
  • 42. Annual Report Arch Survey, 1904
  • 43. Origin and Mutation in Indian Architecture, Loc. cit., p. 23
  • 44. Loc. cit., p. 231. 122, Pl. 9
  • 45. Annual Rep. Arch. Survey, 1915-16, p. 83
  • 46. Journal of Indian Art and Industry No
  • 47. The temples of Nepal are known as ‘Shringa Mandirs’ (spire temples)
  • 48. M. Chakravarti in J. A. S. B., Vol. (N.S.) 1909, p. 141, et. seq
  • 49. J. B. O. R. S., Vol. II, pt. II, p. 119, et. seq
  • 50. M. Chakravarti, loc. cit., p. 147.
  • 51. Ibid, fig. 6, p. 157
  • 52. Pravasi, Baisakh, 1329, pp. 74, 76
  • 53. Dr. Spooner in J. B. O. R. S., loc. cit., p. 124
  • 54. Dr. D. B. Spooner, loc. cit., p. 128, and Pl. 9
  • 55. Ibid
  • 56. Sex-worship and symbolism of Primitive Races, Sanger Brown II, pp. 37, 41, 42
  • 57. Paulinus’s voyage to the East Indies, translated from German by W. Johnstone, London, p. 379
  • 58. This fact has also been noted by Mr. Simpson, in his very valued contribution on the subject already referred to
  • 59. Havell’s Indian Allegory, Art and Architecture, pp. 8-9
  • 60. Havell’s Ancient and Mediaeval Architecture in India, p. 63
  • 61. Havell’s paper on Gupta style of Architecture and the Origin of Shikhara, Bhandarkar Commemoration, Volume, p. 444
  • 62. Ostasiastische Zeitschrift, Vol. III, p. 387
  • 63. Vide the article on Hinduism in Hasting’s Encyclopaedia of Religion and Ethics, Vol. VI, p. 688 ff. referred to by the late B. G. Tilak in his article on Chaldean and Indian Vedas, Bhandarkar commemoration volume, p. 30. The Mitani who worshipped Vedic deities like Mitra, Varuna and Indra, is said to have flourished in N. Mesopotamia in 1400 B. C. and at an earlier date. Mr. Tilak held apparently on these data that the intercourse between the Indians and the Turanian races took place at a remote period of antiquity and the indebtedness was mutual and not much one-sided between these two almost contemporaneous civilizations. Loc. cit., p. 37 and p. 42
  • 64. Ibid
  • 65. Mr. R. D. Banerjee’s History of Bengal (In Bengali), Vol. I, pp. 20-22
  • 66. Loc. cit., p. 201
  • 67. Transactions of the Royal Society of British Architects, Vol. VII, N. S., fig. 112, p. 244
  • 68. Loc. cit., p. 253
  • 69. Loc. cit., p. 267, et seq
  • 70. Extract from Prof. Macdonell’s paper quoted by Mr. A. H. Longhurst in his article on ‘The influence of Umbrella on Indian architecture’, Journal of Indian Art and Industry No. 122, p. 7
  • 71. In the Mayamata the upper part of a temple is technically called a stupika, probably suggesting its derivation from a stupa. Ed
  • 72. Simpson, loc. cit., p. 232
  • 73. Journal of Indian Art and Industry No. 122, footnote, p. 7
  • 74. Simpson, loc. cit., p. 256, fig. 125
  • 75. Ibid, p. 257, fig. 127
  • 76. Ibid, p. 257
  • 77. J. I. A. I. No. 122, p. 6. Mr. K. N. Dikshit, in his excellent monograph on six sculptures from Mahoba (Memoirs of the Archaelogical Survey of India, No. 8, p. 2), remarks, in connection with the fragment of the backslab of an image, pl. II, that the miniature shrine depicted thereon ,‘with its spire of umbrellas’, ‘illustrates how the stupa of olds was gradually evolved into the tower of the modern temple.’ This piece of sculpture is dated about 11 ---- note here --- th
  • 78. ---- note here --- th
  • 79. Marshall’s Guide to Sanchi, p. 108. Chaity or Stupa in Mohra Maradu supposed to be of the 5 ---- note here --- th
  • 80. Mr. Havell states that the ‘solar symbol’ crowns the spire depicted on the Naram Sin stele
  • 81. Havell’s Ancient and Mediaeval Architecture in India, p. 63
  • 82. J. I. A. I. No. 122 Pl. 10, fig. A
  • 83. Loc. cit., p. 237, fig. III
  • 84. Havell’s History of Aryan India, p. 112, et seq
  • 85. As examples of faithful copies in stone of wooden cars may be mentioned the Vithala temple of Achyuta Raya, a photograph of which had been reproduced in Longhurst’s Hampi Ruins, p. 131, and the stone car in front of the temples of Tiruvarur, a picture of which has been given by Messrs. Ananthalwar and Rea in their work on Indian Architecture, Vol. II, p. 198. None of these two Ratha temples, it may be remarked, has any Indo-Aryan shikharas
  • 86. Les Inscription de Piyadasi, p. 113
  • 87. Artha Shastra translated by Prof. Shama Shastry, p. 175. Ed. 1915
  • 88. Legge’s Fahien, pp. 18-19
  • 89. Parisishta Parvan (Bib. Ind.) Ed. Jacobi, pp. 282-284
  • 90. Paolino, p. 390
  • 91. Ramayana, II, 6-11. Ed. Bangabasi, p. 149) quoted by A. Govinda Carya Svamin in his paper on Ramayana and Temples, J. B. B. R. A. S., Vol. XXIII, p. 250
  • 92. Macdonell’s Sanskrit Literature, p. 309
  • 93. Original Sanskrit texts, Vol. V, p. 276. Mr. Simpson has referred to the passage in this paper
  • 94. J. R. A. S., July 1915, p. 523
  • 95. A comparison of the pictures of Tiruvadamudur Ratha and the great temple of the Sun God at Konaraka, would certainly bring home to the reader the striking similarity due to original kinship.

    For illustrations, see ‘Three Temples’ (Mandirer Katha), Butterworth, Vol. I, Pl. 32; Vol. II, Pl. 34.

  • 96. Progress Rep., W. Circle, 1905, p. 56
  • 97. Progress Rep., W. Circle, 1914, p. 65
  • 98. Dr. D. R. Bhandarkar’s paper on Chaumukh temple at Ranpur, Annual Report Arch. Survey, 1907-8, PI. LXXXI (a)
  • 99. Ibid, Pd. LXXXI (b), opposite p. 212