Pramāṇa · Search · About · API

History of Indian and Indonesian Art c8.s2: Kuṣāna and Later Āndhra, ca. 50—320 A. D.

History of Indian and Indonesian Art c8.s2

Language
English

Kuṣāna and Later Āndhra, ca. 50—320 A. D.

c8.s2.1

The Yue-chi tribe originally occupied a part of N. W. China. Driven thence about 165 B. C. they first occupied the territories of the Scythian (Śaka) nomads, and later took possession of Bactria, about 10 B. C. By about 50 A. D. under Kadphises I, the first Kuṣāna king of N. W. India, they had occupied Gandhāra, i. e. most of Afghānistān, and the Pañjāb as far as Taxila. Kadphises II (A. D. 90—110) and Kaniṣka (ca. 120—160) extended the Kuṣāna dominion certainly as far as Mathurā and probably as far as Benāres; the eastern territories were governed by viceroys (satraps), Kaniṣka's winter capital being at Puruṣapura (Peshāwar), and his summer capital at Kapiśa in Afghānistān. The ancient University city of Taxila, on the Indian side of the Indus, lay within easy reach of Peshāwar. The Kuṣāna dominions included also Kaśmīr, and in India proper, Mathurā and the Ganges Valley as far as Bihār.

c8.s2.2

It should be noted that the date of Kaniṣka has been the subject of a great controversy; the dates given above are those now accepted by a majority of Indianists, including the late Vincent Smith (4) and Marshall (6) and in accordance with the results of the excavations at Taxila. Rapson, however, adheres to A. D. 78, the initial year of the Śaka era.

c8.s2.3

In centuries preceding the Christian era the Indo-Greek and Indo-Scythian kings of Gandhāra and the Pañjāb had already come under the influence of Indian religions; we have a remarkable instance of this in the pillar erected by Heliodora, ambassador of Antialkidas of Taxila at the court of Vidiśā (Besnagar), about 126 B. C. in honour of Vāsudeva (Kṛṣṇa); Heliodora calls himself a Bhāgavata (Vaiṣṇava). Indian religious symbols appear not only as before on native punch-marked coins, but on various coins of Greek and Indo-Scythian kings from Agathocles (ca. 200 B. C.) onwards. In many cases these symbols had, no doubt, a Buddhist significance, but all are common to Buddhist and Brāhmaṇical usage, and many represent the deities of the cities in which they were struck (e. g. the elephant deity of Kapiśa, and the bull deity, probably Śiva, of Puṣkalāvatī). Religious benefactions by Śaka satraps or their queens are recorded in the Taxila copper plate (ca. 72 B. C.), the Mathurā lion capital inscription (ca. 30 B. C.

c8.s2.4

or near the beginning of the Christian era), and a Jaina votive tablet of about 16—17 B. C., likewise from Mathurā. In the first century of the Christian era figures of Indian deities appear on the coins of Gondophares and Kadphises II, followed in the second century by representations of Buddha, and of Śiva with four arms, on coins of Kaniṣka¹. *Stūpa* bases of the Śaka period, associated with coins of Azes, appear at Taxila.

c8.s2.5

¹ Kaniṣka is represented in Buddhist literature as a Buddhist emperor like Aśoka. His eclecticism is evident, however, from the fact that the deities represented on his coins include Hindū, Buddhist, Zoroastrian, Elamite and other types.

c8.s2.6

The Graeco-Buddhist art of Gandhāra must now be discussed in greater detail. The whole subject is highly controversial, and even the most important points at issue depend upon a balance of evidence rather than upon positive data. It will be helpful to distinguish, as M. Foucher has done, the Indian from the Hellenistic elements of the iconography. The former, which predominate, include a part evidently of Indian origin, and another part that belongs to the common Indo-Īrānian inheritance of "Early Asiatic"; for present purposes these can be considered together. The following motifs exemplify those current in pre-Gandhāran Indian art (cf. figs. 85—95): Types and compositions: *Jātakas* (cf. figs. 47 and 93); the Sun god; Atlantes; Indra, Brahmā, and Yakṣas as probable prototypes of Avalokiteśvara, Maitreya, and Vajrapāṇi²; "woman and tree" (*Yakṣī* or *Vṛkṣakā*) motif³; figures of donors; lotus-seat, &c. Architectural forms: *stūpas*⁴; the double-roofed *vihāra*; the *caitya*-window arch; Buddhist railing; "Persepolitan capital"; battlements. Animals: Lion, elephant, bull, horse; winged lion, centaur, and other monsters; *haṁsa*, *garuḍa*, *makara*, &c. Floral: rose-lotus forms; blue-lotus derivatives (palmette and honeysuckle); vine (already at Sāñcī); various trees. Patterns: diaper, dog's tooth, reel and bead, checker, &c. Symbols: lion, elephant, bull, horse; wheel, *triratna*, &c.

c8.s2.7

² Spooner, 1, remarks that the evolution of Indra and Brahmā "was an accomplished fact prior to any form of the Gandhāra school with which we are acquainted". One aspect of Vajrapāṇi too, seems to derive from old Indian *yakṣa* forms. See also Bachhofer, 2 and Grünwedel, *Athene-Vajrapāṇi*, Jahrb. k. Preuß. Kunstsamml., 1915.

c8.s2.8

³ This most important characteristic of Gandhāra art will be discussed after the Kuṣāna art of the Mathurā school has been described.

c8.s2.9

⁴ At least three types are now current, (1) the early hemispherical, which is becoming rare, except in Ceylon where it persists at least to the thirteenth century, (2) the type with high cylindrical drum and two *vedikā* courses (e. g. Vogel, 13, pl. V) which occurs already at Beḍsā and has been compared above with Phoenician forms, and (3) the type with a bulbous dome (Vogel, 13, pl. IV). A fourth type which must have differed considerably from these, was that of the pagoda-like wooden relic towers (see p. 53). Cf. Finot and Goloubew.

c8.s2.10

Costume: Indian *dhotī*, &c., and jewelry, turbans, &c., while the leading Gandhāran forms which do *not* occur at Sāñcī, Bhārhut, and Bodhgayā, &c., include: Types and compositions: the Buddha figure; many *Jātaka* scenes and scenes from the life of Buddha which now appear for the first time¹ (but in many cases the composition — e. g. the “Visit of Indra” — is practically that of old Indian art, with the Buddha figure inserted); Hāritī (?); garland-bearing Erotes (fig. 89); the nimbus (see pp. 41, 57 ff.). Architectural: the three Greek orders, especially the Corinthian. Floral: acanthus. Costume: various classical, Īrānian, and Central Asian garments and jewelry.

c8.s2.11

¹ It is rarely safe to assume on negative evidence that a composition first met with in Gandhāran art must be of Gandhāran origin. Compare Figures 47 and 93; if we did not possess the first, we might easily have been misled to suppose that no Indian prototype existed for the second. What we know is only a part of what was produced in stone, and what was executed in stone was only a part, probably a very small part, of the total production.

c8.s2.12

The dating of Gandhāra sculptures is a matter of great uncertainty. As remarked by Marshall, “Not one of the thousands of known images bears a date in any known era, nor do considerations of style permit us to determine their chronological sequence with any approach to accuracy”². Foucher and others have attempted to prove that the school developed in the first century B. C., relying partly on the supposition that the best examples must have been the earliest, partly on the Bīmarān reliquary and the few sculptures which are dated in undetermined eras. The Bīmarān reliquary (fig. 88), if we can rely upon Wilson’s (Masson’s) account published in 1841, was associated with coins of Azes³; it is a golden casket, with standing figures of Buddha and worshippers in relief in niches of Indian form, the base engraved with an Indian lotus. This is generally cited as the earliest example of Graeco-Buddhist art⁴; but coins merely provide a *terminus post quem*, and Wilson himself concluded that the *stūpas* of Afghānistān “are undoubtedly all subsequent to the Christian era” (loc. cit. p. 322). A headless standing Buddha figure from Loriyān Tāngai in the Calcutta Museum is dated 318, which, if the Seleukid era is to be understood, gives 6 A. D.; another figure, from Hashtnagar, is dated 384, equivalent by the same reckoning, to 72 A. D.¹. Unsculptured reliquaries and parts of buildings undoubtedly date from the latter part of the first century B. C., e. g. part of the Dharmarājika *stūpa*², and the Jaina *stūpa* base at Sirkap³. Thus, all that can be safely said is that the Gandhāra school of Graeco-Buddhist sculpture may date from the first century B. C., probably antedates Kaniṣka, and certainly attains its greatest expansion in his reign⁴, and that it continues an abundant production in the third and fourth centuries, with increasing Indianisation both there and in Kaśmīr. Gandhāra art is iconographically in part, plastically almost altogether, a local phase of Hellenistic (not Roman — Roman art is cousin, not parent), descended from the art of the Greek period in Afghānistān and the Pañjāb, but applied to themes of Indian origin. It may be described from one point of view as representing an eastward extension of Hellenistic civilization, mixed with Īrānian elements, from another as a westward extension of Indian culture in a western garb.

c8.s2.13

² Marshall, 6, p. 31. For the whole problem consult Adam, Bachhofer, Burgess (8, 9), Codrington, K. de B., Foucher (1, 3, 4), Grünwedel, Marshall (5, 6, &c.), Rapson (2), Rawlinson, Smith (1, 2, 4), Spooner (1—5), Vogel (3, 7, 13), Wilson; also Goloubew, reviewing Foucher, 1, in B. É. F. E. O., 1923, pp. 438 ff., and Coomaraswamy, do. in O. Z., N. F. 1, 1924, and *Indian origin of the Buddha figure*, J. A. O. S., 1926. Goloubew, loc. cit. remarks: “Rien n’empêche en effet, dans l’état présent de nos connaissances, de supposer que le buddha indo-grec du Gandhāra soit une création plastique postérieur de quelques années au buddha indien de Mathurā.” Smith (4, p. 255) calls the Lahore Pallas Athene type (Smith, 2, fig. 66) “the earliest known Indo-Greek sculpture”, but elsewhere points out that the type is Indianised, and may be late (Smith, 2, p. 116).

c8.s2.14

³ Wilson, H., *Ariana Antiqua*, 1841.

c8.s2.15

⁴ Bachhofer; Marshall places it about the beginning of the Christian era, and this is possible.

c8.s2.16

¹ Vogel, 3; Bachhofer, pp. 24, 25.

c8.s2.17

² Marshall, 6, 7.

c8.s2.18

³ Marshall, 6, p. 73, and pl. XII.

c8.s2.19

⁴ "It is a point on which most authorities agree, that the palmy days of Buddhism and Buddhist art in Gandhāra coincide with the reign of the great Kuṣāna kings, and more especially with that of Kaniṣka. This is somewhat more than a hypothesis" (Vogel, 3, p. 258).

c8.s2.20

It should be observed that while the Gandhāran Buddha (figs. 89, 90, 94) is stylistically Hellenistic, it follows Indian tradition, verbal or plastic, in every essential of its iconography. The whole conception of the seated *yogī* and teacher is Indian, and foreign to western psychology, while the Indian Yakṣas afford a prototype for the standing figures. The *uṣṇīṣa* is found already at Bodhgayā, the lotus seat at Sāñcī; indeed, the Gandhāran type of lotus, resembling a prickly artichoke, is far from realising the Indian idea of a firm and comfortable (*sthira-sukha*) seat, and this is really due to the misunderstanding of a purely Indian idea. Nor can the *mudrās*, *abhaya* and *dhyāna* for example, be anything but Indian. All that is really Hellenistic is the plasticity; the Gandhāran sculptor, even supposing his priority in time, did not so much make an Apollo into a Buddha, as a Buddha into an Apollo. He may not have copied any Indian sculpture, but his Buddha type and that of Mathurā are equally based on a common literary and oral tradition⁵.

c8.s2.21

⁵ For the Indian conception of the *yogī* seated in meditation see *Bhagavad Gītā*, VI, vv. 10—21, and the *Samaññaphala Sutta* of the *Dīgha Nikāya*. A seated teacher is represented at Bhārhut, "Digha instructing his disciples", Cunningham, 2, pl. XLVIII, 4; a seated cross-legged figure occurs already on an Indo-Sumerian seal. The westward migration of the Yogī motif is traced by Berstl, who, unacqainted with the examples at Bhārhut, nevertheless inferred its representation in Indian art of the second or third century B. C.

c8.s2.22

The most important remains of Gandhāran art have been found or still exist at Jalālābād, Hadda, and Bāmiyān in Afghānistān, in the Svāt Valley (Udyāna), and at or near Taxila and Peshāwar. At Jelālābād (= Nagarahara, scene of the Dīpaṅkara legend) is the Khaesta *stūpa*, with a magnificent basement; the lower part of the *stūpa* drum is adorned with niches and statues. At Hadda the Tappa Kalan monastery proved to be a veritable museum of Gandhāra sculpture, but nothing that has been excavated has escaped the iconoclasm of the local Musalmāns. At Bāmiyān are many monasteries, innumerable "caves" and some colossal Buddha images; nothing seems to antedate Kaniṣka. One of the colossal images, 53 metres in height, is well proportioned, and slightly "swayed". The trefoil niche in which it stands preserves remains of painting, more or less Indian in aspect. The painting at Bāmiyān, however, exhibits a great variety of styles, and inclines more to the later Central Asian, than to Indian types, as a rule. The Kaniṣka monasteries were built in the open; later in date are the innumerable excavated monastic dwellings¹. Sculptures from the Svāt valley (Udyāna) and many unknown sources are scattered in various collections all over the world, the most important series being that of the Lahore Museum, which includes the sculptures formerly at Peshāwar². The gray slate in which Gandhāra sculptures are executed is supposed to come from an unknown site in the Svāt valley.

c8.s2.23

¹ Godard; Fergusson, 2, vol. 1, pp. 84 ff. Good illustrations of remains in Afghānistān are given by Hayden.

c8.s2.24

² Spooner, 5; and Foucher, 1.

c8.s2.25

Excavations at Takht-i-Bāhi, in the heart of the Yūsufzaī country and centre of Gandhāra have yielded abundant Gandhāra sculptures; the only actual date available is an inscription of Gondophares, A. D. 46, but most of the remains date from the third to fourth century³. The most remarkable monument of Kaniṣka's reign was probably his great *stūpa* near Peshāwar. To sum up the various descriptions of the Chinese pilgrims, it consisted of a basement in five stages (150 feet), a superstructure ("*stūpa*") of carved wood in thirteen stories (400 feet), surmounted by an iron column with from thirteen to twenty-five gilt copper umbrellas (88 feet) making a total height of 638 feet⁴. The monument was probably a transitional form between the simple *stūpa* and the Far Eastern pagoda; a storeyed tower represented on a *toraṇa* architrave in the Mathurā Museum (fig. 69) may perhaps give some idea of its appearance⁵. The site (at Shāh-jī-kī-Ḍherī) has been identified and excavated¹. The total base diameter proved to be some 286 feet, and the monument was thus by far the largest of its kind in India (the base diameter of the contemporary Maṇikyāla *stūpa* is less than 160 feet). In the relic chamber was found the famous Kaniṣka reliquary (fig. 89)². This reliquary consists of a gilt copper alloy cylinder and lid, of total height 7¾ inches. On the lid are a seated nimbate Buddha and two Bodhisattvas, around the rim a series of *haṁsas* with extended wings; and on the cylinder are seated Buddhas, a representation of Kaniṣka, and the sun and moon deities, with garland-bearing Erotes. The inscription mentions the names of Kaniṣka and of Agiśala, the Greek or Eurasian craftsman by whom it was made. The inferior workmanship has provided an argument for regarding the Gandhāran art of Kaniṣka's reign as late in the development of the school, but it is doubtful how far this can be pressed.

c8.s2.26

³ Spooner, 1; A. S. I., A. R., 1912–13, pt. 1, p. 17.

c8.s2.27

⁴ Chavannes, 1 p. 424.

c8.s2.28

⁵ Vogel, 6, 1909–10, pl. XXVII. Some further suggestion of what such a high wooden tower may have been like may be gleaned from the Go-jū-no-tō, Hōryuji, Japan; or from the Sembutsu Ha-no-tō (Hokke *mandara*) of the Hasedara monastery, Japan (*Japanese temples and their treasures*, 11, pl. 209). The *Bukyō Daijiten* (Japanese Buddhist Encyclopedia) explains *tō* as etymologically = *stūpa*. For Indian characteristics in Chinese architecture see also Boerschmann, *passim*.

c8.s2.29

¹ Spooner, 2.

c8.s2.30

² Spooner, 2, 4; Foucher, 1; Smith, 2.

c8.s2.31

The name Taxila (Takṣaśīla)³ covers a number of neighbouring sites. The Bhīṛ mound has already been referred to; the city of Sirkap is Indo-Greek, Scytho-Parthian and early Kuṣāna; Sirsukh is the city of Kaniṣka's reign. The area has yielded remains dating from the Mauryan period onwards; Hellenistic art of the Scytho-Parthian period; and Buddhist art mainly of the Kaniṣka period and later. Of true Gandhāran (Graeco-Buddhist) sculpture not a single fragment occurs in Scytho-Parthian or early Kuṣāna strata. The remains of fifty or sixty *stūpas* and many monasteries have been traced. The following are the chief monuments:

c8.s2.32

³ See Marshall, 6 (the great mass of information contained in this invaluable handbook cannot be adequately condensed in the space here available), and 7.

c8.s2.33

Dharmarājikā *stūpa* (= Chir tope): originally Scytho-Parthian, repaired and enlarged in the Kuṣāna period, and partly refaced in the fourth century. In connection with the building of various periods here a succession of masonry types has been established as follows: rubble and *kañjur* work of the Scytho-Parthian (Kṣatrapa) period, small diaper of the latter part of the first century, massive diaper of the second century, and semi-ashlar of the third and later. The sculptures from this site are all of the later period and include many fine specimens of the stucco Buddha heads of the Indianised Gandhāran type.

c8.s2.34

Chapel G 5 at the Dharmarājikā site is of interest only because of the discovery beneath its floor of the relics of Buddha, accompanied by an inscribed silver scroll dated equivalent to A. D. 78. The Kuṣāna apsidal *caitya*-hall 13 has the end octagonal instead of round. Chapel F 1 had a floor of thick transparent glass tiles, mostly bright blue.

c8.s2.35

In the city of Sirkap, the plan of the great palace has been made out; partly of Scytho-Parthian (Kṣatrapa), partly of Kuṣāna date, it shows a remarkable resemblance to the planning of Assyrian palaces in Mesopotamia, a feature already remarked at Pāṭaliputra. The Jaina *stūpa* base in block F, probably of the Kṣatrapa period, has a façade with niches of three types, Greek pediment, Indian *caitya*-arch, and *toraṇa* : birds are represented as perched on the arches¹, amongst others the double-headed eagle, the oldest known example of this type in India. Temple D is a large structural *caitya*-hall. An elegant female statue in the round, dateable by the evidence of coins about 50 A. D. is of interest in view of the rarity of early dateable examples of Gandhāra sculpture, and of types in the round generally². Amongst the small copper ornaments found are some comma-shaped forms exactly like the well-known prehistoric Japanese *makatama*³.

c8.s2.36

¹ A fact of interest in connection with the resemblance and like usage of the Indian *toraṇa* and Japanese *torii*, and the meaning of “bird roosting-place” assigned to the latter term.

c8.s2.37

² A. S. I., A. R., 1919—20, pl. IX.

c8.s2.38

³ A. S. I., A. R., 1919—20, pl. X. So far as I know, the related *tomo-e* form appears first on a Gupta seal (Marshall, 3, pl. XXI, 120); it is not uncommon as an architectural ornament in later south Indian art.

c8.s2.39

A remarkable temple at Jaṇḍiāl had a lofty central tower and an otherwise flat roof; certainly not Jaina, Buddhist or Brāhmaṇical, it may have been a Zoroastrian fire temple with a *zigurrat* like those of Mesopotamia.

c8.s2.40

The ancient city of Puṣkalāvatī⁴ is probably to be identified with the site known as Mīr Ziyārat or Balā Hiṣār at the junction of the Svāt and Kābul rivers, in the Peshāwar valley. Various sites in the immediate neighbourhood, such as Chārsada (= Haṣṭnagar), Pālātū Ḍherī, Ghaz Ḍherī, &c. have yielded remains of Gandhāran art rather above the average quality, and it is noteworthy that at least five of the very few inscribed sculptures of this school, including two with dates in unknown eras, one however supposed to be equivalent to 6 A. D., have come from this area⁵. That Indian influence in Gandhāra was not exclusively Buddhist is illustrated by the occurrence of a Śiva image (Maheśa, a so-called Trimūrti) from Chārsada; the deity is three-headed, three-eyed, and six-armed, and stands before the bull Nandi, holding the *ḍamaru*, *triśūla* and *kamaṇḍalu*. This type is very close to that of Vāsudeva’s coin, fig. 126. The style is that of the Indianised Gandhāran art of the third century⁶. The same is true of a four-armed female figure from the Momand frontier⁷.

c8.s2.41

⁴ The site of Foucher’s classic picture of the Indian Buddhist visiting the Eurasian craftsman and asking him to make a Buddha.

c8.s2.42

⁵ Bachhofer, 1; Marshall and Vogel; Stratton; Vogel, 3. In the last mentioned the plate references on the first page need correction.

c8.s2.43

⁶ Natesa Aiyar, 1; and A. S. I., A. R., 1914—15, p. 1, pl. XVI d. The type (seated), whether representing Śiva, or as has been suggested, Lokeśvara, reached Kaśmīr and Khotān (Kak, 1, and Stein, 4, pl. 230). Cf. pp. 67, note 1, 99, note 2, and 149. What appears to be part of a three-headed Viṣṇu of the later Kuḷū type is illustrated by Burgess, 8, pl. 22, no. 5. Cf. p. 143, note 4.

c8.s2.44

⁷ Smith, 2, fig. 78.

c8.s2.45

Another important group of *stūpas*, at least fifteen in number, is found at Māṇikyālā, some twenty miles south east of Rāwalpiṇḍī, and several have yielded valuable relics. The largest, and best preserved of all the Pañjāb *stūpas*, is a hemispherical dome of the ancient type, but probably dating from the second or third century and repaired, perhaps in the eighth, when the pillared basement must have been added. Other *stūpas* are met with further down the valley, at Mohenjo-Daro, where fragments of frescoes have been recovered, and at Thūl Mīr Rukhan near Daulatpur¹, Saïdpur² and Mīrpur Khās³, the latter here regarded as of early Gupta date.

c8.s2.46

¹ Cousens, 8, pl. 10.

c8.s2.47

² Bhandarkar, D. R., 7.

c8.s2.48

³ Cousens, 5; 8, pl. 11.

c8.s2.49

Few sites in India are of greater interest than Mathurā. If all that has been excavated had been adequately surveyed at the time, or if all that remains could be made accessible, it is probable that many of the most doubtful problems of Indian political and artistic history might be solved, and much light would be thrown on the early development of the iconography. Even the rich finds, ranging from the Śuṅga to the Gupta period, which are now preserved in the Mathurā, Lucknow and Calcutta Museums, have not been adequately studied⁴.

c8.s2.50

⁴ These finds have been secured very largely through the indefatigable efforts of Pandit Radha Krishna, Honorary Curator of the Mathurā Museum. The inadequate space available for the exhibition is by no means compensated for by the publication of Vogel's Catalogue in 1910. No catalogue of any kind is available at Lucknow, and no recent or illustrated catalogue in Calcutta. Publications fully illustrating all that has been found in Mathurā are one of the first necessities.

c8.s2.51

The pre-Kuṣāna sculptures of the Scytho-Parthian Kṣatrapa period have already been alluded to. Here we are chiefly concerned with those assignable to the reign of Kaniṣka which was the time of greatest production, and those of his immediate successors.

c8.s2.52

The most obvious characteristic of the Kuṣāna school in Mathurā is the fact, by no means astonishing, that it represents in the main a direct development of the older Indian art of Bhārhut and still older art of Besnagar. The position is nevertheless complicated by the development of a new iconography in which the Buddha figure is one of the most important elements, and by the evidence in a few of the sculptures, especially in certain reliefs, of the influence of the contemporary school in Gandhāra.

c8.s2.53

The early Kuṣāna Buddha and Bodhisattva type of Mathurā (figs. 79, 83—85)⁵ is characterised by the following peculiarities: the sculpture is in the round, or very high relief, and always in the mottled red sandstone of Sīkrī or Rūp Bās; the head is shaven, never covered with curls; the *uṣṇīṣa*, wherever preserved, is spiral; there is no *ūrṇa* and no moustache; the right hand is raised in *abhaya mudrā*, the left is often clenched, and rests on the thigh in seated figures, or in standing figures supports the folds of the robe, the elbow being always at some distance from the body; the breasts are curiously prominent, though the type is absolutely masculine, and the shoulders very broad; the robe leaves the right shoulder bare; the drapery moulds the flesh very closely, and is arranged in schematic folds; the seat is never a lotus, but always a lion throne (*siṁhāsana*) without miniature figures, while in the case of standing figures there is often a seated lion between the feet; the gesture and features are expressive of enormous energy, rather than of repose or sweetness, nor is there any suggestion of intended grace. The nimbus is plain or scalloped at the edge in low relief. All of these characteristations apply with equal force to the early Kuṣāna images of Jinas (fig. 86), and the great majority represent the contrary of what is to be found in Gandhāra.

c8.s2.54

⁵ The early inscriptions distinguish by the designations "Buddha" and "Bodhisattva" types which are to all appearances the same; in these cases "Bodhisattva" must refer to Gautama, Śākya Muni, and may be freely equated with "Buddha". In the early Kuṣāna period the iconography is not yet fixed, and there is considerable variety of costume, and it would appear that prototypes of the later crowned Buddhas can already be recognized (fig. 87). It will be noticed that in some cases the right hand raised in *abhaya mudrā* is held sideways (*vyāvṛtta*) in others with the palm forward (*parivṛtta*) as in all later types. The clenched fist should be a symbol of stability, as it is in dance gesture.

c8.s2.55

This is in fact the type of which Vogel remarks that it "cannot be derived from any known class of images in Gandhāra"¹. It is obviously a product of the Indian school, and related by continuous tradition with the type of the pre-Kuṣāna Yakṣas. This is especially evident in the case of the great standing images.

c8.s2.56

¹ Vogel, 6, 1909—10, p. 66. Only for the nimbus and robe has a western origin been suggested (Sahni and Vogel, p. 19). It is hard to believe that the nimbus can have originated outside the classic area of sun-worship. It may be of Īrānian origin, or of Indian origin as suggested on p. 41. The earliest examples in India are found on coins of Hermaios and Maues, thus about 100 B. C. As regards the robe, it is true that the Gandhāran craftsman makes it look like a toga, but the actual shape and use accord with the prescriptions in the Pali canon, and must have been fixed long before the first century A. D. The stylistic handling of the drapery in Indian and Gandhāran types could hardly be more unlike; in India the fleshy form is clearly revealed, ("wet drapery") and the folds of the material are formally arranged, in Gandhāra the body is concealed and the folds of material are loosely and naturalistically treated.

c8.s2.57

The following list includes the more important examples of the Buddha, Bodhisattva and Jina type above described, all either dated in, or dateable in or before, the reign of Kaniṣka:

c8.s2.58

Mathurā (1—6 seated, 7—9, standing): (1) Bodhisattva from the Kaṭrā mound, A 1 of the Mathurā Museum² (fig. 84), with inscription in characters like those of no. 10; (2) Buddha from Anyor, A 2 in the Mathurā Museum, headless, with similar inscription³; (3) Buddha in the Museum of Fine Arts, Boston (fig. 85), without inscription; (4) Buddhas in relief, of small size, not cult images, N 1 and J 24 in the Mathurā Museum¹; (5) Jina from the Kaṅkālī Ṭīlā site, J 39 in the Lucknow Museum (fig. 86); (6) Jinas represented in relief on several *āyāgapaṭas* from the same site, now in the Lucknow Museum²; (7) standing Buddha relief, J 18 in the Mathurā Museum³; (8) A 41 in the Mathurā Museum, like nos. 9—12; (9) headless figure from the Gaṇeśra mound, like nos. 10—12, but with ornaments, now in the Lucknow Museum⁴.

c8.s2.59

² Vogel, 13, p. 46, and pl. VII.

c8.s2.60

³ Vogel, 13, p. 48, and pl. VIII.

c8.s2.61

¹ Vogel, 13, pp. 148, 166, and pls. III a and IV.

c8.s2.62

² Smith, 1, pls. VII, XVII, CI.

c8.s2.63

³ Vogel, 13, p. 146, and pl. III c. Other very important railing pillars with Bodhisattva or similar figures are numbered B 82, B 83, B 88 in the Lucknow Museum; similar reliefs in the Pennsylvania University Museum (fig. 80). Cf. fig. 87.

c8.s2.64

⁴ Smith, 1, pl. LXXXVII.

c8.s2.65

Sārnāth: (10) colossal standing Bodhisattva, B (a) 1 in the Sārnāth Museum (fig. 83) dedicated by Friar Bala in the third year of Kaniṣka (123 A. D.), with a richly carved umbrella, lion between the feet, and traces of original colouring, a magnificent and powerful figure, perhaps the finest of the group, and with no. 1, the most important⁵; (11) headless Bodhisattva of similar type now in the Indian Museum, Calcutta⁶. The two last similar to no. 7.

c8.s2.66

⁵ Sahni and Vogel, pp. 18, 33—37 and pls. VII, VIII; Oertel, pl. XXVI a, b. Cf. fig. 96 (head in the Museum of Fine Arts, Boston). The similar figure B (a) 2, Sārnāth Museum, in Chunār sandstone, is regarded by Sahni and Vogel, p. 37, as a copy of the Mathurā type by a local sculptor. Friar Bala's name occurs also at Śrāvastī and in Mathurā (Sahni and Vogel, p. 36). As regards the tendency to colossal size, it may be noted that a detached left hand found at Mathurā (Cunningham, 4, vol. 1, p. 239) measured a foot across the palm, indicating a figure twenty four feet high. Cf. also the older statues from Pārkham and Baroda.

c8.s2.67

⁶ Smith, 2, fig. 94.

c8.s2.68

Śrāvastī: (Saheṭh-Maheṭh): (12) standing Bodhisattva from the Jetavana site, dedicated by Friar Bala, similar to no. 10, now in the Calcutta Museum⁷; (13) seated figure, all above the waist now missing, but evidently like nos. 1—3, made by "Sivamitra, a sculptor of Mathurā" and set up "in the Jetavana at Śrāvastī" by two brothers "with special regard to the welfare of their parents", the inscription in Kuṣāna Brāhmī script⁸. Pāṭaliputra: (14) Bodhisattva fragment "which can only have been produced by the famous school of sculpture which flourished at Mathurā", also referred to as "at least one large and inferentially elaborate Bodhisattva statue from Mathurā, which is to be assigned probably to about the dawn of the Christian era or a little later"⁹. No reason is given for this early dating. Rājagṛha: (15) fragment of a Kuṣāna Buddha pedestal of Mathurā origin¹. Sāñcī: (16) Buddhas, A 82 and A 83 from Mathurā, assigned to the second century, and a Bodhisattva fragment².

c8.s2.69

⁷ Oertel, pl. XXVI d; Ep. Ind., VIII, pp. 180 ff.; J. A. S. B., XLVII, 1, 1898, p. 278.

c8.s2.70

⁸ Sahni, 4. Dedications, not necessarily of images, "for the welfare of departed relatives" are mentioned in the *Milinda Panha*. The Sāñcī Buddha A 83 was dedicated "for the happiness of the donor's parents and of all creatures". Dedications were generally made *ātmāparahitam*, "for the benefit of oneself and others".

c8.s2.71

⁹ Spooner, 7; and in A. S. I., A. R., 1912—13, Pt. 1, p. 26.

c8.s2.72

¹ Vogel, 6, 1906—07, p. 143, note.

c8.s2.73

² Marshall, 12, pp. 29, 30, and pls. II, XII. A 82 Sāñcī closely resembles A 45 at Mathurā (Vogel, 13, pl. X). See also A. S. I., A. R., 1912—13, Pt. 1, pl. VIII b.

c8.s2.74

In addition to these sculptures the supposed figure of Buddha, but perhaps a king, of the early Kuṣāna type, seated cross-legged, broad shouldered, the left hand on the thigh, the elbow extended, but with some undetermined object held in the raised right hand, appears on certain coins (fig. 119) of a king Kadapha, who is probably to be identified with Kadphises I, and must have reigned near to the middle of the first century A. D.³. Certain of Kaniṣka's coins bear the standing figure of Buddha in loose diaphanous robes, with nimbus and body-halo, and the legend "Boddo" (fig. 123)⁴; others a seated Buddha, apparently with curly hair and in any case of a later type than Kadapha's, with the legend "Goboydo" (Go[tamo] Budo)⁵. Mr. Longworth Dames believed that the much earlier seated broad-shouldered cross-legged figure on the reverse of a coin of Maues (Smith, 6, p. 12, and pl. VIII, 4) was a figure of Buddha⁶. The type is at least as close to a Buddha figure as that of the Kadapha coin, and would be the earliest Buddha figure known; but the identity cannot be regarded as established beyond all doubt in either case.

c8.s2.75

³ For this coin see Whitehead, R. B., pl. XVII, 29; and Smith, V. A., in J. A. S. B., LXVII, pt. 1, 1898; also Marshall, 13, pag. 34 and pl. XXV, 18, 19.

c8.s2.76

⁴ Gardner, pl. XXVI, 8; Whitehead, R. B., pl. XX, 7; enlarged reproduction, Adam, p. 21.

c8.s2.77

⁵ Zeit. für Num., 1879, pl. IX, 1; Marshall, 13 pag. 34 and pl. XXV, 20 and Cunningham, Coins of the Kushans, pl. XVII, 12.

c8.s2.78

⁶ In J. R. A. S., 1914, p. 793.

c8.s2.79

None of these examples, other than the doubtful coins of Kadapha and Maues, can be proved to be older than the reign of Kaniṣka. It would nevertheless by very rash to assume that none of the sculptures can be older, or that any one of them is necessarily the oldest of its kind ever made. In any case Mathurā must have acquired a high reputation as the source of Buddha images before so many colossal figures would have been exported to comparatively distant sites, and this consideration certainly involves the existence of Mathurā Buddhas in the first century A. D.

c8.s2.80

It is evident from what has been said, and from the illustrations, that a type of Buddha image had been created at Mathurā independently of any Hellenistic prototype; and that this Mathurā type was transported to many other sacred sites, for at the very beginning of Kaniṣka's reign we find Mathurā "sending down images to the sacred sites of the Gangetic plains, thus setting examples to the sculptors of Benares and Gayā”¹.

c8.s2.81

¹ Vogel, 13, pp. 28, 34: “There is plenty of evidence that the Mathurā school greatly influenced Buddhist art throughout the period of its existence”. I fail to understand why Vogel regards this circumstance as “not a little curious”. A study of the literature also shows “que l'église de Mathurā eût parmi les communautés bouddhiques une situation privilégiée et qu'elle eût contribué pour une large part au rayonnement du foi“ (Przyluski, p. 9). Mathurā fragments have also been found at Taxila (Marshall, 7, p. 39). The later Kuṣāna type has been recognized in Central Asia (Foucher, 1, figs. 562, 563 = Grünwedel, 3, pl. IV, 1, and Stein, 4, II, pl. LXXXII); and in China (Sirén, 1, pp. 37, 38, 43, 66, 144). See also p. 17, and Coomaraswamy, 16.

c8.s2.82

These facts, taken into consideration with the subsequent continuity of the tradition, and the obvious and natural relationship of Gupta to Kuṣāna types, exclude the possibility of a “Greek origin of the Buddha image” in India. That in certain directions a Hellenistic element, plastic and iconographic, was absorbed into Indian art, and that the presence of this factor is sometimes unmistakeable, is all that can properly be asserted in this connection.

c8.s2.83

All the Mathurā sculptures showing traces of Hellenistic influence, taken together, constitute a very small fraction of the whole production of the school. Given the identity of theme, a greater divergence of the early Mathurā Buddha, Bodhisattva and Jina type from that of the Buddha and Bodhisattva of Gandhāra could hardly be imagined (cf. figs. 94, 96). The wide distribution of the Mathurā type and the fact that it was locally copied show that it was regarded as the orthodox model. Incidentally, this has a bearing on the question of the date of the Gandhāra school; for if the Gandhāran type had been evolved and acquired prestige long before Buddhas were made at Mathurā, the Indian sculptors, who had no prejudice against a foreign style, would surely have made use of it². The only possible conclusion is that the Buddha figure must have been produced simultaneously, probably in the middle of or near the beginning of the first century A. D., in Gandhāra and in Mathurā, in response to a demand created by the internal development of the Buddhism which was common ground in both areas³; in each case by local craftsmen, working in the local tradition.

c8.s2.84

² It should not, however, be forgotten that as Le Coq (3, p. 28) remarks “Allen Asiaten erscheinen Europäergesichter (also auch die der Hellenen) sehr unschön”. Cf. Masson-Oursel.

c8.s2.85

³ The development of devotional theism, cf. Jacobi, S. B. E., vol. XXII, p. XXI.

c8.s2.86

Only after the local types had been established did each affect the other. Here, indeed, there is a legitimate field for discussion with a view to definition of the influence, and of the extent to which its trace can still be recognized in *the* Buddha type which is definitely established early in the Gupta period (figs. 98, 154, 158—161, 164). The possibility of any further Hellenistic influence having been exercised at that time is of course excluded, as it is well known that the Gandhāra type in Gandhāra and in Kaśmīr was then subject to Indian influence and had already been greatly Indianised.

c8.s2.87

Somewhat later than the type above described is another closely related to it, characterised by the following peculiarities: the general treatment is rather more refined; the robe is often thrown over both shoulders, and in seated figures, both feet are hidden, and more voluminously rendered; figures, probably of donors, appear on the pedestals; and above all, the head is covered with short curly hair, examples of the shaven head gradually disappearing. Amongst examples of these later Kuṣāna figures from Mathurā may be cited:

c8.s2.88

Mathurā: (1) the Sītalā Ghāṭī image, A 21 in the Mathurā Museum¹; (2) standing image A 4 in the Mathurā Museum²; (3) small standing image with shoulder flames 25. 439, in the Museum of Fine Arts, Boston³; (4) Mathurā relief, fig. 104.

c8.s2.89

¹ Vogel, 13, pl. XVI.

c8.s2.90

² Vogel, 13, pl. XV a.

c8.s2.91

³ One of very few known examples, the others all Gandhāran: see A. S. I., A. R., 1921—22, p. 65; 1922—23, pl. XXV; and J. A. S. B., III, pl. XXVI, i. Figures of Śiwa and of kings on Kuṣāna coins also exhibit shoulder flames. Cf. the “essence of fire and hereditary royalty” of the Devarāja cult (see page 197).

c8.s2.92

Śrāvastī: seated image exactly like no. 1⁴.

c8.s2.93

⁴ Marshall, in T. R. A. S., 1909, p. 1065, and pl. III b.

c8.s2.94

Sāñcī: no. 19, Sāñcī Museum, possibly early Gupta⁵.

c8.s2.95

⁵ Marshall, 12, pl. 11.

c8.s2.96

It is this type, though it still closely adheres to the vigorous Kaniṣka formula, with prominent breasts and full features, that shows the first signs of a rapprochement with Gandhāran art, as indeed, specifically remarked by Goloubew with reference to no. 2⁶. On the other hand, a rapid Indianisation of the Hellenistic type begins in Gandhāra, in general, though not invariably accompanying a degeneration of the style, and a substitution of stucco for stone. Thus, about the end of the second or beginning of the third century some mutual influence of the two styles upon each other is traceable, and it is at this stage if at all that a Hellenistic element can be said to have entered into the constitution of *the* Buddha type⁷.

c8.s2.97

⁶ B. É. F. E. O., 1923, p. 452.

c8.s2.98

⁷ A conclusion quite in accordance with the fact that Gandhāran characters are rather more clearly traceable at Amarāvatī than at Mathurā.

c8.s2.99

The exact date at which the type with curly hair appears is uncertain; Sahni⁸, restricts it to the Gupta and subsequent periods, but it certainly appears, both in Gandhāra and Mathurā before the end of the third and probably before the end of the second century. The source of the type is probably to be found in the tradition which first appears in the *Nidānakathā* to the effect that when the Bodhisattva shore his locks, becoming a hermit, his hair “was reduced to two inches in length, and curling from the right, lay close to his head, and so remained as long as he lived”¹. The spiral form of the early Kuṣāna *uṣṇīṣa* must have a like meaning, the later type merely covering the whole head with short curls in place of the original single lock. This later type becomes the universal rule, alike in India and the Far East, only one figure with shaven head (the Māṅkuwār image fig. 62) being known of early Gupta date.

c8.s2.100

⁸ Sahni and Vogel, p. 33, note 8, but contradicted, *ibid.* p. 75.

c8.s2.101

¹ Rhys Davids, *Buddhist Birth Stories*, p. 93.

c8.s2.102

If none of the Buddha images in the round found at Mathurā can be said to be copies of Gandhāra types, and only the later Kuṣāna type exhibits a few Gandhāran characteristics, the same cannot be said in the case of the rarer reliefs illustrating scenes from the Life, particulary the Eight Great Miracles. Good examples of such reliefs include H. 1 (fig. 104), H 7 and H 11 of the Mathurā Museum; here, as remarked by Vogel, the copying of Gandhāran compositions is evident². The same applies to the life-scenes of the Dhruv Ṭilā *stūpa*-drum, N 2 of the Mathurā Museum, described by Foucher as a “caricature lamentablement indianisée” of the *stūpa*-drum of the Lahore Museum from Sikri in the Peshāwar district³. The decorative motif of garland-bearing Erotes, already indianised at Mathurā (fig. 76), and much more so by the time it reaches Amarāvatī, is likewise of Gandhāran origin. The only example of actual Gandhāran sculpture in the well-known blue slate of the Swāt valley certainly found in Mathurā is a late image of Hāritī, the “Buddhist Madonna” and consort of Kubera as Pāñcika, F 42 in the Mathurā Museum⁴. Another group of sculptures, stylistically Indian, is nevertheless strongly suggestive of the West in respect of its themes; well-known examples include the so-called Herakles and the Nemean lion, and the various Bacchanalian compositions, which are really representations of the aforesaid Pāñcika and are of Buddhist significance⁵.

c8.s2.103

² Vogel, 13, pl. VI; Vogel 6, 1909—10, pls. XXV, etc.

c8.s2.104

³ Smith, 1, pls. CV—CVII; Foucher, in J. A., serie X, vol. 11, 1903, p. 323; Vogel, 13, pp. 166—168. Sikri near Peshāwar should not be confused with Sikrī, one source of the red sandstone used at Mathurā.

c8.s2.105

⁴ Vogel, 13, p. 118; Burgess, 8, pls. 56, 57. For the “Buddhist Madonna” and “Tutelary pair” (Hāritī and Pāñcika) see Foucher, 4, and for examples from Mathurā, Vogel, 6, 1909—10.

c8.s2.106

⁵ For the Bacchanalian scenes see Smith, 4, pp. 134—139 and references there cited; Vogel, 6, 10, 13, 16; Coomaraswamy, 9, 2, pl. III; Foucher, 1, vol. II, p. 151.

c8.s2.107

No complete railings have been traced in Mathurā, but parts of many different Buddhist and Jaina railings have been discovered at several sites. Of these the most important are the pillars and pillar-bases from the Jamālpur (“Jail”) mound, also the source of the water-nymph (fig. 74), most of the pillars and the nymph being now in the Lucknow Museum, the bases divided between Calcutta and Lucknow: pillar bases from the Court-house mound; pillars and cross-bars from the Kaṅkālī Ṭīlā site of the Jaina Yaśa-vihāra and *stūpa* of Vāsiṣka: pillars from the Bhūteśar site, divided between the Calcutta, Lucknow and Mathurā Museums: those from Jaisiṅghpura, and some others.

c8.s2.108

The pillars and bases from the Jamālpur mound are of three sizes; some bear dedicatory inscriptions of the reigns of Kaniṣka and Huviṣka. Those from the Kaṅkālī Ṭīlā are associated with inscriptions ranging from the Kuṣāna years five to ninety eight. Older fragments are not unknown, one from the Arjunpura site bearing a Mauryan inscription; but those here considered appear to belong for the most part to the latter half of the first, and to the first half of the second centuries A. D., with some perhaps as late as the end of the second century. It is unfortunate that so many of the Mathurā pillars, and *āyāgapaṭas* cannot individually be more exactly dated. The Mathurā excavations were conducted solely with a view to collecting sculptures and without regard to scientific observation of the sites; and the difficulties of study have been increased by a distribution of sculptures from the same site amongst at least three different museums¹.

c8.s2.109

¹ Fa Hsien describes about twenty Buddhist monasteries with three thousand monks as existing in Mathurā at the beginning of the fifth century. Some of these are represented by the mounds from which sculptures have been extracted. Amongst the buildings for which rather more precise evidence exists may be mentioned the Jaina Guha-vihāra founded by the chief queen of the Satrap Rañjubula towards the end of the first century B. C.; the main Jaina establishment with the Vodva stūpa at the Kaṅkālī Ṭīlā site, which had existed in the second century B. C. and survived into the twelfth A. D.; the Buddhist Yaśā-vihāra and stūpa at the Kaṭrā site, restored by Vāsiṣka, later replaced by a Brāhmaṇical temple and finally by Aurangzeb’s mosque; the Buddhist monasteries and stūpas at the Jamālpur and Court-House sites, with a monastery named for Huviṣka and a shrine of the Nāgarāja Dadhikarṇa; the Brāhmaṇical temple at Māt, mentioned in two inscriptions as a *deva-kula*, source of the royal statues, and perhaps the particular church of the Kuṣāna kings. For a general account of these sites, see Vogel, 13; for other details, Smith, 1, Cunningham, 4, vol. III, and Vogel, 6, 10, 15, 16. An inscription indicating the existence of a Vaiṣṇava shrine with a *toraṇa* and *vedikā* at or near Mathurā in the reign of Śoḍāsa has been edited by Chanda, 4, pp. 169—173.

c8.s2.110

The sculptures represented in high relief on the front sides of the pillars include Buddhas, Bodhisattvas, Yakṣīs (or Vṛkṣakās), toilet scenes and other genre subjects, and a few male figures; the backs of the pillars bear lotus medallions, or in a few examples, *Jātaka* panels. Of the Bodhisattvas or Buddhas, the standing figure from the Jamālpur site, B 83 in the Lucknow Museum (fig. 79) is identical in type with the Sārnāth Bodhisattva from Mathurā, dated in the third year of Kaniṣka. J 18 in the Mathurā Museum is of the same kind. B 82 in the Lucknow Museum (fig. 78) is a crowned Bodhisattva in secular costume, holding the *amṛta* vase in the left hand, and having the representation of a Dhyāni Buddha in the crown. This is one of the earliest examples of an Avalokiteśvara thus unmistakeably designated: for a Gandhāran example see fig. 95. B 88 in the Lucknow Museum is another figure in secular costume, holding a bunch of lotuses in the raised right hand, and probably represents a donor (fig. 77), or may be connected with the Dīpaṅkara legend.

c8.s2.111

The great majority of the remaining figures are female (figs. 73—75). The commonest and most characteristic type, indeed, is that of the nude or semi-nude female figures associated with trees, unmistakeable descendants of the Yakṣīs and Vṛkṣakās of Bhārhut, Bodhgayā and Sāñcī, and ancestors of the Rāmeśvaram verandah brackets at Elūrā, those of the Vaiṣṇava cave at Bādāmī, and many later derivatives. What is the meaning of these sensuous figures, whose connotation and implications are by anything but Buddhist or Jaina? They are certainly not, as they used to be called, dancing girls¹; they are Yakṣīs, Devatās or Vṛkṣakās, nymphs and dryads, and to be regarded as auspicious emblems of vegetative fertility, derived from popular beliefs². Trees, as we have already seen, are closely connected with fertility, and tree-marriages have survived to the present day; the twining of the limbs of the dryads, as in the Bodhgayā pillar, deliberately or unconsciously expresses the same idea³. It will, indeed, have been observed that there is scarcely a single female figure represented in early Indian art without erotic suggestion of some kind, implied, or explicitly expressed and emphasized; nowhere, indeed, has the vegetative sexual motif been presented with greater frankness or transparency, though in certain later phases of Indian art, as at Khajurāho and Koṇārak, more specifically. The railing types are to be connected with and perhaps derived from the early terracottas, which in their turn remind us of the nude goddess once worshipped throughout Western Asia, and of the gold plaque of the Earth goddess from Lauṛiyā-Nandangaṛh. In the presence of these emblems of abundance we must not be misled by modern ideas; their meaning, if not Buddhist or Jaina, is nevertheless religious, and reveals an essential purity of spirit that has at all times preserved the East from many psychological disasters that have overtaken the West. The two polar themes of Indian, indeed, of all experience, are there presented, side by side, though not in opposition; in much later, mediaeval, Vaiṣṇava art we find them unified¹.

c8.s2.112

¹ Even Le Coq, 3, p. 82, makes the mistake of describing Māyādevī's stance in the Nativity scene as “in der Tänzerinnenstellung”; No Indian representation of a dancer in this position can be cited. It may be remarked further, that no sufficient reason exists for the usual description of the female figure on the coins of Pantaleion and Agathokles as an “Indian dancing girl”. For the Bhārhut figures, however, cf. Mitra, K., *Music and dance in the Vimānavatthu-atthakathā*, J. B. O. R. S., XII, 1926.

c8.s2.113

² The later Saṁhitās speak of trees as the homes of Gandharvas and Apsarases. Vogel, 13, p. 44, quotes appositely from the *Mahābhārata*, “Who art thou, bending down the branch of the Kadamba tree? A Devatā, a Yakṣī, a Dānavī, an Apsaras, a Daityā, a Nāginī, or a Rakṣasī?”. The *Mbh.* also speaks of dryads (Vṛkṣakā, Vārkṣī) as “goddesses born in trees, to be worshipped by those desiring children”. The female figures associated with trees in Bhārhut reliefs are labelled as Yakṣīs and as Devatās. On the other hand some of the figures standing under trees are evidently human. The special adaptations of the dryad motif (1) in the case of the Buddha Nativity in the Lumbinī garden, where Māyādevī supports herself by the Śāl tree, and (2) in the poetical fancy of the Aśoka tree, which blooms only when touched by the foot of a beautiful woman, are evidently secondary, though still closely connected with the idea of fertitily. For the general significance of feminine divinities worshipped through the ancient world see Glotz, J., *The Aegean civilisation*, 1925, pp. 243—245.

c8.s2.114

³ Cf. the tree-girl married by the ten Pracetas in *Mahābhārata*, I, 196, 15; and Hsüan Tsang's story of the origin of the name Pāṭaliputra, from the marriage of a student to the maiden of a Pāṭali tree, resulting in human offspring.

c8.s2.115

¹ The subject has been ably discussed by Berstl, who traces the westward migration of both motifs, Vṛkṣakā and Yogī, *via* Alexandria and Syria. See also an admirable article on *Art and Love*, by Eric Gill, in Rūpam, no. 21; and remarks by Keyserling, *Travel Diary of a Philosopher*, vol. 1, p. 97.

First 115 of 159 passages. Read on

Cites (45)