Asian Art: The Great Miracle of Crasvati

The Great Miracle at Cravasti

by Alfred Foucher

Originally published in Journal Asiatique, Jan.- Feb. 1909.
English translation by L. A. and F. W. Thomas in
Beginnings of Buddhist Art (1914)

147

The narratives of the death of Buddha assert that after his cremation eight kings or ruling clans shared his ashes, and that they deposited their several portions under as many stupas. We see no reason for disbelieving tradition on this point: the important thing is not at times to confuse these first eight sanctuaries with the historical «eight grand caityas» (1). We know, indeed, for certain that eight towns of Madhyadeca had finally divided among themselves, not the relics, but the legend, of Buddha. In their immediate neighbourhood were eight specially holy places, supposed to preserve the vestiges of the eight principal miracles of the Master. This implies that they formed as many centres of attraction for pilgrims, the organized exploitation of whom -- one of the few industries which still survive in India -- must have constituted an appreciable source of income. It may easily be imagined that the definite choice of scenes and sites was not accomplished without rivalries and hesitations.

(1) That there is no lack of opportunity for these confusions we find proved at once in Une poesie inconnue du roi Harsa Siladitya, restored from Chinese transcription by M. S. Levi (Actes du Xe Congres Int. des Orient., 1894, I, p. 188, Leiden, 1895) and entitled «Hymn to the Eight Great Caityas», which enumerates still more. The «eight reliquaries» of stanza 5, followed by the stupas of the «urn» and the «ashes», are evidently the 8 + 2 stupas of the Mahaparinibbana-sutta, VI, 62, and have nothing to do with the «eight great caityas» of the title.

148

At least four cities, indeed, received from the first an undisputed recognition. A relatively ancient text, the Mahaparinibbana-sutta, already recommends the pilgrimages to the four sacred places of the Birth, the Illumination, the First Preaching, and the Death, of Buddha (1). On the square bases of the little stupas of Gandhara and the stelae of Amaravati these four scenes are invariably associated (plates II-IV): only we must draw attention to the fact that in the latter case Kapilavastu is usually represented not by the nativity of the child Buddha, but by what might be called his birth into the spiritual life, we mean his «abandonment of home» (2). However, neither the cities of Gaya and Benares, nor certainly the obscure frontier market towns of Kapilavastu and Kucinagara could pretend to monopolize between them the Buddhist legend and the advantages accruing therefrom.

(1) V, 16-22; the Jati, the Abhisambodhi, the Dharmacakrapravartana, and the Parinirvana are similarly associated in the Divyavadana, ed. Cowell and Neil, p. 244 and p 397, 1. 18.

(2) For the little stupas of the northwest, cf. Art Greco-Bouddhique du Gandhara, fig. 208. For the stelae of Amaravati, see J. Burgess, The Buddhist Stupas of Amaravati and Jaggayyapeta, pll. XVI, 4; XXXII, 4; XXXVIII, 5; XLI, 6 (with the departure on horseback; cf. J. Fergusson, Tree and Serpent-Worship, pl. LXXV on the right), and pl. XVI, 3 (with the farewell to Chandaka; cf. Fergusson, ibid., pl. LXXV, to the left). On all those stelae which are complete the Parinirvana is constantly symbolized by a simple slupa. With these one may connect others, in which the Abhisambodhi and the Dharmacakrapravartana are figured by an empty throne under a tree or a wheel (Burgess, ibid., pll. XXXVIII, 3 and 6 ; XLV, 2 and 4; XLVI, 1-3 ; XLVII, 3 ; XLVIII, I; Fergusson, ibid., pll. XCIII; XCIV, I, 4). The most curious of this kind are those which shrink from representing not only the Buddha, but even the Bodhisattva, and wherein the Mahabhinnishkramana is no longer represented, except by a horse without a rider (Fergusson, ibid., pll. XCIII, to the left; XCVI, 3 and XCVIII, 2). It will be observed besides that on several stelas of Benares (Anc. Mon. Ind., pl. 67, 2 and 68, I) the Mahabhinishkramana is associated with the Jati in the same framework.

149

Through the disconnected accounts of the documents we seem to catch the play of the two dominant forces which brought the number of the great pilgrimages up to the sacred figure of eight. Sometimes the preponderant element seems to be the prestige which a certain miracle had very early acquired in the popular imagination. Thus we see the «descent from heaven» separate itself very early from the crowd of traditional marvels; but its localization continues fluctuating, at least if we keep to the letter of the texts (1). On the other hand, the ancient capital of Magadha, Rajagriha, and the wealthy free town of Vaicali easily, by reason of their preeminent role in the Buddhist scriptures, eclipsed the titles of Kaucambi or Mathura: there is, however, no consensus of testimony as to which among all the edifying scenes which had there come to pass it was right more particularly to commemorate. At Cravasti even, where the interest is at once concentrated upon the Jetavana, the Master's favourite sojourn, unanimity of choice does not fall, as might have been expected, on the «great miracle», the triumph whereby its immediate environs had been rendered famous (2).

(1) It is known that the Divyavadana and Fa-hien localize the Devavatara at Sankacya, Hiuan-tsang at Kapitha and Fa-t'ien (cf. S. Levi, loc. cit., p. 190) at Kanyakubja; the Mahavyutpatti (§ 193) and Wou-k'ong (trans. S. Levi and Ed. Chavannes, Journal Asiatique, Sept-Oct. 1895, p. 358) do not give definitely the place of this «Descent» of Buddha.

(2) The Mahapratiharya is indeed mentioned by the text of Harsha and placed, somewhat incorrectly, by Fa-t'ien in the Jetavana of Cravasti (the Divyavadana [pp. 151 and 155,11. 12-14 and 17-18] specifies, in fact, that the theatre of the scene was situated between the town and the park); but Wou-k'ong associates with the Jetavana the preaching of the Mahaprajnaparamita-sutra. In the same way, at Rajagriha, in direct antithesis to the vague «teachings» of Fa-t'ien, he places the preaching of the Saddharmapundarika on the neighbouring hill of tne Gridhrakuta. At Vaicali both agree to call by different names the touching episode of the rejection of life (ayur° or ayuh-samskdra-utsarjana), which supervened three months before the Parinirvana. But we shall see that, guided by considerations of a pictorial and technical order, the artists made from the mass of the traditional accounts a quite different choice from the men of letters.

150

In the face of the capricious divergencies of the texts the concordant precision of the figured monuments has fortunately permitted us to make out the list, and to sketch the traditional scheme of the four supplementary great scenes, the miracle of Cravasti, the descent from heaven at Sankacya, the monkey's offering at Vaicali, the subjugation of the savage elephant at Rajagriha (1). It is true that, in order definitely to fix this scheme, we have availed ourselves chiefly of miniatures in Nepalese and Bengali manuscripts of rather late date (XI-XII centuries). At the most we had been able to compare with them only a few carved slabs, which came from the scene of the «first preaching», at Sarnath, in the northern suburb of Benares, and which date back approximately to the Vth century of our era. Unfortunately, these slabs were quite incomplete : we may be permitted, therefore, to emphasize the interest of the recent discovery at the same place of a stele in fairly good condition, divided into eight panels and consecrated precisely to the eight great scenes (pl. XIX, I). Let us say at once that seven of these bas-reliefs only confirm what we already knew of the subjects which they represent and the conventional manner of treating them. Besides, Mr. J. H. Marshall has completely identified them. He has no hesitation, except as regards one single scene, «of which the identification», he says, «is doubtful, but which appears to have taken place at Cravasti» (2). And it is, in fact, concerning the traditional manner of representing the «Great Miracle» of Cravasti that this new document will furnish us with useful evidence.

(1) See Et. sur l'Iconogr. bouddh. de l'Inde, I (1900, pp. 162-170), summarized, corrected and completed, ibid., II (1905), pp. 113-114.

(2) See Mr. Marshall's (article in J. R. A. S. 1907, pp. 999-1000, and pl. IV, I). We take pleasure in here thanking the very distinguished Director-General of Archaeology, India, for his extreme kindness in putting at our disposal a photograph of the stele in question and authorizing its reproduction.

151

The canonical importance of the mahapratiharya of Cravasti is incontestable. The Divyavadana gives it expressly as one of the ten acts of which every perfect Buddha must necessarily acquit himself before dying (1). It is likewise in this text -- that is to say, as MM. S. Levi and Ed. Huber have shown, in the Vinaya of the Mula-Sarvastivadins -- that we find the most ancient and most detailed account of the miracles whereby on this occasion Cakyamuni triumphed over his rivals, the six chiefs of sects. Thanks to the translation of Burnouf, this account is too well known to need citation here (2). We shall restrict ourselves to bringing out the essential points.

(1) Divyavadana, pp, 150-151; no Buddha of the past has failed in it ibid., p. 147,Il. 24-27); according to the Tibetan testimonies the Buddha of our age accomplished it in the sixteenth year of his ministry (Rockhill, Life of the Buddha, p. 79).

(2) Divyadana, XII and Bornouf, Introduction a l'hist. du Bouddh. indien, pp. 162 sqq. The XIIIth story of the Avadanakalpalata, deplorably edited indeed in the Bibl. Indies, V, 1895 (see, below, p. 174, n. 5), adds, in accordance with the usual custom of Kshemendra, nothing but poetic graces; Rockhil (Life of the Buddha, pp. 79-80, following the Dulva) and Schiefner (Eine Tibet. Lebensbeschr. Sakyamuni's, p. 293) restrict themselves to a reference to Burnoof. For the connections of these various authors with the tradition of the Mula-Sarvastivadins see also the very clear conclusions of Prof. S. Levi, Journal Asiatique, July-August 1908, pp. 102 and 104.

152

After having wrought a few minor miracles, which were mere preliminary trifles, and refusing to allow anyone, monk or layman, man or woman, to be substituted for him, so as to confuse the Tirthyas by the exhibition of a supernatural power, the Blessed One accomplishes successively, on the direct and twice repeated invitation of king Prasenajit, two kinds of miracles. At first he displays what in technical terms is called the yamaka-pratiharya, which consists in walking the air in various attitudes, while emitting alternately flames and waves from the upper and lower parts of his body; in the second place, multiplying images of himself up to heaven and in all directions, he preaches his law. A violent storm, raised by a chief of the genii, completes the overthrow of the heterodox. An immense multitude is converted to the good law.

If now after the Sanskrit version we consult the Pali tradition, we find that the maha-pratiharya of Cravasti is there usually designated «the miracle at the foot of the mango-tree». The Mahavamsa and the commentary of the Jataka, for example, give it no other name. According to the latter, as also according to the Singhalese and the Burmese (1) accounts, Buddha did, in fact, begin by accomplishing the magical operation which the jugglers of India are always endeavouring to imitate : from the stone of a mango planted in the ground he is supposed to have forth-with grown an enormous tree, covered at once with flowers and fruits. But then this is merely a simple extra, scarcely even a curtain-raiser.

(1) Cf. Mahavamsa, ed. Turnour, pp. 107, i8i, 191; ed. Geiger, pp. 137, 241, 254; Jataka, ed. FAUSB0LL, I, p. 77,1. 23 ; 88,1. 20, etc.; ambamule, or gandamba-mule, is written; Ganda has become in the commentary of the Jataka, no. 483, and Gandamba in Sp. Hardy (Manual of Buddhism, Ist ed. pp. 295-296), the name of the gardener who supplied the mango: see also Bigandet, Life of Gaudama, Rangoon, 1866, p. 205.

153

When the great day has come, the divinities assemble, and the introduction to Jataka no. 483 tells us in a single sentence that «The Master, having accomplished the yamaka-patihariya, and having recognized the believing dispositions of a great number of people, redescended, and, seated on his Buddha seat, taught the law» (1).

If we analyse this brief resume of the scene, it is not difficult to recognize in it, exactly as in the over-elaborate version of the Divyavadana, the distinct and successive enunciation of the same two moments, that of the «pair of miracles» and that of the preaching.

Of these two manifestations the first strikes one immediately as the more original and the more picturesque: one would have wagered that it must have thrust itself on the choice of the artists whose duty it was to decorate the Buddhist monuments with edifying scenes, or to compose pious ex-votos for the use of the laity. As a matter of fact, we have found in the ancient school of Gandhara at least one indubitable representation of the «twin miracles»; and even at the present moment the special attribution of this bas-relief to the maha-pratiharya of Cravasti seems to us not in the least untenable, on the sole condition that we mark well its exceptional character (2). It was, besides, the accidental circumstance of this find that prevented our carrying still further our researches on this point.

(1) See Jataka, IV, p. 265, 11. 13-14 ; the English translation (IV, p. 168, 1. 13) of oruyha == avaruyha by « then arose » seems to us to be a lapsus calami, going directly against the meaning. It will be noticed that the Pali, like the Maliavastu, makes use of the technical term of yamaka-p°; Bigandet, loc. cit., p. 207, gives us a very clear description of it (perhaps even two descriptions, cf. below, p. 157); it is also easily recognized through the tejo° and apokaswa-samapatti of Sp. Hardy, loc. at., p. 297.

(2) See Art g.-b. du Gandh., pp. 516 and 535, and fig. 263 (==Anc. Mon. India, pl. 115, 5), where we give the reasons which led us us to prefer this identification to the equally possible one of « the arrival at Kapilavastu ».

154

Nevertheless, as we had already observed, it was the scene of the manifold preaching of the Master that later, if we may judge from the miniatures of the manuscripts, inspired the traditional image of the «Great Miracle» : at least, they represent it regularly by three Buddhas teaching, seated side-by-side upon as many lotuses (1). Now the stele recently exhumed from the ground of the ancient Mrigadava testifies, five or six centuries earlier, to this same manner of conceiving the subject: the compartment which we know beforehand to have been reserved for the miracle of Cravasti shows us, in fact, like the miniatures, three Buddhas seated on lotuses in the attitude of teaching. This is the new fact supplied by this discovery, and it will not be long before its consequences unfold themselves before our eyes.

But, first of all, a question arises as to whether we must restrict ourselves to merely stating, or whether we can succeed in explaining, the unexpected choice of the artists. If we consider only the stele in question (pi. XIX, l), it seems that we may immediately see a reason, although an external one, for the course taken by its author. Let us observe, in fact, that of the four great supplementary scenes of the legend of Buddha there are two which absolutely necessitate a standing posture: they are the subjugation of the wild elephant and the descent from heaven. A legitimate care for symmetry in the alternation of the poses would have demanded a sitting posture in the corresponding scenes, not only in the monkey's offering, but also in the great miracle of Cravasti. Such, at least, is the idea imperiously suggested by an examination of the apportionment of the subjects on the new stele - the only one, let us remember, that we possess with the eight scenes complete (cf. the table below). It is scarcely necessary to remark that this reason, valid for the whole, is inapplicable to an isolated panel.

(1) See Icon. Bouddh. de l'Inde, I, pi. X, i (cf. Bendall, Catalogue of the Buddhist Sanskrit Manuscripts in the University Library, Cambridge, 1883, p1. II, i), and cf. ibid. p. 205, no. 83, and II, pp. 114, no. 4.

155

TABLE

First Preaching	Parinirvana
BUDDHA SEATED	BUDDHA LYING DOWN
Descent from Heaven	Miracle of Sravasti
BUDDHA STANDING	BUDDHA SEATED
Offering of the Monkey	Subjugation of the Elephant
BUDDHA SEATED	BUDDHA STANDING
Nativity	Perfect Enlightenment
MAYA STANDING	BUDDHA SEATED

The reading of the texts will furnish an argument of wider bearing. It does not, in fact, take us long to perceive that they use and abuse the yamaka-pratiharya. The general introduction of the Jataka makes it to be wrought by Buddha as early as the eighth day after the Bodhi, and specifies that he repeated it under three other circumstances, 1) at the time of his visit to Kapilavastu and of his meeting with his father and his relatives, the Sakyas, 2) at the time of his encounter with the heterodox monk Patikaputta, and last, 3) at Cravasti, at the foot of the mango-tree (1). The Divyavadana attributes it further to a simple monk; the Mahavastu to Yacoda or Yacas, the converted son of the banker of Benares; the Sutralankara to the five hundred bhikshunis, companions ot Mahaprajapati; the Jatakamala to a Pratyeka-Buddha;

(1) Jataka, I, pp. 77 and 88; trans. Rhys Davids, Buddhist Birth Stories,

pp. 105 and 123 ; on the first of the three other occasions cf. Mahavastu, III, p. 115, and on the second the Manual of Sp. Hardy, p. 331.

156

finally, the Mahavamsa twice places it to the account of simple relics of the Blessed One, etc, (1). We receive the impression that the yamaka-pratiharya has become hackneyed in consequence of being classic. Moreover when, after having accomplished it, Buddha returns and seats himself in his place, he informs king Prasenajit in a moment of proud modesty that «this kind of magical power is common to all the disciples of the Tathagata» (2). Hence it may be conceived that artists and worshippers were of one mind in no longer finding this banal wonder anything to characterize with sufficient clearness the great scene of Cravasti, and preferred the multiplication of the teaching images of the Master: for it is written that this last miracle is realizable only by the special power of the Buddha and the gods (3).

Finally, if we must conceal nothing, we seem to detect in the texts themselves a tendency to confuse the two kinds of wonders, and even to eliminate the former in favour of the latter. First of all, there seems to have been at times a misapprehension as to the real meaning of the expression yamaka-pratiharya. This technical term «twin miracles» does in fact lend itself to confusion.

(1) Divyavadana, p. 378 ; Mahavastu, III, p. 410 ; Sutralankara, trans. Ed. Huber, p. 399; Jatakamala, IV, 20; Mahavamsa, pp. 107 and 191 (tok-nouk), 137 and 254 (Geiger).

(2) Divyavadana, p. 161, 1. 13 : sarvacravaka-sad'harana. The text of the commentary of Jataka, n° 483, IV, p. 265, 11. 12-13 : Asadharanam savakehi yamaka-p°, which seems to mean the contrary, becomes in consequence most suspicious, at least if the two texts are speaking of the same miracle.

(3) Divyavadana, p. 162, ad. fin. The power of holding a dialogue with a magic double is likewise stated a little further (on p. 166, 1. n) as a privilege of perfect Buddhas only and inaccessible to simple cravakas.

157

We know now from the very explicit descriptions of the Divyavadana and the Mahavastu that it must be understood as the combined alternation of the two opposite wonders of water and fire: but it was not without reason that in 1880 Prof. Rhys Davids understood it to mean «making another appearance like unto himself». In the Burmese narrative translated by Bigandet (1) Buddha does, indeed, begin by making flames or streams gush forth alternately from the upper and lower parts of his body: but very soon he hastens to create a companion for his conversation and his walks, and sometimes it is his turn, and sometimes that of his double, to walk or to sit down, to question or to reply. It is curious to notice that the Divyavadana also makes the magically multiplied images of the Blessed One assume varied attitudes, and whilst some repeat afresh the marvels of water and fire, «others either ask questions or give answers to them». It even goes so far as to introduce most unexpectedly, as an ending to the chapter, a dialogue between Buddha and another self, created expressly for this purpose (2). Thus it manifests at least a certain propensity to amalgamate the two successive moments which it at first endeavoured to distinguish, and to confuse the reduplication of the miracles with that of the images (3). But this is not all. In another passage of the same collection the reverend Pindola Bharadvaja relates to king Acoka this same miracle of Cravasti, of which he represents himself as an eye witness.

(1) See Rhys Davids, Buddh. Birth Stories, p. 105, n. 4; and Bigandet, loc. cit., p. 207.

(2) Divyavadana, p. 162, 11. 17-20, and 166, 11. 3-11 ; cf. the description of plate XXI, 2.

(3) The same confusion seems to be reproduced with regard to the miracles attributed to the monk Panthaka ; as regards these last I am indebted to the obliging friendship and incomparably extensive information of Prof. S. Levi for the following references: Divyavadana, p. 494; Anguttara-Nikaya, I, 14 (p. 24); Visuddhimagga, analysed in Pali Text Society, 1891-1893, p. 114; Vinaya (Chinese) of the Sarvastivadins (c. n), of the Mahicasakas (c, 7), of the Dharmaguptas (c. 12), etc.

158

Now he no longer even mentions the yamaka-pratiharya: «And when, o! Great King, in order to triumph over the Tirthyas, the great miracle was accomplished at Cravasti by the Blessed One, and there was created an array of Buddhas which mounted up to the heaven of the Akanishtha gods, at that time I was there, and I saw these sports of Buddha (1)».

Here it is no longer a question of anything but the second miracle. Finally, we again find this latter, reduced to its most simple expression, in the Buddhacarita of Acvaghosha, whose descriptions are always so close to the figured monuments. According to him (so far as we can trust the English translation made by the Rev. S. Beal from the Chinese translation of the original Sanskrit) Buddha restricts himself to rising into the air and there remaining seated, and «diffusing his glory like the light of the sun, he shed abroad the brightness of his presence». In this version -- by a strange coincidence, but one which in our opinion it would be vain to seek to press further -- the maha-pratiharya quite assumes the characteristics of a Transfiguration: «His face did shine as the sun, and his raiment became white as the light». (2)

(1) Divyavadana, p. 4.01 (cf. Burnoff, Introd., p. 398) : it will be noticed that these are exactly the same terms as are employed on two occasions in the previously quoted sutra (Divyavadana, p. 162, 11. 16 and 26). P. 401, 1. 15, read probably aham instead of mahat.

(2) Sacred Books of the East, XIX, p. 240; Gospel according to St. Matthew, XVII, 2.

159

III

These waverings of tradition, as they are thus indicated in the texts, may help us to understand the at first somewhat surprising choice of the Indian image-makers. Regarding the fact of the choice itself there is, as we said above, no room for doubt. Let us resume the examination of the new panel, no. 5 of plate XIX, I: On a lotus, whose peduncle issues from a ripple of waves rolled into volutes, Buddha is seated with crossed legs in the hallowed posture, and his hands are joined in the gesture of instruction; on his right and left, again, there rises a padma with a long stalk, bearing another smaller Buddha, similar in all respects to the first... Now it is written in the Divyavadana that at that moment -- namely, at a second invitation from Prasenajit and when the first series of miracles was already accomplished -- «Buddha conceived a mundane thought».

Immediately the Gods rush forward to execute it: Brahma takes a place at his right and Cakra at his left, while the two Naga kings, Nanda and Upananda, create entire a wonderful lotus, on the corolla of which the Blessed One seats himself. Then by the force of his magic power, «above this lotus he created another, and on this one also a Buddha was seated with his legs crossed: and thus in front, behind, at the sides... » The crowd of Buddhas, holding themselves in the four consecrated attitudes (erect, walking, seated, or recumbent), soon rise to the highest heavens (1). The bas-relief, unable to juggle, like the text, with numbers and forms, shows us just three of them, all alike seated : but by now there is for us no question that we must see in this restricted space an attempt, however timid, to realize the legendary phantasmagoria.

(1) Cf. Divyavadana, p. 162. We know that the heaven of the Akanishthas is the highest heaven of the Rupadhatu, at the 23rd story of the Buddhist paradises. We remember also that the two kings of the serpents, «Nanda and his junior», play a part in a number of episodes in the life of Buddha, beginning with the bath which followed the nativity. We shall find information concerning them extracted by M. Ed. Huber from the Vinaya, of the Mula-Sarvastivadins in the B. E. F. E. O., VI, 1906, pp. 8 sqq.

160

With this abridged version we may connect immediately other more developed pictures, such, for example, as that which totally covers another stele originating from Sarnath (1), and in which are staged no less than four rows of Buddhas, seated or standing pi. (XIX, 2). On seeing the upward-branching lotus stems which bear these small figures, we might believe ourselves in the presence of a genealogical tree of Buddhas. Thus we are invincibly led to recall those which, either carved or painted, entirely cover great stretches of the walls of several of the subterranean temples of Ajanta. One of these frescoes, of which a copy has been published, very gracefully combines wreaths of flowers and foliage with the dreamy figures of the seated or standing Buddhas (2) : it decorates the wall on the right, in the antechamber of the sanctuary of Cave I; and no one will be surprised to learn that there it forms a pendant to another of the eight great scenes, « the Perfect Illumination », symbolized on the left wall by the Mara-dharshana.

(1) Again let us cite no. (Sarnath) I of the Calcutta Museum (Anderson, Catalogue, II, p. 4; Anc. Mon. India, pl. 68, i), the left upper division of which unfortunately broken) represents similarly the «Great Miracle» of Cravasti opposite lo the « descent from heaven » (cf. below, p. 164, n. I). It will be observed that on two other stelas of the same origin (Anc. Mon. India, pll. 67, 3 and 68, 2 : Art g. -b. du Gandh., fig. 209, and Iconog. bouddh. de l'Inde, I, fig. 29, to the right) analogous representations of the same miracle decorate the borders of the stone and enclose the scene.

(2) See Griffiths, The Paintings of Ajanta, pi. XV (cf. on the plan of the grotto the letter 0 and X, ibid., pi. IV and pi, VIII) and Burgess, Notes on the Bauddha Rock Temples of Ajanta, p. 17; the paintings of this cave are usually attributed to the VIth century. -- In Cave II the walls of the ante-chamber of the sanctuary are likewise adorned with figures of Buddha, of a very inferior make to those of the preceding ones. M. Griffiths counted 1055 of them, measuring about o m. 20 high and covering a surface of 22 square metres : he has reproduced some of them, pi. XXIV (cf. p. 28, and Burgess, loc. cit., p. 35, § XVIII ad fin.'). -- One may immediately connect with these frescoes the «thousand Buddhas» painted on the vault of the grotto no. I of Murtuk, a specimen of which Prof. Grunwedel has reproduced in his interesting Bericht iiber Archsologische Arbeiten in Idikittschari und Umgebung im Winter 1902-1903, pi. XXX: notice the strangely stereotyped character of the support of this Buddha, affecting both a cloud and a lotus.

161

The high reliefs of plate XX merely reproduce it in stone: in imitation of the painting the sculptor has not failed to fill the space between his characters with leaves and buds of pink lotuses, of the same kind even as those which bear his superposed rows of Buddhas (1). Only it will be observed that the stem of the seat of the central figure, at the bottom, is supported with both hands by two kneeling nagarajas, both wearing head-dresses of five serpent heads. As we have just been reading the Divyavadana, their names immediately occur to our minds : they are Nanda and Upananda. Thus we find ourselves in possession of an explanation satisfactory down to the details of the compositions. We have not, as was thought, to do with simple debauches of piously decorative imagery: we must here recognize representations on a vast scale, by reason of the space which the artist had at his disposal, of the «great miracle» of Cravasti (2).

(1) All the necessary particulars concerning this sculpture are given opposite to plate XX. In the Arch. Survey of Western India, vol. IV, pi. XXXVII, 2 (cf. ibid., p. 52), will be found a drawing of the opposite wall of the same vestibule of the sanctuary, with its eight rows of Buddhas, seven of which are rows of seven : the nagardjas are not missing.

(2) Is it worth while to observe that nowhere, either in these representations or in those considered above, have we found any trace of an attempt at an artistic realization of the fancies imagined by the editor of the Hien-yu-yin-yuen king? Never, in particular, do we see rays which open out into lotus bearers of illusory Buddhas burst from the «pores of the skin», or from the «navel», of the principal character, as is written in Schmidt's translation from the Tibetan Dzang-lun (Der Weise und der Thor, pp. 82 and 84).

162

This is indeed, if one reflects upon it, the only orthodox method of explaining the simultaneous presence of several Buddhas in the same picture, when an absolute law says that there shall never bemore than a single one at one time in each world-system. It follows that we must at the outset suspect the existence of this subject every time that we find ourselves in the presence of multiple images of Buddha: not, certainly, where they are isolated in separate sections or merely juxtaposed, but where they are evidently associated in the same action (1). If from this point of view we examine the reliefs and the frescoes of Ajanta, we shall not fail to discover a whole series of replicas, somewhat less prolix, but not less surely identifiable, than the preceding. Here we will restrict ourselves to citing the most typical of these variants. It seems that we shall have to look for them in the immediate neighbourhood of the inner sanctuaries:

"On the back wall, between the left chapel and antechamber [of the adytum of cave II], a large Buddha has seated himself under an amba (mango) tree with an Indra on his right and a Bodhisattva on his left (2).

(1) This restriction is necessary for three reasons. First, we must reckon -- with the progressive crowding together of images of Buddha on the facade or inner walls of the same sanctuary at the expense of various donors. -- Secondly, we must not forget the relatively ancient juxtaposition (cf. pl. XXVI, II) of the seven Buddhas of our age: but we are prepared to believe that there may be a close connection between this motifs and the « grand miracle », either because in the latter the Buddhas prefer to affect the number 7 in rows (ct. p. 161, n. i and p. 163), or because the representations of the « seven Buddhas » are strongly influenced by those of the maha-pratiharya (as is notably the case at Ajanta for the pl. XCI of Griffiths, in contrast to pi. LXI). Finally, we do not pretend do deny that at a fairly late period there may have been sought, in a mechanical repetition of images of the Master, an automatic accumulation of merit : but it is our opinion that the origin of this inept procedure must be sought in the single motif where its employment was canonically justified.

(2) Dr. Burgess, loc. cit., p. 34, § XVII : we think it necessary to make a choice and say; « between Brahma and Indra », or « between two Bodhisattvas » : but that can be decided only on the spot. Let us remark also that a mango-tree cannot be a Bodhi-druma. The letters E-F mark the place of this panel on the plan of Cave II, given by Mr. Griffiths, pi. XX. We must add that the fresco is approximately dated « by an inscription painted in the alphabet of the VIth century ».

163

His feet rest on a white lotus: a worshipper is below a little to the left. Across the top are seven Buddhas in various mudras, each on a lotus, the stalks being brought up from below. On each side of the Bodhi-druma, or sacred tree, are two Buddhas... Below these, on each side, were two pairs more », etc. We borrow this description from the notes of Dr. Burgess : it would not be possible to find a better one for the « great miracle » of Cravasti, including the mango-tree of the Pali tradition. It is again the same subject which in Cave XVIL on the right wall of the vestibule of the sanctuary, forms a pendant to the no less famous miracle of « the descent from heaven » (1); and this replica, unfortunately very much damaged, contains also a topical and rather exceptional detail : « The right end of the antechamber », says Dr. Burgess (2), a is painted with standing and sitting Buddhas; the lower portion, however, is destroyed, except a fragment at each end. The portion remaining at the right side is very curious, representing a number of Digambara Jaina bhikshus helping forward an old fat one, and carrying the rajoharana or pichi, a besom to sweep away insects, etc. Most of them are shaven-headed and stark naked. One or two, who wear their hair, are clothed. On the extreme left are an elephant and a horse with two men. The intermediate painting is completely destroyed ».

(1) The Devavatara is there represented in three stages, as on the pillar ot Barhut (Cunmngham. Stupa of Barhut, pi. XVII) : at the top is seen the « Preaching to the Trayastrimca Gods .:, in the middle the « Descent from Heaven », at the bottom the « Questions to Cariputra ». Only these last two episodes are represented on Griffiths' plate LIV ; for the plan of Cave XVII cf. his plate LIII.

(2) Loc. cit., p. 69, § XXXIII.

164

By now it is not difficult for us to recognize - exactly as on plate LXVIII of the Anc. Mon. of India (1) -- at the left at least an indication of the royal presence of Prasenajit, to the right the demoralized troop of Tirthyas, and doubtless the obese and naked old man, whose steps these have to support, is the Purana Kacyapa whom the Buddhist legend denotes as their leader and whose defeat is about to have for penalty an ignominious suicide (2). It is again he whom we believe we can identify on the left side of the new panel of plate XIX, I by his shaven head, his nakedness and, especially, by his strange backward posture, in striking contrast to the devout attitude of the Buddhist monk who forms a pendant to him on the other side. But, on the whole, representations of monks belonging to other sects are rather rare in Buddhist art, even where their presence would be most expected : and the pictures of the Master's triumphs willingly dispense with the not very edifying spectacle of the vanquished. It would be only the more desirable that we should possess a good reproduction of what is still to be found of this Ajanta fresco. Lacking this, we must content ourselves with giving a sketch of one of those which adorn the principal archway of Cave IX (pl. XXI, I). We know the curious aspect of that little subterranean chapel, with its three naves, its portal gallery and its stupa marking the position of the altar: the warm, ruddy tones of its frescoes give the finishing touches

(1) At the bottom of the upper compartment on the left we perceive, indeed, in addition to the two nagarajas who are holding up the stem of the central lotus, Ist, at the left of the spectator. King Prasenajit, who is recognizable by his parasol-bearer and his elephant, and 2nd, facing him, also seated upon a stool, Parana Kacyapa, in the form of a fat, naked man, with shaven head, who is supported from behind under the arms by one of his companions. We may connect with this type that of the same person in Art g.b. du Gandh. (fig. 261 and 225 c), and read, ibid., pp. 529 and 537, remarks on the rarety of these representations of « sectarians ».

(2) Divyavadana, p. 165

165

to the illusion of an ancient basilica. Above the pillars, where the triforium should be, ranges a series of paintings representing hieratic groups (1). One, almost complete, which is represented by our plate, has the advantage of uniting only the essential elements of the subject, namely, the three Buddhas with their feet placed on lotuses, and -- at each side of the one in the centre of the picture, who is teaching, and of whom the two others are, and can only be, illusory emanations -- the two traditional divinities, voluntarily reduced to the humble role of flyflapholders. Is it necessary to observe that this is exactly the same distribution of persons (2) that we find again on the lower row of plate XIX, 2?

All the specimens of which we have just been speaking, both from Benares and from Ajanta, can in bulk be dated, in accordance with the alphabet of the inscriptions on some of them, as of the Vth or VIth century of our era. We shall not hesitate, in spite of time and distance, to connect with them the numerous groups which decorate the principal wall of the highest sculptured gallery of Boro-Budur (IXth century). Almost the whole of this wall is covered with variations on the theme of the a Great miracle » of Cravasti; and this profusion of replicas is sufficiently justified by the enormous surface which the sculptors of the monuments had received instructions to decorate.

(1) Cf. Griffiths, Paintings of Ajanta, pll. XXXVIII and XXXIX.

(2) The only differences to be observed consist, Ist in the somewhat capricious detail (cf. p. 167) of the orientation of the acolyte Buddhas, turned or not towards the central Buddha ; 2nd in the fact that the latter has a lotus not for a seat, but only for a footstool. This kind of throne and this sitting position « in the European mode » are current peculiarities of the local style, although they are not unknown to the school of Benares and although we may have found them even so far as in the great Buddha of the Chandi Mendut near Boro-Budur in Java. They constitute all the less an obstacle to the proposed attribution since the central lotus, while treated as a simple little bench, is nevertheless usually supported by the two classical nagarajas (cf. for example, in Arch. Survey West. India, IV, pi. XXXVI, 2, the Buddha carved on the stupa of cave XXVI of Ajanta, and below, p. 168).

166

We content ourselves here with reproducing the group placed at the left of the eastern staircase, which we know was that of the facade (pl. XXII). On the other side an analogous group forms a pendant thereto, except that it is still more complex and contains no less than seventeen images of the Blessed One. The general arrangement of these compositions is a compromise -- doubtless imposed by the dimensions of the rectangular panels, which were much wider than they were high -- between the line taken by plates XIX, 2 and XX and that by plate XXI, I: but on one side or the other all the topical features are to be found. This Symmetrical reduplication of Buddhas, supported by lotuses and surrounded by divinities, suffices to establish not only the undeniable relationship of the schools, but also the fundamental identity of the subjects.

Inevitable, again, is the connection with many of the great rock-sculptures of northern China, less remote in time, but not less distant in space, from their Indian prototypes. We shall note especially, among the gigantic images which decorate the grottos of Ta-t'ong-fu (Vth century), those recently published by M. Chavannes, which, as he informs us, owe the possibility of their being so clearly photographed to the fact that the crumbling of the rocky facade has left them open to the sky (pl. XXI, 2).

167

The presence of a second Buddha standing at the left of the great seated one, -- the acolyte on the right has disappeared in the fallen debris -- is sufficient to recall the maha-pratiharya: and the innumerables figures of the Blessed One, superposed upon a kind of band, which form nimbuses and aureoles on the flamboyant background of the tejas, finally convince us that we have to deal with a representation of this miracle in the traditional form of the multiplication of Buddhas (l).

All these works of art, painted or carved, whether Chinese, Japanese, or Indian, represent more or less, in fact, -- to make use of the expression employed in literature, -- the vaipulya method of sculptured tradition. Let us return to our starting point, I mean to the quite summary lesson presented to us by the stele of the Archaeological Survey (pl. XIX, I): we shall see connected with it also a series of replicas no less sober than itself. A carving, which we believe to be unpublished, will furnish us with a type of them, at least as far as Magadha is concerned (pl. XXIII, I). A great Buddha, seated, in the attitude for teaching, on a lotus whose stem is flanked by two Nagarajas, is inserted between two other images of himself, with feet also resting on lotuses. The only novelty introduced is that the two acolyte Buddhas, instead of confronting the spectator, as in plate XIX, I, or being turned towards the central person, as in plate XXI, I, or slightly turned from him, as in plate XIX, 2, are looking in exactly opposite directions. This slab, of rather rude workmanship and late date (2), will serve as a perfectly natural transition to the miniatures of the Nepalese or Bengal manuscripts of the XI-XIII centuries,

(1) We should like to connect with these groups from Ta-t'ong-fu others somewhat later, which decorate the grottos of the pass of Long-Men (Ho-nan), of which also M. Chavannes has brought back photographs taken in the course of his last mission in China (see, already, T'oung Pao, Oct. 1908, fig. 4; cf. Journal asiatique, July-August 1912, figg. 1-4; Bull. Ecole fr. Extr.-Or., V, 1905, fig. 36) : but here the two acolyte Buddhas have been changed into two simple monks! The transformation might in strictness be explained by scrupulous orthodoxy (cf. above, pp. 161-162).

(2) For a reproduction of an analogous group, of the same provenance and likewise preserved in the Museum of Calcutta, see Et. sur l'Iconogr. bouddh. de l'Inde, I, fig. 28, where these three Buddhas are placed just below a representation of the Nativity.

168

where the representation of the « great miracle » of Cravasti by three Buddhas back to back has become the constant rule (1). The identification of our plate XXIII, I, which already flowed naturally from the analogy of the new stele of Sarnath, receives, on the other hand, an interesting confirmation in extremis from these latest indigenous manifestations of Buddhist art. Whilst definitely taking this turn in eastern India, it became in the West by degrees stereotyped under a form equally abridged, but sensibly different. The place occupied by Elias and Moses in the Christian pictures of the Transfiguration is now, in the representations of the Buddhist « great miracle », no longer held by the two acolyte Buddhas, but by two divine attendants. The imagery of the valley of the Ganges had reduced their part to almost nothing, or even omitted it entirely: here, on the contrary, they end by figuring alone at the side of the Master, standing on lateral lotuses and retaining in their hands their fly-flappers. As to the central Buddha, at one time he continues to sit in the Indian manner upon a padma like that of plates XIX-XX; at other times, and more frequently, he is installed on a throne after the manner of Europeans, as in plate XXI, I, and only uses the necessary lotus as a footstool: but nevertheless the two traditional Nagarajas continue to hold up its stem. We borrow from a mural sculpture of Kuda the most reduced type of the first variant (pl. XXIII, 2): a no less summary specimen of the second would be furnished by one of the caves of Kondivte (2). But, above all, we must recognize that all the cave-temples of western India are covered with representations of this kind.

(1) Cf. above, p. 154, n. i.

(2) See Burgess, A. S. W. I., IV, pi. XLIII, I, left part (cf. ibid., p. 71}.

Cf. the Fuller Replicas of Kanheri, ibid., fig. 22; Buddh. Art in India, fig. 60, and Cave Temples of India, pi. LVI (cf, ibid., p. 358), etc.

169

On this point it is sufficient to refer to the testimony, which no one will think of challenging, of Fergusson and Burgess. Along with them we might gather an ample harvest of replicas of the « great miracle ». If we do no undertake to draw up a list from their descriptions or from the too cursory notes which we formerly found occasion to take, it is because on these sculptures of a late period there is always reason to fear contamination of subjects (1).

We have followed up the evolution of the subject and its variants from the Vth century of our era to the final extinction of Buddhist art in India. Could we not now, after having brought the course of its history as far down as possible, endeavour to remount towards its origin and seek in the preceding schools, beginning with that of Gandhara, the prototypes of the monuments which we have just identified? The enterprise imposes itself upon us, and there seems to be no way of escape.

(1) In fact these contaminations have not failed to take place. The Buddha of the mahapratiharya of Cravasti makes the gesture of instruction, exactly as does the Buddha of the Dharmacakra-pravartana of Benares: nothing further "was required to provoque confusions and exchanges between the two motifs originally characterized, the one by the lotus with the Nagarajas, the other by the wheel with the gazelles. On plate 164 of Anc. Mon. India, by the side of the subject of our plate XXIII, 2, we find some « First Preachings » treated as « Great Miracles », except that the gazelles have replaced the Nagarajas on each side of the lotus; on the facade of the great temple of Karll (ibid., pi. i6S) the gazelles have even been intercalated above Naga-rajas! From this it may be conceived with what precautions we must surround ourselves before risking a firm identification from descriptions

170

Such fortunately is, so far as Buddhist icoriography is concerned, the routine force of tradition, that, in order to succeed in this second part of our task, it will suffice to determine with exactitude the distinctive feature common to all the verified representations of the mahapratiharya. Now, if you turn over the plates afresh, you will very soon observe that what characterizes them above all is the special form of this lotus « with a thousand petals (1), as broad as a chariot wheel, of solid gold, with a diamond stem », standing out entirely from the plinth. Whether supported or not by the two Nagarajas, whose masterpiece it is, it constantly serves as a throne -- or at least as a footstool -- to a Buddha seated in the attitude of teaching. By this sign we must henceforth retrospectively identify a whole series of Greco-Buddhist stelae, the greater number of which have already been published, but not explained, and which for the convenience of the reader we have here collected together before his eyes (pll. XXIV-XXVffl, i).

The most sober type (and the one which most closely resembles that of plate XXIII, 2) presents to us a Buddha, flanked simply, in addition to the usual worshippers, by two standing divinities (2), who, like him, are sheltered under parasols, adorned with garlands (pl. XXIV, I).

(1) Divyavadana, p. 162, 11. 9-n. Cf. the epithet of Buddha in Kshemendra's Dafdvataracarita, IX, 54: Bhunirgata-pratata-kancana-padma-prsthca-padmasinastha...

(2) We may connect with this group that of the British Museum, reproduced by Dr. burgess (Journ. of Indian Art and Ind. no. 62, 1898, pi. 8, 2 == Anc. Mon. India, pl. 92, in the middle): the teaching Buddha and the two divinities are seated, or standing, on the enlarged pericarp of a lotus flower. In the acolyte at the right we recognise Brahma by his head-dress and his water vessel, in the one on the left Cakra by his diadem. The two worshippers are withdrawn to the bottom of the stele and separated by what is usually the stalk of the central lotus, but is here treated as a pyre. -- We pay no regard to another image (that of the Calcutta Museum), likewise published by Dr. Burgess(J. I. A, I., no. 69, Jan. 1900, fig. 24 = Buddh. Art in India, fig. 112): here Buddha is indeed seated between the two worshippers on the characteristic lotus, but -- by an exception which, for the rest, is since the last excavations of Takht-i-Bahai (cf. below, p. 172, note I) not unique -- he is making the gesture of meditation, instead of that of instruction.

171

On plate XXIV, 2 we scarcely divine the suggestion of the lotuses on which rest the seat of the Master and the feet of his two acolytes : on the other hand, two other busts of the Blessed One are interposed in the hollows delimited by the lines of their shoulders : except for the exchange of place between the two gods and the two magical Buddhas, it is evidently the same group as on plates XIX, 2 (first row) and XXI, I. At other times the ingenious art of the sculptor erects graceful architectures (pl. XXV) above the three principal characters: doubtless we must here recognize the pratiharya-mandapa, built expressly for the occasion of the miracle; but we remain free to admire in it, together with the Mula-Sarvastivadins, the royal munificence of Prasenajit, or, with the Theravadins, the divine skill of Vicvakarman (1). At one time (2) it is a simple portico that presides above the three seated figures (pl. XXV, 2). At another time bolder constructions lodge beneath their domes or arches images of Buddha or even accessory episodes (pll. XXV, I and XXVI, I). On this last plate the two divinities, again standing, have each provided themselves with a long garland, which we shall find in their hands on all the reproductions that we still have to examine (pll. XXVI, 2-XXVIII, i). The latter, like those first cited, place the scene -- or rather, the vision -- in the open sky : at the most, they shelter some small figures under aerial aediculse. However, the number of divine spectators increases in a striking manner. Now they are placed one above the other on their lotus supports, profiting by all the liberty which a picture of apparitions allows to be taken with the laws of perspective.

(1) Divyavadana, p. 155, 1. 18 ; Jataka., IV, p. 265,1. lo.

(2) From the point of view of the arrangement of the attendants we may connect with this plate the fragment published by Dr. J. Ph. Vogel in Archaeol. Survey Report, 1900-1904, pi. LXVIII b (with the Nagarajas) and c.

172

At the same time the central Buddha becomes bigger, and his figure still more disproportionate to his surroundings. The garlands which used to hang above his head no longer suffice; there is now added a crown, borne by two little genii, with or without wings; once even other marvellous beings, with their busts terminating in foliage, hold still higher a parasol of honour. Lastly, among the images which have emanated from the Blessed One, some, as if better to emphasize their supernatural and magical character, are surrounded by an irradiation in the form of an aureole composed of other Buddhas (1). These specimens are more than are required to prove that we have not to deal with the fancy of some isolated artist, but, in reality, with a traditional subject, constantly reproduced for the edification, and at the request, of the faithful.

(1) See the two upper corners of plate XXVIII, i and compare fig. 78 of Art. g. -b. du Gandh., and especially the panel recently discovered by Dr. D. B. spooner at Takht-i-Bahai and published by Mr. J. H. Marshall in the J. R. A. S., Oct. 1908, pl. VI, 3. Here again we recognize the maha-pratiharya. The lotuses which once decorated the bottom of the slab have almost disappeared through the defacement of the stone; but it is not so with those which support the characters above, that is, five little seated Buddhas (three of whom are at the top among foliage), and the two divine garland-bearers. By way of an exception the principal Buddha affects the pose of meditation. The front of his parasol is curiously adorned with a crescent moon, doubtless in order to emphasize the aerostatic character of the miracle. But the point which specially holds our attention is the indication on each side of his body, between the knee and the shoulder, of four little Buduhas, standing on lotuses and arranged obliquely like the outspread feathers of a peacock's tail. -- It is known that Sir Adrel Stein found this procedure in use also on the sculptures of Rawak in Chinese Turkestan (Ancient Khotan, I, figg. 62-65 ; cf. Sand-buried Ruins of Khotan, frontispiece).

173

The series of these examples adjusts itself without effort in all its characteristic features - seat, attitude, gesture, surroundings of Buddha, etc. - to that in which we have already with certainty recognized versions of the « great miracle » of Cravasti. By virtue of the close relationship which we have often had an opportunity of noting between the Greco-Buddhist sculpture and the tradition of the Mula-Sarvastivadins we must more than ever appeal to the Divyavadana for information concerning the identity of the various personages. In the two « kings of the serpents », who at times support the stem of the great lotus (pll.XXV, 2; XXVII; XXVIII, I), we naturally continue to greet our; old acquaintances « Nanda and his junior », either accompanied or not by their wives. From these « fallen beings » we pass to the human bystanders. It has been asked whether the two lay devotees without nimbuses and of different sexes, who on plate XXVIII, i surround the seat of Buddha, are not merely donors of the stele (l). But it will be noticed that their point of support is, like that of the rest of the figures, the enlarged pericarp of a lotus : they appear, therefore, to form an integral part of the scene. For the same reason we must refuse to see in them anonymous worshippers : rather should we seek here -- exactly as in their kneeling counterparts on plate XXIV, I --

(1) This identification was proposed incidentally by Dr. J. Ph. Vogel, A. S. I. Rep., 1903-1904, p. 257 : but, in a general way, we believe it safer to look for donors only on the bases ofstelas (cf. pll. XXV, I; XXVI, I, and XXVII) or the pedestals of statues. - On the other hand, the hypothesis of Dr. Vogel (ibid, n. 3) which suggests the identity of the four nimbused figures seated on the lower row of the same stele (pi. XXVIII, i) with the four Lokapalas, seems to us most probable and confirmed by analogy with plates XXVI, 2, and XXVII.

174

that Luhasudatta and his wife, « the mother of Riddhila » (1), who in turn and in vain proposed to the Blessed One to accomplish the miracle in his stead. Likewise, on plate XXV, 2, the text expressly invites us to recognize in the monk and nun kneeling at each side of the Master the agracravika Utpalavarna (2) and the agracravaka Maudgalyayana, who also asked, and saw themselves successively refused, the same authorization. It is, then, these same four personages, rather than commonplace worshippers, whom we should prefer to recognize on plate XXIV, 2. We should be equally ready to find King Prasenajit, the impartial (3) president of this public manifestation : but, even where the number of spectators is increased, his royal equipage never appears, as later, to betray his incognito (4). In front of the four men of good caste seated at the bottom of plate XXVI, 2, it seems that we are rather, as on plates XXVII and XXVIII, I, in the presence of the four guardian gods of our terrestrial horizon. Among the crowd of divinities shall recognize immediately on plate XXVII, above the right shoulder of Buddha, his faithful companion Vajrapani, to whom also by certain texts a part is given in the story, he being made to intervene in order to hasten the denouement (5).

(1) On this upasaka and upasika information taken from the Vinaya of the Mula-Sarvastivadins will be found in the already quoted article of M. Ed. Huber (B. E. F. E.-O., VI, 1906, pp. 9 sqq.).

(2) For this title given to Utpalavarna, cf. for example, the commentary on the Dhammapada, ed. Fausboll, p. 213,

(3) For this impartiality cf. Divyavadana, p. 146, 1. 23.

(4) Cf. above, p. 164, note i.

(5) According to the Divyavadana (pp. 163-164) the yaksha-senapati who, understanding the impossibility of otherwise overcoming the obstinacy of the Tirthyas, raises a violent storm to disperse them, is called Pancika; but the Bodhisattvavadana-kalpalata calls him Vajrapani (XIII, 57). Only we must warn the reader that this stanza vasantatilaka, as it is given in the Bibl. Indica, I,v, p. 427, has no kind of plausible meaning. Prof. S. Levi has kindly restored the text for us, by the help of the Tibetan translation on the opposite page. It should read (the corrections are indicated by the italics):

Atrantare Bhagavatah satatam vipaksaw

Sarvatmana ksapanakow avadharya Yaksah

A^ipto^ravafavrtavarsavaraic cakara

Vidravya randhracaranan bhuvi Vajrapanih

We should translate: « In the meanwhile, perceiving that the Sectarians persisted in remaining obstinate adversaries of the Blessed One, the Yaksha Vajrapani, raising a violent storm accompanied by rain, dispersed them, and forced them to seek a shelter in the hollows in the earth ».

175

The feminine figure facing him would perplex us greatly, did not her crown of towers signalize her at once as the incarnate nagara-devata of Cravasti, an edified witness of the miracle which will hence-forth assure her fame; it is in no other form that, for example, the native town of Buddha is seen on other Greco-Buddhist bas-reliefs (1). But the most interesting feature to be observed is that, if we are to credit the Divyavadana, the two chief divine acolytes can be no other than Brahma on the right of Buddha and Cakra on his left. As a matter of fact, on several replicas the sculptors obviously emphasize this identification by the aid of the usual procedures of the school: to the much bejewelled turban of Indra they oppose, as is the custom, the chignon of Brahma, or they even endeavour to designate the latter expressly by the indication of a water-vessel or of a book (2).

(1) See Art g. -b. du Gandh., figg. 183-184 a, and p. 360.

(2) Cf. the procedure of distinction employed ibid., fig. 152, 154-156, 1640 (cycle of the nativity), 197 (march to Vajrasana), 212 (invitation to the preaching), 243 (preaching to the Trayastrimcas), 264 (descent from heaven), where we know that we have to deal at the same time with Cakra, the Indra of the Gods (cf. fig. 246), and with Brahma, the Cikhin. In the particular case with which we are concerned their positions are at times exchanged from one stele to another (cf. plate XXIV with plate XXV and p. 170, note 2), either because on this point the tradition was uncertain or because there had been a confusion,which is always easy, between the right and the left of the statue and those of the speculator.

176

It would take too long to enter further into the details of each variant; and besides on this point we may refer to the notices which accompany the plates: only, we should wish to be allowed to make three remarks of a general character. The first bears on the importance which already in the school of Gandhara we have been led to attribute to the lakshana, or sign of recognition : it seems indeed that here we find a fresh proof of the antiquity and wide extension of this proceeding (1). In this very case it is a lotus with a stem rising from the ground or from the waters, that serves as a distinctive mark for a whole series of monuments and has allowed us to follow the series for more than a thousand years, through the four corners of the peninsula. It is quite exceptional that, as on plate XXV, 2, the peduncle of the flower should be hidden and its pericarp covered by a cushion : and, if the artists of western India prefer that Buddha should cause his teaching to be heard from the height of a throne (simhasana), the typical padma is retained at least as a stool for his feet. Henceforth, therefore, we may rank this a « lotus emergent and usually attended by two Nagarajas », to use heraldic terms, side by side with, for example, the « wheel flanked by two gazelles, either back to back or face to face », among the specific symbols of the great events of Buddha's life. In the second place, this identification seems to us to confirm another rule which we had thought ourselves in a position to lay down, and in accordance with which there is scarcely any Gandharian bas-relief, however passive and motionless the characters therein may be, wich does not, even under the most strictly iconographic appearances, conceal the story of some episode in the legend of Buddha.

(1) Art gr.-b. du Gandh., p. 607.

177

We shall be the more readily excused for recalling the fact, inasmuch as we are the most to blame for having once ranged among the simply decorative motives, in default of finding a better place, several of the stelae which now assume for us a definite meaning and one of legendary value, as being versions of the « great miracle » at Cravasti (1).

But at the same time -- and this third observation is the most important of all -- it is to be feared that we must relinquish the idea of indubitably distinguishing, in the whole repertory of the Greco-Buddhist school, an iconolatric group of « Buddha between two Bodhisattvas ». As far as concerns the great scene of the descent from heaven at Sankacya, the texts had already forced us to recognize in the two divine acolytes of the Master the gods Brahma and Cakra. Here again ought not the same evidence to constrain us to accept the same identification? Then will disappear our last hope of discovering by the side of the Blessed One an Indo-Greek Avalokitecvara or a Manjucri, as plates XXIV, I and XXVI, I seem specially to invite us to do. In fact, all that we can say is that we believe we discern already on these stelae in the type, head-dress, attributes, meditative or pensive pose of the attendants the suggestion of the procedure which later served to represent,

and to differentiate from one another, the great Mahayanic divinities: but methodically we may not go further and light-heartedly oppose to the peremptory assertions of the texts any quasi-gratuitous conjectures. Even the sign of the urna, so distinctly marked on the forehead of the acolytes in plates XXIV, I and XXV, 2 fails to induce us to lay

(1) Cf. ibid., figg. 76-79 and p. 479.

178

aside this prudent reserve. So long as the sculptures do not furnish us with an image bearing a written inscription, the verbal statements of the Scriptures will always take precedence over their mute velleities of expression. Likewise, the more we advance in familiarity with the old artists of the north-west of India, the nearer are we to believing that the names of Avalokitecvara and Maujucri were as strange to their thought as to that of the compilers of the Divyavadana and the Mahavastu.

It will be felt how far this question passes beyond the limits of the present article, and we will not here insist upon it further. All that remains to ask ourselves, in order to complete the study of the representations of the mahapratiharya, is whether it was represented or not on the most ancient monuments of central India. Now it seems indeed that the old native school had already essayed in regard to it one of those conventional and summary pictures of which it possessed the secret. The pillar of the southern entrance in the railing of the stupa of Barhut has three of its faces decorated. Of the three upper bas-reliefs (1), the first represents, we believe, by the symbol of the Bodhi-tree, the « perfect illumination » ; the second, by the symbol of the stupa, the parnirvana; the third, by the symbol of the garlanded wheel, the « great miracle ». This, at least, is suggested by two inscriptions on the last named, from which we are not certain that all the admissible inferences have hitherto been drawn (see pi. XXVIII, 2).

(l) Cunningham, Stupa of Barhut, pl. XIII.

179

At the bottom a king issues from his capital, mounted in his quadriga: the epigraph, by informing us that he is called « king Prase-najit of Kocala », gives us at the same time the name of the town and localizes the scene at Cravasti. Now this king and his suite are going in the direction of a building of imposing appearance, which shelters a wheel surmounted by a parasol, and bearing a heavy garland suspended from its nave. For all students of ancient Buddhist art the allegory is clear : but, for fear the spectator should conceive the slightest hesitation, a second helpful inscription informs him that it is indeed « the wheel of the Law of the Blessed One » which is represented. The symbol, therefore, if translated into the style of the later schools, is the exact equivalent of an image of an instructing, and consequently converting, Buddha. On each side, standing in a devout attitude with joined hands, is a personage in splendid lay costume, such as India has always indifferently conceived its kings or its gods (1). Accordingly it is impossible for us in the presence of this group not to think of Buddha attended by Indra and Brahma, in the presence of this edifice not to think of the mandapa constructed for the purpose of the « great miracle ». Cunningham, with his accustomed instinct, has already connected with this bas-relief the passage in the Divyavadana translated by Burnouf, which does precisely on this occasion make the king of Kocala betake himself « in his good chariot » to the presence of the Master: but he did not follow out the identification to the end (2). In truth, we see no reason for stopping half way.

(1) For some quite similar images of gods on this same balustrade of Barhut see also Cunningham, loc. cit., pl. XVII.

(2) Ibid., pp. 90-91. - It will be noticed that the visit of Ajatacatru to Buddha, which on the pillar of the western entrance forms a pendant to this one, is likewise of importance from a legendary point of view (ibid., pi. XVI and p. 89).

180

Evidently it was not a question of an ordinary visit, but of a meeting having a solemn character. We know from a sure source, namely the inscriptions, the exact locality of the scene, that is Cravasti, the capital of Kocala, and the names of the two principal actors, Prasenajit and the Blessed One ; the bas-relief shows us the devout ardour of the one, and suggests the converting gesture of the other; finally, the accessory details of the two attendants standing beside the invisible Buddha and the great hall which shelters him

harmonize equally well with the traditions relative to the << great miracle ». We shall not escape the conclusion that such indeed was the subject which the sculptor had proposed to himself. The counter-proof is easy: let us imagine that precisely this task had been set him; granted the customary procedure of the old school, we do not see how he could have accomplished it otherwise (1).

Thus we should end by restoring to this subject of the maha-pratiharya the sphere which legitimately belongs to it and which until now had been too parsimoniously measured out. We are now in a position to sketch its history from the earliest to the last surviving monuments. Treated allegorically -- and with good reason -- by the old native school, it is not long in utilizing for its own advantage the type of Buddha created by Indo-Greek art.

(1) Again an interesting replica of our plate XXVIII, 2 will be found on plate XXXI, i, of Cunningham. We should be quite willing to connect with it the representations of wheels on pillars, like that of plate XXXIV, 4 (cf. at Sanchi, Fergusson, Tree and Serpent Worship, pl.XLII, I). Perhaps it would even be necessary to see a reference to the maha-pratiharya in the wheel which, according to the evidence of Fa-hien and Hiuan-tsang, surmounted one of the two columns raised at the entrance to the Jetavana.

181

From the outset it adopts that mudra of instruction (4) and especially that particular lakshana of the lotus with a stem, both of which it will retain as characteristic signs from end to end of its evolution. Under its most restricted aspect, as at Barhut, it counts only two attendant divinities: but on other replicas these latter multiply themselves and mingle with apparitions of Buddhas. It is chiefly these latter that are retained by the stelae of Benares, and, after their example, by the later productions from the basin of the Ganges, whilst western India to the very end reserves the best place for the divine acolytes. At the same time, the composition, which had finally on the vast walls of Ajanta attained a disproportionate development, returns, with the ultimate decadence, to the soberness of its commencements. All being taken into account, without going outside the Indian publications, and leaving aside the already identified miniatures of the manuscripts, we propose henceforward to inscribe the rubric of the « great miracle of Cravasti » under the following reproductions :

I. Barhut, pi. XXVIII, 2; Stupa of Barhut, pi. XXXI, i, perhaps XXXIV, 4, etc. (Ancient Indian style, 2nd century B. C.);

2. Gandhara: pll. XXIV-XXVIII, i; J. Ind. Art. and Ind., no. 62, 1898, pl. 8,2=Anc. Mon. India, pl. 92(in the middle); Arch. Survey Report, 1903-1904, pi. LXVIII, b and c; Art g.-b. du Gandhara, fig. 78; (with an exceptional mudra) J. I. A. I., no. 69, 1900, fig. 24= Buddh. Art in India, fig. 112, and J. R. A. S., Oct. 19o8,pl. VI, 3 (Indo-Greek style, Ist and 2nd centuries A. D.);

(1) For the only two exceptions known to us cf. p. 170, n. 2, and 172, n. 1.

182

3. Benares: pl. XIX; Anc. Mon. India, pl. 68, I (in the left upper compartment); (on the lateral borders) 67, 3, and 68, 2 (Gupta Style; 4th-6th centuries);

4. Ajanta: pll. XX-XXI, I; Paintings of Ajanta, pll. 15, 24, 39; Arch. Survey. West. India, IV, pi. XXXVII, 2 (Calukya style, 6th-7th centuries);

5. Magadha: pl. XXIII, I; Et. sur l'Iconogr. bouddh. de l'Inde, I, fig. 28(Pala style, 8th-10th centuries);

6. Konkan: pi. XXIII, 2; Arch. Surv. West. India, IV, pi. XLIII, j, and fig. 22 = Buddh. Art in India, fig. 60; Cave Temples of India, pl. LVI(Rashtrakuta style, 8th-10th centuries).

Henceforward the picture of the maha-pratiharya would not be missing from any school: we await only that of Mathura. This is just what might be expected from the importance assumed by the episode in the legend, as a compulsory prodigy of every « Blessed One » worthy of his name. It would have been too astonishing, considering the constant parallelism between the two forms, written and figured, of the tradition, if no ancient illustration had corresponded on this point to the texts. Our hypothesis fills a real gap; and it is only just that « the great miracle of Cravasti » should advantageously, as far as the number of known replicas is concerned, bear comparison with the three other great scenes from the teaching career of Buddha.

Why then -- and this is the last point on which we are conscious of owing the reader some explanation -- why has it been so tardily and so laboriously recognized, whilst its three pendants were identified long ago and at first sight?

183

To this question we may reply, first of all, that the maha-pratiharya, especially in the preaching form which had prevailed, does not lend itself, as we have abundantly experienced, to anything more than a picture almost void of movement, if not of picturesqueness; to effect its instant recognition, it has neither the exceptional role of the monkey or the

elephant, nor the characteristic decoration of the triple ladder : and here we have, doubtless, an excellent reason.

There is room, in our opinion, for adding another. We are so accustomed to utilize the archaeological information of the Chinese pilgrims in India, that we no longer think of being grateful to them for it; in order to measure the value of their help, we have to be once without it. That is the case on this occasion: Fa-hien and Hiuan-tsang, so explicit as regards the three other episodes, scarcely mention the one which interests us here. The places where Cravasti and the Jetavana had been, the favourite sojourn of the Master, evoked too many remembrances pell-mell for the « Great Miracle » not to be swamped in the crowd of those which on all sides, through the mouths of the guides, solicited their devout interest. We must likewise reckon with the fact that the story of the rivalry between the Master and the Tirthyas was on the spot inevitably entangled with the calumny of the noviceCinca, or with the assassination of the courtesan Sundari: and these dramatic stories could not tail to encroach upon the miracle of Buddha, which was after all so neutral and quasi-passive.

Thus, when the pilgrims finally arrive at the temple which marked the locality of the purely doctrinal and magical conflict, they both specify indeed that a statue of the Blessed One was seated (1) there;

(1) We believe, in fact, after careful reading, that the mahd-caitya of Cravasti, marking the locality of Buddha's victory over the other chiefs of sects, was the temple (vihara), 60 or 70 feet high, -which Fa-hien and Hiuan-tsang both saw and mentioned at the west (that is to say, at the right) of the road leading to the south of the town towards the Jetavana, about 60 or 70 (Chinese, therefore double) paces in front of the eastern gate of the park, opening from the same side upon the same road (trans. Beal, I, p. xlvii, and II, p. 10; Watters, I, p. 393). It will be noticed that this situation corresponds fairly well with the indications of the texts (cf. above, p. 149, n, 2): it seems that it is expedient to set aside in its favour the « preaching hall » built by Prasenajit, which was to be found in the centre of the town, and the stupa next to that of Cariputra, which is mentioned by Hiuan-tsang only. As regards the latter, Walters states that he did not know where to place the « tope » of the « great miracle »; he forgets that the eight great caityas are not necessarily all siupas; we know, for example, that that of the Sambodhi at Bodh-Gaya is a temple, and the same is explicity told us by Fa-hien and Hiuan-tsang concerning the Devavatara.

184

but they both forget to tell us on what kind of seat and accompanied by what attendants. Accordingly, do not ask why the connection between the narratives and the representations of the « Great Miracle » has been so tardily realized. Cease likewise to be astonished that we are still even at the present time posed by the question whether the two divine acolytes retained to the very last (as we are certain they did in the representation of the « Descent from Heaven ») their names of Brahma and Cakra, or whether they ended by transforming themselves, in the eyes of the faithful, into Bodhisattvas, and, in that case, at what moment the transformation took place. Fa-hien and Hiuan-tsang tell us nothing concerning this. One feels how valuable their testimony would have been to us, by reason of its mean date as also of the central situation of the country from which they would have borrowed its elements, forming a bridge between the ancient works of the northwest and the later, but identified, productions of eastern India. If we have been able ultimately to dispense with it, this is because the stele recently discovered at Sarnath and immediately published by Mr. Marshall put into our hands precisely the missing middle of the conducting wire, and thenceforward all that we have had to do has been to follow its direction, downwards to the disappearance, upwards to the sources of Buddhist art. For this let us thank the Archaeological Survey!

CHRONOLOGICAL TABLE OF PLATES XIX-XXVIII

Plate No	Origin	Approx. Date
XXVIII, 2	Barhut	IInd Century B.C.
XXIV-XXVIII.I	Gandhara	Ist-IInd Centuries A.D.
XIX	Benares	Vth Century A.D.
XXI.2	Ta-t'ong-fu	Vth Century A.D.
XX-XXI.I	Ajanta	VIth-VIIth Centuries A.D.
XXIII.2	Kuda	VIIIth Century A.D.
XXII	Boro-Budur	IXth Century A.D.
XXIII	Magadha	Xth Century A.D.