The following essay is the 5th and last part of the first essay in the book: The Beginnings of Buddhist Art and other essays in Indian and Central-Asian Archeology by A. Foucher, who herein discusses the important aniconic phase of Buddhist art, giving a reasonable thesis for its production and development. I have illustrated the essay with some files from Wikimedia. It was originally published in French in 1911, and later translated by L.A. Thomas and F.W. Thomas
This is not all. The sculptors of the second century verify our hypothesis not only in what they reproduce and in what they imitate of the works of the past: we may maintain that they do this, also, indirectly, in what they innovate. However unreflecting and mechanical their submission to custom may have been, the forced absence of the protagonist from the scenes of his own biography could not help but inconvenience them considerably. Let the career of the Blessed One be no more than a monotonous tissue of conversations more full of edification than movement; yet only a small number of episodes allowed of being portrayed independently of the principal personage. With the aid of what subjects were the artists to cover the numerous medallions, the long stretches, or the high gates of the Stūpa balustrades?
Buddha Meditating (Gandhāra)
The first expedient of which they  bethought themselves was to turn to the previous existences of the Master, at the time when under all animal forms, and later under all social conditions, he was qualifying by means of perfections for the final attainment of the Bodhi. Thereby we explain why the sculptors of Barhut preferred to dip into this treasure of tales and fables. In treating this new matter they were no longer trammelled, as when illustrating the last life of the Master, by a custom which had been elevated into a law. Accordingly they have no scruples in representing the Bodhisattva in each scene, and it is with a perfect liberty of mind that, at the time of his penultimate terrestial existence, they give to Viśvantara the features which they so jealously abstained from lending to Siddhārtha (cf. above, p. 6). Representations of Jātakas are far from being unknown at Sāñchī: but the decorators of the gates had recourse once again to another stratagem in order to slip between the links of tradition.
It goes without saying that in all the scenes posterior to the Parinirvāṇa the absence of the figure of the Blessed One became perfectly justified and at the same time ceased to be an inconvenience to the artist. Thus, they soon took pleasure in cultivating this part of the Buddhist legend. According to all probability they began by illustrating the famous “war of relics”, which the death of the Blessed One nearly precipitated. Encouraged, apparently, by this trial, they did not fear to attack even the cycle of Aśoka, and to represent at one time his useless pilgrimage to the stūpa at Rāmagrāma, and at another his solemn visit to the tree of the Sambodhi. Thus, under the pressing incentive of necessity, the native school, incapable of openly shaking  off its slavery, had artificially created for itself a double means of escape, in the legends previous to the last renaissance or posterior to the final death of Buddha. For our part, we do not doubt that, if it had continued to develop normally and according to its own rules, we should have seen the number of these sham historical pictures or these illustrations of popular stories increase at the expense of the old fund of pious images.
It is no longer a secret to anyone that the regular sweep of this evolution was brusquely interrupted by a veritable artistic cataclysm. The Hellenized sculptors of the northwest, strangers to the native tradition of Central India, satisfied to the full; and even outwent, the wishes of their Buddhist patrons by creating for their use the Indo-Greek type of Buddha. Immediately their colleagues of the low country, seduced by this wonderful innovation, greeted with no less enthusiasm than the laity the rupture of the magic charm which had weighed so heavily and so long upon the ancient Buddhist school. We have already remarked upon the fact of the rapid diffusion of the new type (p. 7): it is now clear to us that its adoption did not come into direct collision with any dogmatic prejudice. Always docile interpreters of current ideas, the texts set themselves henceforth to guarantee, by the aid of apocryphical traditions or an abundance of miracles, the authentic ressemblance of those portraits whose possibility they were a moment ago denying. 
The reason is  that, in reality, the new mode did not expressly infringe any ritualistic prohibition: it did nothing but overthrow the artistic procedures of composition, and the bonds which fell were of a purely technical kind. We have seen clearly enough how the image-makers of the basin of the Ganges had slowly suffered the spider’s web of custom to weave itself around them, and how, not daring to tear it apart, they had already endeavoured to free themselves from it. Under the stroke of the revelation which came to them from Gandhāra their emancipation was as sudden as it was complete: but even through this unexpected development we are prepared to follow up the test to which we have submitted our theory and from which it seems to us to have so far issued with honour.
The history of the ancient regime in Buddhist art prior to the Gandhārian revolution may, in fact, be summed up somewhat as follows. We have every reason to suppose that there was, first, from the fifth century onwards, local production at the four great centres of pilgrimage, and conveyance into the interior of India, of rude delineations copying the “sacred vestiges” actually still visible above ground in the sites of the miracles. It was these naturally unpeopled tableaux which, thanks to time and distance, ended by being regarded as systematic representations of the four principal episodes in the life of the Blessed One, and which, joined to some routine variations composed in accordance with the same formula, served, before as well as after Aśoka (middle of the third century B.C), for the decoration of religious foundations; finally, on the monuments of the second century (still before our era)  we remark already tentatives towards freedom from the tyranny of the ancient customs by recourse to subjects previous or subsequent to the last existence of Buddha.
However, the school of the north-west comes on the scene. By reason of the very fact that it has been almost entirely removed from these traditional influences, it must, in our system, present characteristic signs quite different from those of the ancient school. Now, the conclusions of an extensive study which we have long dedicated to the Greco-Buddhist bas-reliefs, seem to have conspired in favouring, point for point, the reverse of the preceding propositions. What we have observed at Gandhāra is, first, the almost total disappearance of legendary scenes later than the cycle of the Parinirvāṇa, as also a marked diminution in the number of Jātakas; in the second place, there is an indefinite multiplication of episodes borrowed from the youth or the teaching career of the Master, whose corporeal image occupies now the centre of all the compositions; finally and correspondingly, there is an extreme rarity of symbolical representations. 
In any case – and this is our concluding argument – the old emblems do not disappear completely. Not only at Gandhāra, but even on the latest productions of medieval India, not to mention the Lamaist images of the present day, these survivals of a former age continue to manifest themselves. If the stūpa is regarded as having on nearly all the new representations of the Parinirvāṇa become superfluous, the Tree of Knowledge never falls to rear itself behind the Buddha of the Sambodhi, whilst the wheel between the two gazelles, either back to back or face to face, continues to mark the throne of his First Preaching. And thus  the decline of Buddhist art is linked to its most distant visible origins, the only ones (need we specify?), which have been taken into consideration here. 
Such, at least, is the theory which we could not refrain from submitting to the appreciation of Indianists. Taken altogether, it is only an attempt at synthesis, an effort first to coordinate logically, then to organize in accordance with the laws of an historical development, a series of facts already known. In this sense there is not one Buddhist archeologist, commencing with Fergusson and Cunningham, who has not contributed to it, and it may be found more or less devoid of originality. Our whole ambition would be precisely that it should give, when read, the impression of being already public property. That would be the best of symptoms; for none is better adapted to produce a belief that – except for the retouches which the progress of research will inevitably give to it – it is destined to endure.
- By apocryphal tradition we mean those relative to the statue of sandal wood, carved even during the life-time of Buddha and attributed by Fa-Hian (trans. Legge, p. 56) to Prasenajit of Śrāvastī, and by Hiuan-tsang (trans. Stan. Julien, I, pp. 283 and 296) to Udayana of Kauśambī, whose example had only been imitated by Prasenajit (cf. Beal, Records, I, p. XLIV and 235; II, p. 4). As regards the miracles, see those which are related to us concerning the image of the temple of Mahābodhi by Hiuan-tsang (trans. Stan. Julien, I, p. 465; Beal, II, p. 120) and Tāranātha (trans. Schiefner, p.20). ↩
- Cf. Art Greco-bouddhique du Gandhāra, pp. 266, 270, 427 etc. ↩
- Cf, Art greco-bouddhique, fig. 208 and 209 and Iconographie bouddhique de l’Inde, fig. 29 et 30: the latter is a representation of the Parinirvāṇa, still surmounted by a stūpa. ↩