p. 1 Many Indian fables are anterior to Buddhism. Some of then must be anterior even to the invasion of India by the Aryans.(1) The apostles of Buddhism had quickly realized the utility of these tales in appealing to the heart of man, and they had adapted the old fables to the needs of their propaganda. Such is the origin of the Jataka, and other moral tales, which, on account of their antiquity and uncertain sources, did not take long to be considered as "the sayings of Buddha" (Buddhavacanam) and have been classed as such in the Sutta-Pitaka after the codification of the Scriptures. Other tales, which are less ancient, have a different character. They are no longer popular tales handed down orally from generation to generation, but literary compositions which were written by a known author and the paternity of which could not in consequence be attributed to Buddha. What was to be done with these new productions? One --------------------- 1. On the subject of Austro-Asiatic elements contained in Indian folklore,cf. Le prologue-cadre des Mille et une Nuits et le theme du svayamvara, JA, 1924, II, pp.101 ff.; and La Princesse a l'odeur de poisson et la nagi, Etudes Asiatiques publiees a l'occasion du 25e Anniversaire de l'Ecole Francaise d'Extreme Orient, II, pp.265 ff. p. 2 could either exclude them from the Canon, or include them in one or the other Basket (Pitaka). The alternative adopted has been different according to the schools and the sects. Let us examine that which the Sarvastivadins have chosen. I purposely pass over the sectarian ramifications of this impor tant group, and I confine myself to a consideration of the two principal schools, viz., that of Mathura and that of Kasmir. One reads in the Ta tche tou louen, which is the Chinese translation of a commentary (attributed to Nagarjuna) on the Pancavimsati-sahasrika- prajnaparamita, (1) "That which, is called Vinaya is (a statement) of the faults committed by the bhiksus: according to the precepts laid down by Buddha, this must be done; this must not be done; in doing this one commits such a fault. The abridged statement is in eighty chapters. There is besides a second part. On the one hand, in the Vinaya of Mathura, the Avadanas and the Jatakas are in eighty chapters. On the other, the Jatakas and the Ava-danas are excluded from the Vinaya in the country of Kasmir (Ki-Pin). (The latter) does not contain anything more than the essential, which has been divided into ten chapters; (but) there is a vibhasa in eighty chapters which comments on it." It appears that the Vinaya of Mathura and that of Kasmir differ considerably. At Mathura, the Vinaya, properly so called, was in eighty chapters, while at Kasmir we have to distinguish between a text in ten chapters and a vibhasa eight times longer. Again, while the fables (Jatakas and Avadanas) remain excluded from the Vinaya of Kasmir those tales constitute at Mathura a collection as voluminous as the Vinaya itself. Several facts support the indications furnished by the Ta tche tou louen. The Mahavastu, the complete title of which is Mahavastv- --------------------- 1. Cf. Tripit., Tokyo edition, xx, 5, p.105, col.2. This text was kindly communicated to me by M. Pelliot and I have published a first translation of it in my Legends of the Emperor Asoka (p.214). It is this translation that I reproduce here with modifications in certain places. p. 3 avadana,(1) actually contains a large number of fables, avadanas and jatakas. We know that this collection was attached to the Vinaya-pitaka of the Lokottaravadins of the Mahasanghika School. The Tibetan catalogue of Tanjur supplies a fact of the same kind for the Sarvastivadin School. Volume XC of Mdo'grel contains, among other pieces: No. 17 Suvarnavarnavadana No. 18 Kunalavadana No. 19 Aryanandimitravadana No. 20 Saptakumarikavadana Now, according to the index (f.129, 2-3) cited by Cordier (Bstan 'gyur, III, p.416), the above four avadanas are mentioned to be included in the cycle('khor) of the Vinaya-Pitaka ('dul-ba'i sde-snod). In short, the Vinaya-Pitaka of the Kasmir School, like that of most of the sects, contains nothing more than the texts of Discipline, while the Vinaya-Pitaka of Mathura, like that of the Lokottaravadins of the Mahasangha, contains in addition an important collection of fables in eighty chapters. According to the Ta tche tou louen, these fables were of two kinds, jataka and avadana. They perhaps formed two series, which I propose, in order to fix the ideas, to call avadanamala and jatakamala. We should thus have, in the Vinaya-Pitaka of the Sarvastivadins, the equivalent of the two collections, Jatakamala and Pratyekabuddhamala, which, according to the Report of Nandimitra, were, in another Canon, attached to the Abhidharma-Pitaka.(2) What was the original word which the translator of the Ta tche tou louen has rendered in Chinese by pou(class, group or category) and which I have provisionally translated ---------------------- 1. M. H. Zimmer insists rightly on this point: Zum Mahavastu-Avadana in Zeitschr. f. Ind. u. Iran,, vol. III, pp, 201 ff: 2. Sylvain Levi et Ed. Chavannes, Les Seize Arhat Protecteurs de la Loi, JA, 1916, p.20 of extract. p. 4 by "chapter"? We are told that the Vinaya-Pitaka of Kasmir was in ten pou's. On the other hand, we know that the Vinaya of the Sarvastivadins was called the "Vinaya in ten narratives (adhyaya)" in Chinese: che song liu. It is therefore probable that the ten pou's of the Vinaya of Kasmir were adhyayas or "narratives." We have to remember here that the Divyavadana, which likewise belongs to the literature of the Sarvastivadins, is found in several recensions, of which at least one has preserved the ancient division into adhyayas. Ms. 88 of the Bibliotheque Nationale of Paris contains a Divyavadanamala which consists of only twenty-one avadanas distributed among thirty adhyayas.(1) On the other hand, the printed Divya vadana contains thirty-eight avadanas, of which only one, the Sardulakarnavadana, is still divided into adhyayas. Divyavadana and Divyavadanamala are so entirely different that we cannot draw any conclusions regarding the contents of the collection which is doubtless their (common) source. But it does not seem improbable that both of them were attached to the ancient avadanamala of the Sarvastivadins. The epithet "divya" given to the collections actually known implies that this collection stood on a higher level than the others. Its success and its popularity sufficiently account for the rehandlings which it has undergone in the course of the centuries, and consequently also for the diversity of our collections. The Divyavadana which contains a Mahayana sutra(2) has probably been rehandled lately. The Avadana mala of Mathura perhaps contained the four avadanas of Mdo 'grel, three of which are not to be found at the present day in our Divyavadana. At first sight, the above considerations seem to be in contradiction with the very plausible opinion, according to which the Vinaya of the Mulasarvasti- vadins has been ---------------------- 1. Cf. Divyavadana, edited by Cowell and Neil, Appendix C. 2. The 34th avadana, entitled Danadhikaramahaya- nasutra. p. 5 compiled in Kasmir.(1) This monumental Vinaya contains a large number of avadanas, but we also find that the Vinaya of Kasmir discarded, according to the Ta tche tou louen, the avadanas. The diffculty is not insoluble. The Vinaya of the Mulasarvastivadins seems to be a sort of reservoir, into which have flowed all the currents of the Sarvastivadin literature. It is the Total in which have been doubtless incorporated, by the side of an older Vinaya, the Avadanamala and the Jatakamala of the Sarvastivadins. We should not therefore be surprised to find in it pieces of diverse character and especially some important tales extracted from the Divyavadana. MM. Edouard Huber and Sylvain Levi, the first authors who have noted the pieces common to the Vinaya of the Mula sarvastivadins and the Divyavadana hold that the first work was the origin of the second. Quite recently, again, M. Sylvain Levi has written that the Divyavadana is "a collection of tales sliced out, almost all of them, from the huge Vinaya of the Mulasarvastivadins." (JA, July-September, 1927, pp. 103 ff). This view is probably incorrect. It does not appear that the Vinaya is the source of the Divyavadana. The compilers of the Vinaya of the Mulasarvastivadins are likely to have borrowed from the Divyavadana, or, to be more precise, from the ancient collection of fables of which our Divyavadana is only a late recension. ---------------------- 1. The North-west of India in the Vinaya of the Mula- sarvastivadins, JA, 1914, II, p.493 ff.