INTRODUCTION. as regards the limited number of stanzas explained by him, his text was verbally identical with ours. The frequent statements of the Brāhmaṇas concerning the number of verses contained in a hymn or liturgical group agree with the extant text of the Rgveda. The explanatory discussions of the Brāhmaņas further indicate that the text of the Rgveda must have been regarded as immutably fixed by that time. Thus the Satapatha Brāhmaṇa, while speaking of the possibility of varying some of the formulas of the Yajurveda, rejects as impossible the notion of changing the text of a certain verse of the Rgveda as proposed by some teachers. 3 Probably soon after the completion of the actual Brāhmaṇas the hymns of the Rgveda were fixed in the phonetic form of the Samhitā text; and after no long interval, in order to guard that text from the possibility of any change or loss, the Pada text was constituted by Sakalya, whom the Araṇyakas or appendixes to the Brāhmaṇas, the Nirukta, and the Rgveda Prātiśākhya presuppose. By this analysis of the Samhita text, its every word, stated in a separate form as unaffected by the rules of euphonic combination, has come down to us without change for about 2,500 years. The Samhita text itself, however, only represented the close of a long period in which the hymns, as originally composed by the seers, were handed down by oral tradition. For the condition of the text even in this earlier period we possess a large body of evidence corresponding to that of Mss. for other literary monuments. It was then that the text of the other Vedas, each of which borrowed extensively from the Rgveda, was constituted. With each of them came into being a new and separate tradition in which the borrowed matter furnishes a body of various readings for the Rgveda. The comparison of these variants, about 1200 in number, has shown that the text of the Rgveda already existed, with comparatively few exceptions, in its present form when the text of the other Vedas was established. The number of instances is infinitely small in which the Rgveda exhibits corruptions not appearing in the others. We have thus good reason for believing that the fixity of the text and the verbal integrity of the Rgveda go several centuries further back than the date at which the Samhita text came into existence. As handed down exclusively by oral tradition, the text could hardly have been preserved in perfectly authentic form from the time of the composers themselves; and research has shown that there are some undeniable corruptions in detail attributable to this earliest period. But apart from these, the Samhitā text, when the original metre has been restored by the removal of phonetic combinations which did not prevail in the time of the poets themselves, nearly always contains the very words, as represented by the Pada text, actually used by the seers. The modernization of the ancient text appearing in the Samhitā form is only partial and is inconsistently applied. It has preserved the smallest minutiae of detail most liable to corruption and the slightest differences in the matter of accent and alternative forms which might have been removed with the greatest ease. We are thus justified in assuming that the accents and grammatical forms of the Rgveda, when divested of the euphonic rules applied in the Samhita text, have come down to us, in the vast majority of cases, as they were uttered by the poets themselves. Though the tradition of nearly all the later Samhitās has in a general way been guarded by Anukramaṇīs, Prātiśākhyas, and Pada texts, its value is clearly inferior to that of the Kgveda. This is only natural in the case of DURGA), Calcutta 1882-91 (Bibliotheca | ¹ See OLDENBERG, op. cit., 352. 2 See OLDENBERG, op. cit., 380 f. 1*
