The text of the Atharvaveda is preserved only in the Saunaka school. Its Sanhita consists in the present edition of it, of 20 Kein'd'as, or books. Of these, the first 18 are subdivided into 34 Pea pit' halms, or chapters, with, altogether, 94 Anuvdkas, or sections, each containing a number of mantras (the 17th Kan'd'a consisting of a single Prapil fhaka). The 19th Kan'd'a is not divided into Prapanakas, hut into Anurdkas, of which it contains seven; and the 20th likewise divided into Anurakas, has nine, of which the third is subdivided into three Paryayas.—The text of this Sanhita has been edited by profs. H. Both and W. D. Whitney (Berlin, 1856).
The only existing Brahman'a of this Veda is the Saunkaka- or Gopatha-Brdhman'a. "That this Brahman'a," prof. Muller observes, " was composed after tile schism of the Charakas and Vajasaneyins, And after the completion of the Vajasaneyi-Sanhita, may be gathered from the fact, that where the first lines of the other Vedas are quoted in the Gopatha, the first line of the Yajurveda is taken from the Vitjasaneyine, and not from the Taittirlyas."—Ancient Sanskrit Lit., p. 452. Each of these Vedas received iu time Anukraman'is, or indices, which give the first word of each hymn, the number of verse's the names of the deities, the name and family of the poets, and the meter of every verse. The principal treatise of this kind is the Sarvii nukraman'l, or " The General Index," ascribed to the authorship of Saunaka. For the theosophical works which grew out of these Vedas, see the article UPANISHAD ; and for the works which were composed in order to secure a correct reading and understanding of the Vedic texts, and a correct performing of sacrificial acts, see the article VEDINGA. —At a later period the name of Veda was also bestowed on Rad sas—legends or legendary works—and Pura' n'as (q.v.), col lectively; but in this sense it never obtained real currency. Upavedas, or minor Vedas, are also mentioned in the Charan'avyitha and other works, and explained by them in the following manner: The Upaveda of the R'igveda, they say, is the Ayurveda, or the Veda on medicine—probably the well-known works of Charaka and Sus'ruta; the Upa veda of the Yajurveda is the Dhanurveda, or the Veda on archery; the Upaveda of the Samaveda is the Giindharvaveda, on music; and the Upaveda of the Atharvaveda is the Silpas'elstra, a work on mechanical arts, or, according to others, the Arthds'dstras, works on practical subjects, comprising polity, mechanical science, the training of elephants and horses, and fencing.
In the preceding brief outline of the four Vedas, the question as to the date at which they were composed has not been raised, because, in the present condition of Vedic philology, an answer to it could only be hypothetical. From astronomical facts, based on a statement in a Vaidik calendar, Colebrooke concluded that this calendar was writ ten in the 14th c. before the Christian era (Atiscell. Essays, vol. i. pp. 109, 110); and though subsequent writers have questioned the full correctness of this conclusion, those most reliable nevertheless admit that the error, if any, could not lessen the antiquity of this calendar by more than 100 or 200 years. As this calendar must have been composed
after the R'igveda had been arranged, and as such an arrangement itself must be poste rior to the date of its last hymn, a full scope is left for imagination to fill up these inter vals. But let it be understood that imagination alone would have to perform this task, since scientific research has as yet not yielded any means to check it, or prompt it on, as the case may be; nor is there any real prospect that future discoveries in Sanskrit litera ture will supply this want. A safer basis, however, may be looked for, if future research restricted itself to the question as to the relative age of these Vedic writings. Much vaIn ble evidence has been already brought forward in this respect to prove that there are R'ishis ancient, and less ancient (see, for instance, J. Muir's Original Sanskrit Texts, vol. ii. p. 205, ff.); that there are R'igveda hymns older than others (for instance, in Mailer's Ancient Sanskrit Literature); but, on the other hand, much confusion has also been pro duced by starting a theory that all the Brahman'as belong to one period, and all the hymns to another period preceding it, of which, again, two stages were thought to be discernible, and by assigning dates to the Brahman'a period, as well as to each of the two stages of the Mantra period. For, apart from the purely imaginary value of such dates, and apart from the circumstance that no evidence whatever has as yet been brought forward to justify an assumption of only two stages of hymns, each of which would comprise only 200 years, it is clear that the similarity of subject-matter alone— such as it marks the literary character of the Brahman'as—cannot be a criterion for determining that all the Brahman'as must be more recent than all the Sanhitas. That a Brahman'a of the R'igveda must be posterior to those hymns of the R'igveda Sanhita which it mentions, but to those alone—again, that a Brahman'a of the Samaveda must be younger than the hymns of the Satnaveda on which it relies, and so on—cannot be mat ter of doubt; but as the Sanhita of the Samaveda, for instance, must be more recent than that of the R'igveda, and as no fact whatever has been adduced to show why the Aitar eya Brahman'a, or other Brahman'as of the R'igveda, could not have appeared before a Samaveda-Sanhita was made, and so forth in the case of the other Vedas, it follows that it would be entirely unsafe to infer that all the Brillitnan'as must be more recent than all the Sanhitas; nay, even that all the Brahman'as must be later than all the hymns of the R'igveda, since not all of them need have existed before the oldest Brahman'a of this Veda was composed. A result like this is, unhappily, purely negative, but it may have the advantage of counseling caution and stimulating research.