The marked minor additions and other minor changes.—In a work like this, involving so great a mass of multifarious details, it was inevitable that a rigorous revision, such as the author could not give to it, should detect many statements requiring more or less modification. Thus at xix. 40. 2, the author, in his copy for the printer, says: "We have rectified the accent of sumedhā́s; the mss. and SPP. have sumédhās." In fact, the edition also has sumédhās, and I have changed the statement thus: "⌊in the edition⌋ we ⌊should have⌋ rectified the accent ⌊so as to read⌋ sumedhā́s." The changes in the last two books are such that it was often best to write out considerable parts of the printer's copy afresh: yet it was desirable, on the one hand, to avoid rewriting; and, on the other, to change and add in such a way that the result might not show the unclearness of a clumsily tinkered paragraph. To revise and edit between these two limitations is not easy; and, as is shown by the example just given, there is no clear line to be drawn between what should and what should not be marked. As noted above, it is evident that all these matters would have been very simple if the author could have seen the work through the press.
The revision of the author's manuscript. Verification.—The modifications of the author's manuscript thus far discussed are mostly of the nature of additions made to carry out the unfinished parts of the author's design, and are the modifications referred to on the title-page by the words "brought nearer to completion." The work of revision proper has included a careful verification of every statement of every kind in the commentary so far as this was possible, and a careful comparison of the translation with the original. This means that the citations of the parallel texts have been actually looked up and that the readings have been compared anew in order to make sure that the reports of their variations from the Atharvan readings were correct. This task was most time-consuming and laborious; as to some of its difficulties and perplexities, see below, p. lxiv. Verification means further that the notes of Whitney's Collation-Book and of the Bombay edition and of Roth's collation of the Kashmirian text were regularly consulted to assure the correctness of the author's reports of variants within the Atharvan school; further, that the text and the statements of the Major Anukramaṇī were carefully studied, and, in connection therewith, the scansion and pāda-division of the verses of the Saṁhitā; and that the references to the Kāuçika and Vāitāna Sūtras were regularly turned up for comparison of the sūtras with Whitney's statements. Many technical details concerning these matters are given on pages lxiv ff. of the General Introduction. Since the actual appearance of Bloomfield and Garbe's magnificent facsimile of the birch-bark manuscript
