character. Such are, for instance, ii. 12, 14, sasamânám instead of sasamânám; i. 124, 4, sudhyúvah, in the Pada, instead of sundhyúvah; and the substitution in several places of a short u instead of a long û in such forms as sûsávâma, when occurring in the Pada; cf. i. 166, 14; 167, 9.

It is clear from Sûtra 819 and 163, 5, that the two words ûtî índra in iv. 29, 1, should not be joined together, but that in the Sanhitâ the hiatus should remain. Hence ûtîndra, as printed in my edition and repeated in Professor Aufrecht's, should be corrected, and the hiatus be preserved, as it is in the fourth verse of the same hymn, úti itthẩ. MSS. S. 1, S. 3. are right; in S. 2. the words are joined.

It follows from Sûtra 799 that to double the y in vaiyasva is a mistake, but a mistake which had to be pointed out and guarded against as early as the time of the Prâtisâkhya. In viii. 26, 11, therefore, vaiyyasvásya, as printed in my edition and repeated in Professor Aufrecht's, should be changed to vaiyasvásya. MSS. S. 1, S. 3. are right, likewise P. 1, P. 2; but S. 2. has the double mistake vayyasvásya, as described in the Prâtisâkhya; another MS. of Wilson's has vaiyy. The same applies to P. 1. admits the mis

viii. 23, 24, and viii. 24, 23.

taken spelling vayyasva.

Some corrections that ought to be made in the Padapâtha only, as printed in my edition, are pointed out in a note to Sutra 738 of the Prâtisâkhya. Thus,

according to Sûtra 583, 6, srûyah in the Pada text of ii. 10, 2, should be changed to sruyah. MSS. P. 1, P. 2. have the short u.

In v. 7, 8, I had printed súkih shma, leaving the a of shma short in accordance with the Prâtisâkhya, Sûtra 514, where a string of words is given before which sma must not be lengthened, and where under No. II. we find yásmai. Professor Aufrecht has altered this, and gives the â as long, which is wrong. The MSS. S. 1, S. 2, S. 3. have the short a.

Another word before which sma ought not to be lengthened is mavate. Hence, according to Sûtra 514, 14, I ought not to have printed in vi. 65, 4, shmâ mấvate, but shma mavate. Here Professor Aufrecht has retained the long â, which is wrong. MSS. S. 1, S. 2, S. 3. have the short a.

It follows from Sûtra 499 that in i. 138, 4, we should not lengthen the vowel of sú. Hence, instead of asya û shu na úpa sâtáye, as printed in my edition and repeated by Professor Aufrecht, we should read asya û shú na úpa sâtáye. S. 1, S. 2, S. 3. have short u.

In vii. 31, 4, I had by mistake printed viddhí instead of viddhĩ. The same reading is adopted by Professor Aufrecht (ii. p. 24), but the authority of the Prâtisâkhya, Sûtra 445, can hardly be overruled. S. 1, S. 2, S. 3. have viddhĩ.

While in cases like these, the Prâtisâkhya is an authority which, as far as I can judge, ought to overrule the authority of every MS., however ancient,

we must in other cases depend either on the testimony of the best MSS. or be guided, in fixing on the right reading, by Sâyana and the rules of grammar. I shall therefore, in cases where I cannot consider Professor Aufrecht's readings as authoritative improvements, have to give my reasons why I adhere to the readings which I had originally adopted.

In v. 9, 4, I had printed by mistake purú yó instead of puru yó. I had, however, corrected this misprint in my edition of the Prâtisâkhya, 393, 532. Professor Aufrecht decides in favour of purú with a short u, but against the authority of the MSS., S. 1, S. 2, I, S. 3, which have pură.

It was certainly a great mistake of mine, though it may seem more excusable in a Romanised transcript, that I did not follow the writers of the best MSS. in their use of the Avagraha, or, I should rather say, of that sign which, as far as the Veda is concerned, is very wrongly designated by the name of Avagraha. Avagraha, according to the Prâtisâkhya, never occurs in the Sanhitâ text, but is the name given to that halt, stoppage, or pause which in the Pada text separates the component parts of compound words. That pause has the length of one short vowel, i. e. one mâtrâ. Of course, nothing is said by the Prâtisâkhya as to how the pause should be represented graphically, but it is several times alluded to as of importance in the recitation and accentuation of the Veda. What we have been

in the habit of calling Avagraha is by the writers of certain MSS. of the Sanhitâ text used as the sign of the Vivritti or hiatus. This hiatus, however, is very different from the Avagraha, for while the Avagraha has the length of one mâtrâ, the Vivritti or hiatus. has the length of mâtrâ, if the two vowels are short; of 1⁄2 mâtrâ, if either vowel is long; of & mâtrâ, if both vowels are long. Now I have several times called attention to the fact that though this hiatus is marked in certain MSS. by the sign s, I have in my edition omitted it, because I thought that the hiatus spoke for itself and did not require a sign to attract the attention of European readers while, on the contrary, I have inserted that sign where MSS. hardly ever use it, viz. when a short initial a is elided after a final e or o; (see my remarks on pp. 36, 39, of my edition of the Prâtisâkhya.) Although I thought, and still think, that this use of the signs is more useful for practical purposes, yet I regret that, in this one particular, I should have deviated from the authority of the best MSS., and caused some misunderstandings on the part of those who have made use of my edition. If, for instance, I had placed the sign of the Vivritti, the s, in its proper place, or if, at least, I had not inserted it where, as we say, the initial a has been elided after e or o, Professor Bollensen would have seen at once that the authors of the Prâtisâkhyas fully agree with him in looking on this change, not as an elision, but as a contraction. If, as sometimes

happens, final o or e remain unchanged before initial short a, this is called the Pañkâla and Prakya padavritti (Sûtra 137). If, on the contrary, final o or e become one (ekîbhavati) with the initial short a, this is called the Abhinihita sandhi (Sûtra 138). While the former, the hiatus of the Pañkâla and Eastern schools, is marked by the writers of several MSS. by the signs, the Abhinihita sandhi, being a sandhi, is not marked by any sign*.

i. 3, 12. ragati (Aufr. p. 2) instead of rågati (M. M. p. 75) is wrong.

vol. i.

i. 7, 9. ya ékah (Aufr. p. 5) should be yá ékah (M. M. vol. i. p. 110), because the relative pronoun is never without an accent. The relative particle yathâ may be without an accent, if it stands at the end of a pâda; and though there are exceptions to this rule, yet in viii. 21, 5, where Professor Aufrecht gives yáthâ, the MSS. are unanimous in favour of yatha (M. M. vol. iv. p. 480). See Phit-sûtra, ed. Kielhorn, P. 54.

i. 10, 11. â tď (Aufr. p. 7) should be a tu (M. M. vol. i. p. 139), because â is never without the

accent.

i. 10, 12. gúshtah, which Professor Aufrecht specially mentions as having no final Visarga in the Pada, has the Visarga in all the MSS., (Aufr. p. 7, M. M. vol. i. p. 140.)

* As to the system or want of system, according to which the Abhinihita sandhi takes place in the Sanhitâ, see p. xxxv seq.