page 220 note * The dual vrishnau occurs only when the next word begins with a vowel. Before an initial a, â, i, the au is always changed into âv in the Sanhitâ, (i. 108, 7–12; 116,21; 117,19; 153,2; 157,5; 158,1; 180,7; vii. 61, 5). Before u the preceding au becomes â. in the Sanhitâ, but the Pada gives au, in order to show that no Sandhi can take place between the two vowels (vii. 60, 9; x. 66, 7). Before consonants the dual always ends in â, both in the Sannitâ and Pada. But there are a few passages where the final â occurs before initial vowels, and where the two vowels are allowed to form one syllable. In four passages this happens before an initial â. (i. 108, 2; vi. 68, 11; i. 177, 1; ii. 16, 5). Once and once only it happens before u, in viii. 22, 12.

page 230 note * Cf. i. 58, 7: saptá guhvãh

page 233 note * Âs, mouth, the Latin os, oris, has been derived from a root as, to breathe, presented in the Sanskrit as-u, spirit, asu-ra, endowed with spirit, living, the living god. Though I agree with Curtius in admitting a primitive root as, to breathe, from which as-u, breath, must have sprung, I have always hesitated about the derivation of âs, and âsya, mouth, from the same root. I do not think, however, that the lengthening of the vowel in as is so great a difficulty as has been supposed (Kuhn, Zeitschrift, xvii. 145). Several roots lengthen their vowel a, when used as substantives without derivative suffixes. In some cases this lengthening is restricted to the Anga base, as in anadvâh; in others to the Anga and Pada base, as in visvavât, visvavâbhih etc.; in others again it pervades the whole declension, as in turâshât. (See Sanskrit Grammar, § 210, 208, 175.) Among ordinary words vâk offers a clear instance of a lengthened vowel. In the Veda we find rtîsháham, vi. 14, 4, and ritîshâham (Sanhitâ), i. 64, 15. We find vâh in apsu-vâh (Sâm. Ved.), indra-vâh, havya-vâh. Sah at the end of compounds, such as nri-sah, pritanâ-sah, bhâri-sah, satrâ-sah, vibhâ-sah, sadâ-sah, varies between a long and short â. (See Regnier, Etude sur l'idiome du Vedas, p. 111.) At all events no instance has yet been pointed out in Sanskrit showing the same contraction which we should have to admit if we derived âs from av-as, or from an-as. From an we have in the Veda âná, mouth or face, i. 52, 15. From as, to breathe, the Latin omen, originally, os-men, a whisper, might likewise be derived.

paage 236 note * Rv. i. 61, 10; 122, 5; 134, 2; 139, 6; ii. 1, 10; iv. 29, 5; 32, 9; v. 59, 1, 4; vi. 65, 3; viii. 25, 20; 45, 10; (92, 6); 46, 25; 27; 63, 5; 69, 17; 70, 12; ix. 93, 4; x. 32,5; 44,7; 50,7.