I. ALLGEMEINES UND SPRACHE. 4. VEDIC GRAMMAR. = a. In the RV. ǎ+i is once contracted to ai- in práiṣayúr (1. 1205) — þrá iṣayúḥ (Pp.)¹; in the SV. āti is once contracted to ai- in áindra á indra (I. 2. 1. 45); and in the AV. and VS. the preposition à contracts with ṛ to ār in ārti- à-ti- 'suffering', and árchatu á-rcchatu 2. The last three instances are perhaps survivals of an older con- traction. That à is not otherwise contracted with i'r to ai au ār, is doubtless to be accounted for by the previous shortening of à in hiatus 3. 64 = b. Occasionally a followed by i remains uncontracted in the written text of the RV.; thus jyá iyám (VI. 753); pibā imám (VIII. 17'); raṇayā ihá (VIII. 34¹¹) 4. When ā is followed by r, it is either shorteneds or nasalized; e. g. tátha r-túḥ for táthā r-; kadám rtacid (v. 39), = kadá r-; vibhvāṁ ṛbhúr (IV. 33³)6 — vibhvā y-; vipanyáṁ7 | rtásya (IV. 112) vipanyà rtásya. c. On the other hand, in many instances where the contraction e oro is written, á the original simple vowels must be restored with hiatus; e. g. éndrāgnī (1. 1084) indragnī; subhágosah (1. 487) = subhágā uṣaḥ8. = 2. Final a contracts with a following e or ai to ai; and with o or au to au; e. g. dibhiḥ for á ebhiḥ. But though the contraction is written, the original vowels must sometimes be restored; e. g. áisu (1. 61¹) must be read á eșu. a. Final å, instead of being contracted with e and o, is in a few instances elided before those diphthongs; thus tatar' evéd (VII. 333) tatara evéd; iv' étayaḥ (x. 914) = iva étayaḥ; aśvin' evét (VII. 99) aśvina | evét; yáth' ohişe (VIII. 53) yátha oh s șe; úð eșatu úpa esatu 9. An example of a compound with this elision seems to be dášoni- 10 dáśa-oni- 'having ten aids'. = = b. Final å, instead of being contracted with is, in a few instances, nasalized before that diphthong; thus aminantam évaiḥ 11 (1. 792) for -a é-; śášadānāṁ éși (1. 12310) for à é-; upásthāṁ | ékā (1. 356) for -à é-. = 3. When å remains after a final y or s has been dropped, it does not as a rule contract with the following vowel. Nevertheless such contraction is not infrequent in the Samhitās. In some instances it is actually written; thus sártavájáu (III. 326) Pp. sártauái ājáu; vàsáu (v. 173) = Pp. vái asáu; and the compound rájesitam (VIII. 46²8) Pp. rájaḥ-isitam; in the later Sam- hitãs are found kŕtyéti (AV. x. 15) Pp. kŕtyah iti; pivopavasānanām (VS. XXI. 43) Pp. pivaḥ-upavasanānām ¹². In other instances the contraction, though not written, is required by the metre; thus ta indra (vII. 21°), Pp. te indra, must be read as tendra; pṛthivyá antárikṣāt (AV. IX. 19), Pp. prthivyāḥ, as pṛthivyántárikṣāt¹³; gosthá úpa (AV. IX. 4²³), Pp. gosthé úpa, as gosthipa. 12 a. = I Several instances of this contraction occur in B and later. 2 The TS. extends this contraction to pre- positions ending in a: úparchati, avárchati; see WHITNEY, APr. III. 47f., TPr. III. 9f. In the post-Vedic language this contraction was extended to all prepositions ending in ǎ. = There appear to be several other instances of such written contraction, which are however otherwise explained by the l'adapāṭha; thus rīyótá (X. 93¹⁰j rayé utá, 3 Cp. BENFEY, GGA. 1846, p. 822. 4 The Pp. explains pibā and ranayū as imperatives (piba, raṇaya); but the à here may represent -ās of the subjunctive (cp. WACKERNAGEL I, p. 311 mid.). Occasionally å remain uncontracted because the editors regarded them as representing aḥ-e, or āḥ, as in ranta itya (VII. 363) jmayá átra (VII. 393). Cp. RPr. II. 28 f.; BENFEY, SV. xxxf.; WACKER- NAGEL 1, 267 a ß. 5 See p. 63, note 10; à is shortened before in the AB.; see AUFRECHT's ed. 427. 6 vibhvām occurs thus three times; see LANMAN 529. In osition to the Mss. MAX MÜLLER, RV2. reads vipanyám rtásya because Sāyaṇa appears to favour that reading. 8 See ARNOLD, Vedic Metre 123. 9 This is a precursor of the post-Vedic rule by which the å of a preposition before initial e and o of verbs (except eti etc. and edhate etc.) is elided. 10 Though the AV. has páñcaudana- pánca odana- the elision of a before -odana- often takes place in the Sūtras and later. 11 The old hiatus is here treated as it would be at the end of an internal Pāda. The TS. (III. 1, 115) retains the hiatus in the same verse without nasalization; cp. OLDENBERG 469 ff. 12 See WACKERNAGEL I, 268 b. 13 In the Paippalāda recension this con- traction is actually written.
