VI. DECLENSION. NOUNS. VOWEL STEMS. is clearly due to the influence of metre. The - here seems never to be shortened to -, as is so frequently the case in the N. A. n. pl. of -an stems *. Nor does it avoid hiatus (like the -a of the N. A. du. m.), though coalescence with a following vowel sometimes takes place ². Examples of the commonest forms are: 1. havyá (44), bhúvanā (36), durită (31) 'distresses', sávanā (30), ukthá (25) ‘praises'. -2. bhúvanani (57), vrtráni (36) 'foes', vratáni (34) 'laws', havyáni (25), kṛtáni (20) ‘done'. A. m. The ending of this case is not the normal -as, but -n, before which the final vowel of the stem is lengthened ³, e. g. áśvā-n. The form is frequent, being made from more than 250 stems in the RV. That the ending was originally -ns is shown by the treatment of an in Sandhi, where it becomes -am before vowels and the sibilant itself occasionally survives before c- and t-4. I. m. n. In this case there are two forms, the one adding the normal ending -bhis (before which the final vowel of the stem becomes -e), while the other ends in -ais (which does not appear in any other declension). The form in -ais is only slightly commoner in the RV., being made from 221 stems, while that in -bhis is made from 211. In the AV., however, the former is 5 times as frequent as the latters. The two forms often appear in the same Pāda; e. g. upamébhir arkáis (L. 33²) 'with highest songs'. The choice is often due to the metre; e. g. yatám áśvebhir aśvina (VIII. 57) 'come with your steeds, O Aśvins', and ādityáir yātam aśvinā (vIII. 35¹³) ‘with the Adityas come, O Aśvins'. In the RV. the m. forms are roughly twice as numerous as the n. Examples of the most frequent forms are: I. arkáis (43); uktháis (35), yajñáis (34), deváis (31), ášvais (30), stýmais (25). 2. devébhis (52), stómebhis (26), vájebhis (21). 261 D. m. n. This case is formed with the normal ending -bhyas, before which the final -a of the stem appears as -e. In the RV. it is made from over 40 stems in the m., but from only one in the n. In about half the forms occurring the ending has to be read as a dissyllable -bhias. The forms occurring are: ajárebhyas 'unaging', ámavattarebhyas 'mightier', áranebhyas (Kh. v. 12) 'foreign', arbhaktbhyas 'small', adityébhyas, ašinébhyas 'ageď', àsv-àpas- tarebhyas 'working more quickly', úmebhyas 'helpers', gárbhebhyas infants', grhébhyas, jánebhyas, jivébhyas 'living', jñātebhyas (Kh. III. 16¹) 'known', tāvakébhyas 'thy', dása-kakşyebhyas 'having ten girths', dása-yoktrebhyas 'having ten traces', dáśa-yojanebhyas 'having ten teams', devébhyas, pajrébhyas 'strong', párvatebhyas 'mountains', pasprdhanebhyas 'striving', pitu-kyttarebhyas 'procuring more nourish- ment', písunebhyas 'treacherous', putrebhyas 'sons', puruṣebhyas 'men', párvebhyas 'former', badhitebhyas 'oppressed', bharatebhyas 'descendants of Bharata', mártye- 2 On some probable mistakes made by the Pada in contracted forms see LANMAN 348. - sávanā puruni (111. 368); urdhvá socímși prásthita 5 In the independent Mantra portions of rájāmsi (III. 44); rabhasá vápumsi (III. 18). the TS. the proportion is about the same There seems no sufficient reason to as in the AV. The following four forms with assume that in vísvéd áha (1.92³) as com--bhis occur: étasebhis (1. 2. 4¹), devébhis (111. pared with áha víšvā (1 130²), áha is formed 1. 43), rudrébhis (II. 1.11²), su-yámebhis (IV. 7. from the transition stem áha- rather than from áhan-; cp. LANMAN p. 348. 153). In the Khilas 5 forms in -ais to 7 in -ebhis occur. The latter are: amiva-cátanebhis (1. 117), artavébhis (111. 166), ukthebhis (V. 6³), svébhis (111. 17), kárvarebhis (1. 5¹), ráthebhis (1.117), sāraghébhis (1. 117). In the later language the form in -bhis survives in the pronominal ebhis alone. 3 This lengthening is at least Indo-Iranian: BRUGMANN, Grundriss 2, p. 672². 4 See above 77; and cp. LANMAN 346 on the Sandhi of an in general. 6 On the origin of the two forms cp. BRUGMANN, Grundriss 2, p. 717.
