I. ALLGEMEINES UND SPRACHE. 4. VEDIC GRAMMAR. -lohita- 'red', beside lodhá- 'reddish animal', rudhirá- (AV.) ‘reď'; also rauhiná- beside rudhikrá-, names of demons; the evidence of cognate languages further shows that represents dh in grhá- 'house'¹. h moreover represents dh in the adverbial suffix -ha in višvá-ha beside visvá-dha 'always'; in sahá 'together', beside sadha- in compounds; in ihá 'here', as shown by the Prakrit idha; and doubtless also in sama-ha 'somehow', though there is no direct evidence. The evidence of cognate languages, more- over, shows that in the endings of the 1. du. -vahe, -vahi, -vahai and the 1. pl. -mahe, -mahi, -mahai h represents dh (cp. Gk. -peda, etc.) ². 52 b. takes the place of bh in the verb grah, grh- 'seize' (with its derivatives gráha- 'ladleful', gráhi- 'demoness', grahyì- 'to be grasped', grhá- 'servant', hasta-gŕhya 'having take the hand') beside grabh-, grbh- (with the derivatives grábha- 'taking possession of', grābhá- 'handful'); and in the nouns kakuhá- 'high', beside kakúbh- 'height', kakubhá- (VS. TS.) 'high', m. (AV.) a kind of demon; bali-hrt- 'paying tax' (beside bhṛ- 'bear'); probably bárjaha- 'udder' (?), if formed with suffix -ha = -bha (in rsa-bhá-, etc.) ³. c. h takes the place of dh in bárbrhi, 2. sing. impv. of the intv. bárbṛh- of the root byh- 'make strong', for *barbṛḍhi (= *barbrṇḍhi, with the cerebral dropped after lengthening the preceding in pronunciation) 4. d. The rule is that appears for dh, dh, bh, only between vowels 5, the first of which is unaccented; e. g. ihá, kakuhá-, grhá- bárjaha-, bali-hŕt- rauhiná-, sahá. Similarly in grabh- ‘seize', h alone appears after unaccented y in RV.1-rx6; while on the other hand, in the 2. sing. impv. of graded roots, -dhi regularly appears after strong and therefore originally accented vowels; e. g. bodhi (from bhu- be'), y'dhi (from yu- 'yoke'), yuyodhi (from yu- 'separate'); śiśādhi, but śiśīhí (šā- ‘sharpen'). Again, in rudh- 'rise', the unaccented form of the root is regularly ruh-, while when it is accented, forms such as ródhati appear beside róhati, and in nominal derivatives dh predominates in accented radical syllables, the RV. having rýdha- and ā-ródhana- only, but the AV. riha- and aróhaṇa-7. a. At the same time dh and bh remain in a good many instances after an un- accented vowel; thus in the 2. sing. impv. krdhi, gadhi, śrudhi, śṛṇudhi; în adhás ‘below', adhamá- 'lowest', abhi 'towards'; rbhú- 'deft'; midhi- 'reward'; medha- 'wisdom'; vidháti 'adores'; vidhi- 'solitary'; vidháva- 'widow'; vadhi- bride'; sādhi- 'right'; rudhirá- (AV.) 'red'; derivatives formed with -dhí-, -bhá-, etc. The retention of the dh and bh here is partly to be explained as an archaism, and partly as due to borrowing from a dialect in which these aspirates did not become h, and the existence of which is indicated by the Prakrit form idha beside the Vedic ihá here'. The guttural aspirate media is probably to be similarly explained in meghá- ‘cloud' and arhá- ‘sinful'.. ¹ Cp. WACKERNAGEL 1, 217 a. 2 The in the perf. aha, ahúr, may represent dh; not, however, in náhyati binds', in spite of naddhá-; nah- rather stands for IIr. nažh- (cp. Lat. necto); the pp. instead of
- nadhá-, became naddhá through the in-
fluence of baddhá- from bandh- 'bind'. Cp. WACKERNAGEL I, 217 a, note, where several other wrong explanations of h dh are discussed. — 3 Some erroneous etymologies in which his explained as bh are discussed in WACKERNAGEL I, 217 b, note. 4 Cp. 12 and 17, 5. 5 It appears initially in hitá- -dhita-, but originally it was doutbless preceded in this position by a final vowel. That h — appears, though a vowel does not follow, in grhnatu (IV. 57) and hasta-gfhya, is due to other forms of grabh- in which is followed by a vowel. 6 In RV. x, however, gráha- and grähi- occur; cp. above, 2 b. 7 Exceptions like róhita-, višváha and inflected forms such as gáhate, are due to normalization. 8 Where the RV. fluctuates between h and dh bh (above 2, a, b), h appears throughout in the later Samhitās; thus always -hi, after vowels in imperatives except edhi 'be'; thus for śrnudhi (RV. VIIL. 843), SV. has śrnuhi. But a new h of this kind hardly occurs. Cp. WACKERNAGEL I, 219 a.
