________________
168
deviating most from Sanskrit.
A frequent dilemma occurs in connection with the endings of the nominative and accusative neutre plural, -āi(n), the genitive plural, -āņam) etc., and the instrumental plural, -ehim) etc., when found at the end of the hemistich. This position is metrically neutral so that it does not make any difference whether the text reads, for instance, -ãi or -ain. On this point, though, there is a significant difference be- . tween Ma, Ti and Tp, on the one hand, and Weber's MSS, on the other. The former, with only a few exceptions, throughout have the endings without Anusvāra, while the latter favour those with Anus vāra. The actual situation in Weber's MSS is admittedly unclear. K, edited in the Abh., indeed largely favours the endings with Anusvāra, but this is to be expected in a MS in which the Anusvāra is inserted apparently at random and without regard for the metre (see salak ārāņań in *3 and land aha in *8). In Retr. all deviations have been noted for B, P, T and S; from K, which for these MSS gives the same picture: generally the endings occur with Anusvāra. However, in Ed. Weber refrains from noting this type of variation for the MSS, or parts of MSS, edited there. Moreover, he seems rather careless on this point. Thus, in 354 he reads muheņań , against all the MSS, which have muhena, except possibly Bh, for which in ISt. he does not note a variant (p. 109). This may, however, also be taken as an indication of the fact that he did not note this type of variation of Bh from the text of Ed. either.
Before continuing the discussion of these endings in hemistichfinal position a few remarks may be made regarding their occurrence within the Gathā. There the variation between, for instance, -āi and -āin is entirely regulated by the metre. The endings without Anusvāra are considerably more frequent than those with Anusvāra. The latter are specifically found at the end of the uneven Padas where u-u does not occur. In this position one also finds -ena (relatively frequently) and -es un (once, in 77, a Gathā which probably is a later addition anyhow). -es un is otherwise not found. -enam is otherwise found only once, namely in 525. -ain, -āņam and -ehin are otherwise rare too, and a comparatively large number of these instances is found in particular before the enclitics cia, pi and va, -ehin in fact only so.
While within the Gathā the occurrence of the short and the long
