________________
180
the other Gāthās. The same would seem to be necessary for Gatha *1. Thus, the description of a goddess and her husband seems out of place in a text which is mainly concerned with common people such as farmers and hunters, despite the parallelism in their respective love-lives. A further argument is its form, which is that of a Stotra. It should be noted that most of the other Gathās which have the form of a Stotra, praising either Siva, Vişnu or Ganesa, are in all likelihood later additions to the text, as follows from their absence in the Third South-Indian recension. See, for instance, 448 (Siva), 151, 406 and 411 (Vişnu), and 403 (Ganesa). In some of these, 403, 406 and 411, erotic connotations are lacking completely. However, Gathā 455, a Stotra praising Siva, like the one under discussion, goes back to the archetype.
My translation of the Gathā differs considerably from Weber's. The latter scholar connects the first word of the Gātha, pasuvaiņo, with sanjhāsalilańjalin which he takes directly as the object to namaha, and arrives at the following interpretation: 'Bow before the Sandhya-water in Pasupati's hands' (see 'Retr., p. 347). He may have been led to do so by the position of the particle of comparison, which is here, contrary to common practice, not found after the Upamāna but after a compound which describes the Upamāna. It should be noted that this is not uncommon in Pkt (see Gathā 93: siviņaap Tena va pāņieņa), nor in Skt (see Speyer § 430).
The -e- in Tp gehia, for gahia, is probably due to the South-Indian pronunciation of -a after certain consonants as -e. This phenomenon has been described for Malayā] am by Godavarman (1935-37) and by Kunhan Raja and Ramachandra Sarma in their edition of the Prākrtaprakāśa (p. IX). These authors restrict it to -a occurring after voiced plosives and after C-, y-, r- and 1-. Indeed in the South-Indian MSS of the Sattas at one comes across *7 lehada Tp, 60 deia Ma and hypercorrect 132 davvena Ma and 532 gahnai Ma, Ti, TP, T. But in these same MSS -e for -a or hypercorrect -a for -e is also found after voiceless plosives and after h-. E.g. 189 teha Ma, 412 teio Ma, 642 kendofța Ma and 183 pecchanna Ma (but in the last instance probably under the influence of the following palatal plosive; cf. 373 pacchaîto Ti, 457 pacchai Ti); hypercorrect -a in *17 kuppajja Ma, Tp, 4?0 uvakkhio Ti, Tp (written upakkhio), 445 uppakkhimo Ti. After h-: 914 hero Ma (+(b)haro) and 930 hea Ma..Further
