80 PAUMACARIU in it to decide in favour of any particular alternative. Of course Svayambhu and Hemacandra lay down' the Gana-schemes for all the Pādas containing from seven to seventeen moras, but there are seve- ral and overlapping schemes for each Páda and in several cases there 19 disagreement between the two metricians. Hence this does not help us in making the choice out of the four possible alternatives as pointed out above. A reference to the illustrations adduced by Hercandra and Svayambhū for different kinds of Ghattâs only helps to bring the complicated nature of our problern into relief. The text of the Sva- yambhūcchandas is considerably corrupt in its Ap. section. Hence it can prove of limited help only in so far as we are able to check the illustrations from PC. So some of the Ghattā illustrations from the Chando'nusasana we shall consider first. In an overwhelming majority (but not all) of the cases the end syllable in the Ghattā illustrations of Ch. is to be regarded as long irrespective of its natural value, but in several cases (VI 19 4, 11 bd; 21a, 44; 20 78, 115; 23; 30; 14 1 cf.; 15 2, etc.) it is to be given its natural value, i.e. if it is short, it is to be counted one-moraic; if long, two-moraic. Further, in two Pādas of the same metrical value, the final short is to have its actual value at one place, is to be counted as long at another. Thus in the following illustration of the Campakakusumā (7+8) anga-cangima, jai gorangihim/ campaya-kusuma, tā kaha agghahim// (Ch. VI 19 4) the end syllables of the 7-moraic Padas (i.e., a, c) are to be count- ed as short, but in the following illustration of the very next variety Sämudgaka (74-9) jai bollai, ghana ukkanthia/ sā muddau, muhu kalayaņthia/ (Ch. VI 19 5) the end syllables of the 7-moraic Pādas are to be regarded as long. Thus the Pādas which are theoretically equivalent turn out to be actually different. Similarly though the even Pädas of the above- quoted illustrations of the Campakakusumā and the Samudgaka have to all appearance the same mora-contents, yet their end sylla- bles are to be regarded as short in one case, long in the other. Thus the Pādas that are actually equivalent turn out to be theoretically different. Now let us glance at the Svayambhūcchandas. Svayambhū's general practice also in SC. appears to be to consider the end sylla- ble as long. But here too some Şațpadis (e.g. VIII 20-21, V 7) and Catuşpadis (e.g. Kāmiņihāsaa VI 112 a c, Chabbhania VIII 15 etc.) are treated as ending in a short. One illustration from the actual practice of another epic poet will be illuminating on this point. The commencing stanza of the 6. Sandhi of Puşpadanta's Mahapurāna is called Malayavilasiyā in in the text itself. Ch. VII 66 and Chandassekhara, 234 define a Dvi- padi called Madanavilasitā, but its scheme is given as 5+3, which does not apply to the Malayavilasiyās of MP. But SC. VII 10 des- cribes a Dvipadi called Malaaviasia which has the scheme 6+2, which is satisfied by the MP. stanzas in question. Now on examining these stanzas, we find that in three cases (MP. 61, 63, 6 7) all the four Padas end in l. In the rest, the end is a long. ix- moraic Gana a heavy syllable is avoided in the place of evenodd (1) BC. VI 28, 53, 76, 95, 110, 124, 134, 141. 144, 146, 147; Ch. VI 4-14,
