________________
38
INDIAN ANTIQUARY, VOL. XXXIII, 1904; APPENDIX.
[$ 17,
suffer no material changes. Occasionally, howerer, as in kya (42, II, IV), kye (42, III), gyā (42, VI), and gye (42, VII), a single vertical stroke does duty both for the upper and the lower consonant, just as in the modern ligatures , th, and so forth; compare also the Kharosthi ligatures, 11 above,-0, 3.
(2) But there are cases of greater irregularities, especially in Girnår, where (a) the second sign is sometimes greatly mutilated or made cursive, as in vya (44, II), mya (44, VIII), sti and stu (45, VIII, IX); (b) the sign for the second consonant is sometimes placed first (Girnăr and Siddāpura) for convenience sake, an in sta, sti (42, VIII, IX), tpa, tpă (43, IX, X), vyā (44; X, Y); and (c) in ligatures with ra, this sign is either (both in Girnār and Siddāpura) inserted in the vertical lines of the other consonant (kera, 9, X; tram, 23, X; dra, 25, XII; bra, 30, X; vra, 86, X; aru, 89, X), or (in Girnar alone) is indicated by a small hook at the top of the combined sign (trai, 23, IX ; pra, pra, 28, IX, X; &c.). The position of ra always remains the same, whether it is to be pronounced before or after the combined consonant, and thus 36, X, bas the value both of rva and of ora. The insertion of ra in the left vertical of ba in bra (30, X) probably goes back to the period when the writing went from the right to the left. Otherwise it ought to stand in the right vertical.
17. - The Drāviçi of Bhattiprolu; Plate II. To the remarks on the value of the Drāvidi of Bhattiprola for the history of writing in India (above, page 8), and to the explanations of its peculiar signs (above, & 6, A, 3, 7, 12, 15, 18; B, 4 c, 5; and C, 2), I have now to add the reasons for the assumed reading of the sign in pl. II, 88, XIII-XV. It seems to me certain that originally it had the value of $. For there can be no doubt that it expresses a sibilant, and that the Dravidi is, like the Brähmi, an alphabet invented in order to write Sanskrit (see above, $ 6, 0, 2). As sigus for two of the three Sanskrit sibilants are easily recognisable, - the palatal in 37, XIII, XIV, and the dental in 89, XIII, XIV, XV,- the third sign can only bave been intended to express the lingual sibilant. But it is a different question, whether in the words of the Prākrit Bhattiprolu inscriptions, in which the sign occurs, the lingual sibilant was actually pronounced, or whether, owing to the negligent orthography of the clerks, the sign has been pat where the pronunciation was á ors. A certain answer to this question is for the present impossible. It could be given only if we knew more about the ancient Prakrit of the Kistna districts (39) than is actually the case. But the correct use of ea in éamanudesanain Bhattiprola, No. X) indicates that the dialect possessed two sibilants; and it can only be doubted, whether $ bas been pat erroneously for é, as often happens in the Jaina inscriptions from Mathurā (compare EI 1, 376), or whether it was still the lingual sibilant. Another point in the character of the Dravidi, which requires special mention, is, that its signs, which agree with those of the Brāhmi, in several cases present characteristio peculiarities of the southern variety. "I his may be seen (1) in the angular A, 7; (2) in the kh (10, XIII, XV) consisting, like that of Girnär, merely of a vertical, with hook at the top; (3) in the dh, which has the same position as that of the Jangada separate edicts and the Nănågbåt inscriptions ; (4) in m, wbich, though tarned topsy-turvy, retains the angle of the ma of Girnār; and (5) in s, which mostly has the straight side-limb, as in Girnär and Siddäpura.
As the inscription on the crystal prism (No. X), found with the stone vessels, shows the ordinary Brāhmi except in the da opening to the right, it follows that the Dravidi was not Deed exclusively even in the Kistna districts, but together with the common old Indian alphabet.
The small pumber of the inscriptions hitherto found, makes it impossible to say anything definite regarding the spread of this alphabet. And it is equally difficult to fix with certainty the time and the duration of its use. As king Kubiraka or Khubiraka (Knbera) is not known from other sources, we can only fall' back on the never absolutely certain paleographic indications.
10. TENE, Gurupājākaamid 96, thinks that these groups should be read trä, tal, as the are written.
