________________
No. 47.)
RECORDS OF THE SOMAVAMSI KINGS OF KATAK.
329
not compelled to place the record, on this account, before A.D. 794-95 ; for, the same system was used in the grant (from somewhere in Bengal) of the Maharaja Vinayakapala, the date of which falls in that year (Ind. Ant. Vol. XV. p. 140, Plate). And the force of this feature is impaired by the fact that decimal figures also occur in all the six charters. This latter feature, indeed, does not oblige us to place these records after A.D. 862; for, decimal figures occur in the Deógadh inscription (Central India) of Bhôjadêva of Kanauj, of that year (Archæol. Suru. Ind. Vol. X. Plate xxxiii. No. 2). But it points to about A.D. 800, as the eatliest possible time for these charters.
The next point to which attention may be given, is the use of the virima, in conjunction with the full form of t in B. to F., and once in conjunction with the full form of n in B.; this feature, however, being qualified by the use also of final.forms of t, n, and m, more or less throughout the whole series. The virama perhaps occurs with t in samvat, line 6, in the Deógadh inscription of A.D. 862. It certainly occurs, in conjunction with a half final form of t,-& complete t, except for the absence of the matrå or horizontal top-stroke,- in pañchafat, line 20.4twice), in the Gwalior inscription (Central India) of the same king, of A.D. 875-76 (Ep. Ind. Vol. I. p. 160, Plate). It appears in conjunction with the full forms of t and m in the Pehoa inscription (Panjab) of Mahendra påla, of about A.D. 900 (ibid. p. 244, Plate); see, for instance, sdragabhrit, line 2, and tápam, line 3. And it occurs in conjunction with the full forms of t, n, and m in the Khajuraho inscription (Bundelkhand) of Yaśðvarman, of A.D. 953-54 (ibid. p. 124, Plate); see, for instance, vardt, line 1, tyagavdn, line 3, and niroyitim, line 28. So far, therefore, as the use of the viráma is concerned, the present charters need not be placed after A.D. 900. And the use of final forms also might seem to necessitate oar placing them before that date. But, though the records mentioned above shew no final forms at all, a final form of t occurs in the word samvat, lines 32 and 33, of the Harsha inscription (Rajputânå) of Vigrahardja, of A.D. 973 (id. Vol. II. p. 120, Plate); and the Dewal inscription (North-West Provinces) of the Chhinda prince Lalla, of A.D. 992-93 (id. Vol. I. p. 76, Plate), shews, e.g. in bhuvam, line 8, a half final form of m,- a complete m, except for the mdtrd, with viráma attached to it. And therefore, in spite of the use of final forms, the charters may also, without objection, be placed as late as A.D. 973.
Another instructive feature is the use of the avagraha in E. and F.; properly, in E. line 38, to denote the elision of an initial a after an 6, and incorrectly, in E. line 29, and F. line 35, to mark the absorption of an initial a into a preceding &. This sign does not occur in any of the other records mentioned above. And the earliest other instances of the use of it, that I can quote, are vásnumódéta in line 43 of the Barðda grant (Gujarat) of Dhåråvarskas Nirupama-Dhruvaraja, of A.D. 834 or 835 (Ind. Ant. Vol. XIV. p. 200, Plate), and vuddhudsemad in line 25 of the Ujjain grant (Mâlwa) of Vakpatiraja of Dhårå, of A.D. 974 or 975 (id. Vol. VI. p. 52, Plate), in both of which places it is used, improperly, to mark the absorption of an initial a into a preceding 4, and filpésntaráyaḥ, line 4, dadhésvasddam, line 13, and lanchhitôsoir, line 18, where it is used correctly, and frutvdsnyatha, line 19, where it is used incorrectly, in the Deopara 'inscription of Vijayasena, attributable to the end of the eleventh century A.D. (Ep. Ind. Vol. I. p. 308, Plate). It is a sign which is emphatically opposed to any particular amount of antiquity.
Still more instructive is the use of the fully developed måtrås or horizontal topstrokes of the letters, almost all through these charters. In A. and F., this feature is very anmistakable, plain straight lines being used. In B., O., and D., the upright strokes are capped with triangular tops, cloven on the upper side; the måtrds were formed by running
1 The Aphend inacription was constructed in such a way as to avoid both the birdma and final forms; except in one word, varayan, line 9, where an ansatdra was wrongly substituted for the #, either as a final, or in sandhi with the following i.
20
