________________
766
EPIGRAPHIA INDICA
(VOL. XXX
cells or groups of cells'. This system can be traced to oiroa 200 A.D. only if Hoernle's view regarding the date of the Bakhsali manuscripts is accepted. This scholar assigned the manuscripts to the third or fourth century, although the suggestion has been disputed by Kaye.
But, considering the imperfect nature of the system of writing fractions by symbols, it is possible to conjecture that the custom followed in the Bakhsali manuscripts is really very old. It is, however, certain that some fractions at least were expressed by symbols, as it is done even today, when the old system of writing numerals by symbols was prevalent and this is definitely suggested by the present inscription belonging to the sixth century. In this inscription, the symbol for five exactly robembles the akshara na while that for ten looks somewhat like ndā (with the loop of n having an opening above and the lower part of not being so flat as in ordinary cases), and, in writing one-half and one-fourth, a symbol exactly resembling the akshara vi bas been employed before those for ten and five respectively. Thus the symbol indicating one-half looks somewhat like vindā and that indicating one-fourth exactly like vina. Whether these are developed and modified forms of the symbols originally used in writing one-half and one-fourth cannot be satisfactorily determined and the occurrence of the symbols for five and ten in those respectively for one-fourth and one-half is also not very easy to explain. It should, however, be pointed out that vi may be taken to stand for the word visati meaning 'twenty'. In that case it can be suggested that vi 5 means '5 in relation to 20', i.e. 1.-t, and vi 10 indicates 10 in relation to 20', i.e. 18-1. In this system may have been expressed with vi placed before the two symbols jointly indicating 15. It is clear, however, that all fractions could not be written with the help of 20 alone. But whether a fraction like was expressed as 5 in relation to 40 ' is more than what we can say in the present state of our knowledge.
The language of the inscription is Sanskrit. The record is composed throughout in prose. There are a few grammatical errors (of. pratibhuvēna in line 10; anäprishtva in line 17 ; etc.). Interesting from the orthographical point of view is that final m has been changed to anusvāra before vowels in a large number of cases (cf. lines 4, 10, 12, 18, 23, etc.). The word dharmmikēna, correctly spelt in lines 11, 13-14, is written with n in lines 22-24, 27. Sandhi, which is optional in prose composition, has not been observed in many cases (cf. lines 8, 10, 12, 15, 20-21, 24, etc.). Consonants like g, gh, j,, t, d, n, m, y and v have been reduplicated after r ; but th and dh after the same letter have been similarly reduplicated only occasionally. The record is highly interesting from the lexical point of view, as it abounds in words which are not to be traced in the standard Sanskrit lexicons and are in many cases extremely difficult to interpret. As will be seen from our discussion below, some of these words are of Prakrit origin and may be explained with the help of the vocabularies of some forms of the Prakrit speech. A number of obscure and technical words, which could not be traced to any other work, have been conjecturally explained.
The document is dated on the 5th day of the bright half of Śrāvana in the year 649 and the endorsement on the 7th day of the dark hall of Kārttika in the year 357. Both the dates are expressed in symbols instead of numerical figures of the decimal notation. We have to note that the year 357 refers to a date later than that indicated by the year 649 as, it will be clear from our discussion below, the endorsement must be some years later than the document itself. Now, considering the palaeography of the document, the year 649 can only be referred to the Vikrama era and taken as corresponding to 592 A.D. This is also supported by the fact that the later dute, year 357, which, considering the palaeography of the endorsement, can only be referred to the Traikūta ka-Kalachuri-Chēdi ers or the Gupta-Valabhi era, corresponds either to
Und. Ant., Vol. XVII, p. 38. Relying on Hoernle, Böhler suggested that the use of the decimal system in Indis may be as old us the beginning of the Christian era or even carlier. Cf. ibid., Vol. XXXIII, Appendix, p. 82.
* JASB, 1907, pp. 475 ff.
