hope shortly to make public. In one of these, the date was entered at full in the words ^^ FcR^rfsprFqfa % ' in the samvat year three hundred and ninety-four.' A few lines below this, the word 1 1 ^q*<|pc I I again occurred, followed by three symbols ^ffl^i* which must of course be numerals : they are more exactly copied in Plate XX ; and according to the preceding statement should be 394. On a second plate in the same manner, the date in words was ^T^x im ^^f»r<^fV^- 3flffNf jg^'WFWf, « in the 15th of Kartik, samvat 380/ and in figures ^^[03 ^Tlffi^f F |Jr<^. On a third plate the date in words was Fjrr^^xf^rfVj^fVij% ^TTfw Kartik full moon, samvat 385' and in figures ^TQQ^ and djr<^ as before : in both of which the same symbols occur for 1, 3, 8, and 5 ; and the latter figure, much resembling the ancient letter na, but slightly altered was again observed on a fourth plate sent me by Dr. Burn from Gujerat, which did not contain the date in words, thus, ^
Much pleased with this new train of discovery, I turned to Mr. Wat hen's paper in the fourth volume of the Journal, in which I remembered his interpretation of the date on a similar grant by Ski Dhara Sena, as being in the ninth year of the Valctbhi Samvat of Tod, corresponding with A. D. 328. Here the translator had no written entry to guide him, nor had he any clue whereby to recognize the numerals which followed the abbreviated Samvat, thus, ^T 3 which we now perceive to be 300, + some unknown unit. I immediately wrote to Mr. Wathen and to Dr. Burn, requesting them to examine carefully the dates of all other plates in their possession, and from them in return I received all the examples which are inserted in the accom- panying plate. From the whole series combined we may venture to assign a certain value to the 1, the 3, the 4, the 5, the 8 and the 9.
The last of these, I could not but remember as the symbol on one of the JBhilsa inscriptions which led to so many conjectures a year ago. In the form of we have evidently our fg , or the year 9, but the three strokes at the side would appear to modify its value, or to be themselves a numeral, perhaps the o. Then, as we find the preceding Jj has not a dot above it, we may use that also as a numeral and under- stand the whole JJ88r as 2 or 6 or 790 according to the value to be hereafter assigned to jty.
Again in the second Bhilsa inscription (page 458, pi. XXVI.) the figure 3 with another is perceived, following the word ^^?T and