and on the Babylonian Lapidary Characters. 239 
gular of the imperative, corresponding to $: and d’hi, while tw and ¢d are, in 
the other persons, corresponding to 7, 7e, ¢u, ta. Lest, however, any one 
should persist in thinking that the letters which I call secondary should have an 
aspiration, because 6: and 6v, d’hi and d’hu, have, I observe, further that d, in the 
old Persian, in all its forms, primary and secondary, corresponds to the Greek & 
as well as to @; to the Sanscrit d, as well asd’h. The old Persian dadémi 
corresponds equally to déd@pe and to t’Snue; to the Sanscrit daddmi and dad- 
hdmi; and so do the secondary forms of d ; dhuzitiyam, as the Major writes it, 
duwitiyam, as I write it, corresponds to the Greek devrepov, the Sanserit 
dvitiyam; while his yatiya, my yad?, is the Sanscrit yadi. With respect to the 
alleged interchanges of Nos. 7 and 10, I observe, that if they be not errors of the 
sculptor, which I believe them to be, they are interchanges of di and #i, not 
interchanges of a primary letter with its secondary one. The instances are two. 
In one (III. 14), the imperative singular terminates in ¢i instead of di. The 
same sentence in the plural had occurred not long before, and the sculptor was 
apparently repeating it. After he had written the ¢ he perceived his mistake, 
but he could not well correct it: in another imperative, however, connected with 
this by the conjunction ‘‘and,” he wrote dz. The other instance is the name of 
a month, which occurs twice; in one place it is written with ¢, and in the 
other with di. It is uncertain which was right, and a confusion between sylla- 
bles so like in a long word should have no stress laid on it. 
The Bisitun inscription does not appear to me to throw any new light on 
the value of No. 12.* As to No. 18, it seems at first sight in favour of its having 
the value ¢h ; since, with a after it, it corresponds as a termination of pronominal 
adverbs to the Sanscrit tas or d’has (the former the more frequent, but the 
latter, I believe, the original form), and the Greek Sev (for Ses, as in the first 
person plural of verbs, ev for pes). To this, however, it may be objected that, 
* It may, indeed, be said, and with some appearance of reason, that the argument drawn 
from the Median transcription, ¢atsharam, is affected by the rectifications of the Median alphabet 
made in the present paper. I now read the word ?’.ta.sha.ra.m. Consequently, I cannot rely 
on the argument drawn from it, to prove that No. 12 was notj. This, however, is abundantly 
evident from other considerations ; and the argument that it was 22, and not 2, appears to me 
unaffected by the change. 
