sign was followed by either a or f, and that a different sign was nsed before u. Taking these facts into consideration, Eawlinson thought he observed sufficient regularity to justify him in formulating the general law that, for some unexphiined reason, the grade of surds in each class were expressed by two signs, one used before a and 2, the other l^efore a; the grade of aspirates by one sign only, equally available before any of the thn*e vowels; and the grade of sonants by three signs, each applied to one vowel only. Tie admitted that there were numerous exceptions to the rule; indeed, the class of dentals is the only one where the series is complete, but the exceptions he was inclined to attribute chiefly to the incompleteness of the alphabet.'
When the letters of Ivawlinson's original alphabet were distributed into the various classes of gutturals, palatals, and so on, and among the subdivisions of surds, aspirates, and sonants, they were found sufficient to suggest the existence of some such law in the cases that have been named. With Holtzmann's correction of 19 (^yy) from t to f/, he had the three cVs required to complete his sonants of the dental class. His own list gave him the tliree nis required for the sonants of the nasal class; and he alreadv knew that one was used onlv bc^fore /, and the other onlv before u. He knew also that the as^nrates of three of the classes were to be found indifterentlv before anv vowel. In the case of the surds, he had found that five of them \k y^, ch ^, t ^yyy, n ^^. r ^y) are always to be found before either an a or an /. lie knew also that his second sign for /? {^ (43) and his second sign for r
- -^^ were only to be found before //; and he ol)served
that the signs he still read [^ kh and (yyy^ th were also oidy found before it. It required, therefore, no
' j.n.A.s. X. 176.
