Oct. 1858.— Mae. 1859.] In Roman Characters. 205 



become obsolete from its incompatibility with the Arian 

 phonetic system and that the various changes it has undergone 

 in the examples quoted, are so many make-shifts to get rid 

 of it. If we wish therefore to represent it clearly, we must 

 assign to it a special sign and for this purpose, I am decidedly 

 of opinion that we should retain Ellis expedient of writing 

 zh, an arbitrary sign it is true, but one incapable of being 

 mistaken for anything else and at least as appropriate as any 

 other that has been suggested. For let any one follow ex- 

 actly the mode of pronunciation laid down by Native gram- 

 marians as explained by Beschi, Caldwell, and Pope without 

 any preconceived idea in favor of a particular letter and the 

 sound emitted by simply forcing the voice through the organs, 

 will be neither a liquid I nor a trilling r but a gushing utter- 

 ance which comes near to ajh or azh than any other com- 

 bination of sounds that I can conceive. 



Aeabic Alphabet. 



This is the most difficult character to deal with, both on 

 account of the larger number of sounds peculiar to itself, which 

 it contains, and because these are pronounced so differently 

 by different races* which make use of it. The Muhammadans 

 of India altogether fail in enunciating the true sounds of 



6, u°> u^i ^? which are for the most part reduced 

 by them to s and z, as among ourselves. 



I have already stated my objections to Wilson's represen- 

 tation of £ 



i 



J 



I also object to his exchange of h from to ^ as an un- 

 necessary innovation and as less fitting in itself: & has 



* The Egyptain Arabs pronounce the Jj as g. e. g. galh for qalb " the heart;'* 

 and the Basarah to and many of the Syrian Arabs invariably gave the sound of 

 ch to £ as chalb for kalb " a dog." 



