2 On the Poiver of the Letter jo. [No. 9, new series. 



it ; for instance with sesar ' a bank'; and, after inserting the link- 

 particle jjdr, we shall have the form ^ppii&eorr. In this form 

 the power of the ps would be said to be that of hard ' t's : and 

 yet, from analogy, this doubled form of the letter ought not to be 

 anything more than an intensification of the power of the single 

 letter. The object of the present paper is to endeavour to recon- 

 cile this apparent anomaly with the usually beautiful philosophical 

 uniformity observable in similar cases in the structure of the Tamil 

 language; and if possible, to arrive by this means at the true power 

 of this letter p. 



It is perhaps worth remarking thus early, that it is expressly 

 stated in the Nannool that this letter has no equivalent in the 

 Sanscrit alphabet ; and when it is remembered that the author of 

 the Nannool, by casting so much of his grammar in a Sanscrit 

 mould, has evidenced an intimate acquaintance with that language, 

 some reliance may be placed on his statement in this respect. 



It might be supposed to be conclusive that this letter is a hard 

 ' r' from the fact that it is usually called eueodHesr psn\l by Native 

 scholars, which expression literally means just so much : because 

 it would be tacitly assumed that the converse of this expression 

 would refer to the soft ' r\ The fact is, however, that the ex- 

 pression Qux&eSlesr jr<sjri2> would never be used by a Tamil scholar : 

 for this reason, that the expression euebsSI&sr pssnl refers only to 

 the fact that p belongs to the class of letters called oicuaW [hard] ; 

 while its corresponding Qldsqg&sw& [soft letter] is ' n' : the r 

 [soft ' r'] being referred to the QenL-uSlssr or ' middle letters' with- 

 out any relation to p. 



This threefold division of the consonants of the Tamil alphabet 

 is highly scientific, and speaks well of the patient investigation of 

 sounds, the correctness of ear, and the powers of classification of 

 its author in the far- distant age in which he lived. The names, 

 however, which he has handed down to us for these classes are not 

 precise enough for the present requirements of philology. Let 

 us substitute for them the common and more definite terms, 

 surds, nasals, and semi-vowels. For our present purpose we need 

 only to speak of the two former of these. Assuming this primary 

 division to be unexceptionable, we may tabulate the twelve con- 



