4 On the Power of the Letter p. [No. 9, new series. 



followed by another consonant of the surd class, it bears its hard 

 pronunciation : for instance, « in such circumstances is the equi- 

 valent of the English * k.' 



"When a surd follows its corresponding nasal, or any other con- 

 sonant of the nasal class, it bears its soft pronunciation : « in 

 such a position is equivalent to the English ' g.' 



Lastly, when a surd follows a vowel or consonant of the semi- 

 vowel class, it bears its aspirated pronunciation : our s has then 

 a light guttural pronunciation, much lighter than the Welsh or even 

 the German * ch.' 



The][different surds are susceptible of the aspirated pronuncia- 

 tion, however, in different degrees : in the case of the cerebral i— 

 the aspirated pronunciation is imperceptible, and undistinguish- 

 able from its soft pronunciation. : in the case of the labial u that 

 aspirated pronunciation ('ph') begins to grow distinct in some 

 words : in the case of the palatal & it assumes a more constant 

 form : while in the cases of the guttural a, and the dental 

 this aspirated pronunciation is perfect and invariable. The other 

 two pronunciations, namely, the hard and the soft, are invariable 

 admitting of no exceptions and of no degrees. 



Applying these rules to the letters jd and ear, we find that jd, 

 when it occurs doubled in the middle of a word, has in common 

 usage a pronunciation which at present it is sufficient to say par- 

 takes largely of a ' t' sound ; and when it follows its corresponding 

 nasal «r it submits to the same rule as the other surds, and softens 

 itself into a kind of * d.' When, however, it follows a vowel, in- 

 stead of assuming some kind of aspirated pronunciation of a * t/ 

 singularly enough common usage gives it the pronunciation of a 

 strongly trilled ' r.' Is there not in this exception an intimation 

 that popular usage has in some way corrupted the true pronuncia- 

 tion of the jd in this position ? for it does not seem possible to 

 suppose that a trilled 1 r' can be the aspirated pronunciation of 

 any description of 4 t.' 



2. Let us put this in a different form. The true radical power 

 of the Tamil consonants is to be discovered in the pronunciation 

 which they bear as the initial letters of words, when they are un- 

 influenced by the final letters of preceding words. We shall have 



