266 Substitution of the Roman [No. 8, new series. 



difficulty relates to the first two consonantal nasals, which are 

 commonly transliterated by ng and nj, and to Anuswdra. 



(a). I begin with Anuswdra. This is the character which in 

 Sanskrit, and in all other Indian alphabets except the Tamil, is 

 used in certain conjunctions as the common representative of all 

 nasal sounds. It is sounded like ng before gutturals, like nj be- 

 fore palatals, like n before Unguals, like n before dentals, like m 

 before labials and at the end of certain words ; and before h. where 

 it is most appropriately used, its sound is nearly that of ng. 



In reality this character is not at all required, the great latitude 

 of sound which is accorded to it perplexes, instead of simplifying 

 the alphabet, and each of its sounds might with great advantage 

 be represented by the consonantal nasal which is equivalent to it. 

 The substitution for it of the consonantal nasals is indeed to a 

 certain extent optional in all the Indian alphabets, and what is 

 optional in other languages is the rule in Tamil. The Tamil re- 

 jects Anuswdra altogether, using the consonantal nasals instead, 

 and gets on perfectly well without it. The best course would, I 

 have no doubt, be to follow the example of the Tamil in this mat- 

 ter ; as however the scheme now advocated is that of substituting 

 the Roman characters for the Indian, leaving Indian modes of 

 spelling untouched, the second best course is to provide a repre- 

 sentative for Anuswdra, which shall be used with the same lati- 

 tude until people learn to lay it aside. As we are not allowed to 

 change about from n to m as circumstances seem to require, but 

 must keep to the same letter throughout, it appears to me that m 

 is the most suitable character to be used, to be distinguished from 

 the consonantal m by a dot above. People will learn in time to 

 lay aside the dot where it is not required, as at the end of neuter 

 nouns. 



(5). The nasal of the guttural row of consonants which is pro- 

 nounced like ng in English may best, I think, be transliterated by 

 n with a dot above. This is the best representative of the Anus- 

 wdra before each A and of the obscure final Anuswdra of Hindi 

 words. In those instances people will naturally prefer n to m, and 

 that will help forward the abandonment of m. The objection to 



