572 Transactions. — Miscellaneous. 



particular evidences is estimated. It was seen that, whereas a 

 close verbal resemblance between two nearly related tongues has 

 the balance of probabilities in its favour, one between only dis- 

 tantly related tongues, or those regarded as unrelated, has the 

 probabilities against it . . . There are, in short, two fun- 

 damental rules, under the government of which all comparative 

 processes must be carried on : — (1.) Comparisons must have in 

 view established lines of genetic connection ; and (2) the com- 

 parer must be thoroughly and equally versed in the materials of 

 both sides of the comparison. For want of regard to them, men 

 are even yet filling volumes with linguistic rubbish, drawing 

 wide and worthless conclusions from unsound and insufficient 

 premises." 



It is, I suppose, undeniable that the principles here laid 

 down are thoroughly sound. The concluding language is strong, 

 but not too strong where it applies ; and we ought to be put in 

 a position to show that it does not apply to what I have called 

 the new, but ought to have called the newest, method. 



And it is not only the charge of " insufficient premises" we 

 may have to meet. For it may be further objected that, while 

 in many cases the evidence is insufficient or altogether wanting, 

 in many others what evidence there is tends in the wrong 

 direction. 



Take, for instance, the statement already quoted : '• The 

 Maori kiri, ' the hide,' is English currtj, ' to dress hides.' " The 

 identity, I presume, is declared upon such similarity as there is 

 between them on the surface ; what is the evidence against it ? 

 Kiri, " the skin," is a very widely-spread, and therefore ancient, 

 Polynesian and Melanesian word. What is " curry "? According 

 to Skeat, following Littre, it is from old French con-roi, "gear, 

 preparation ;" a hybrid word, made by prefixing con (= Lat. 

 cum) to old French roi, order ; but this roi is itself of Scandi- 

 navian origin, from Danish rede, " order," or " to set in order." 

 It forms the second part of the word ar-raij : to " curry favour" 

 is a corruption of to " curry favel," to rub down, or get ready a 

 horse, of which favel was an old name. Now, if kiri was an 

 original Maori word, whilst cnrnj was coined in Europe within 

 historic times, it is evident that their identity can only be by 

 virtue of some extension of the doctrine of " pre-established 

 harmony" well worth elucidation. 



Again, Maori rauhi, " to seize," and ranr, " snatch," are 

 coupled with old English ravin, ''to obtain by violence," and 

 raven, " a greedy bird." But, according to Littre and others, 

 ravin comes from Latin rapina, whilst rarcn. Max Miiller says, 

 is from Sanskrit root RU (a general word for sounds of all kinds) ; 

 so that, it would seem, if we looked to find through the Sanskrit 

 a Maori relative for the raven, it should be not rawe or mu-hi, 

 but ruru, " the little owl." If, on the other hand, we were to 



