1867.] 347 [Chase. 



either to ignore the labors of early philologists, or to confine the ap- 

 plication of radical etymology within narrow limits. Independently 

 of all grammatical and other considerations, if there are satisAictory 

 evidences that the root ma is used to denote magnitude, ta to denote 

 tendency, cr to denote curvature, and r to denote radiation, in each 

 of two compared languages, it can be mathematically demonstrated 

 that it is unreasonable to attribute these coincidences to mere acci- 

 dent. Evidence may bo adduced as to their reality, arguments may 

 be urged as to their probable origin, but if their existence is con- 

 ceded, we should seek for some definite cause. A single accidental 

 coincidence may be probable; that there should be more than one is 

 improbable. 



Of the three comparisons which are quoted as samples of '' a batch 

 of atrocious identifications," one (idaro) is given without any com- 

 ment, as having no special importance ; one (abetele), by the re- 

 viewer's own confession, is based on essential roots so nearly identical 

 in form as be, bet ; and only one (bale, or basile) is considered in any 

 calculation of chances. Assuming this as a test example, let us see 

 to what extent the resemblance is traceable, and what probable in- 

 dications can be found of a common origin. 



I. In the first place, there are many evidences of ancient com- 

 mercial intercourse between the Yorubans and Egyptians, accom- 

 panied by traditions of settlements on the western coast of Africa by 

 Egyptian colonies. 



II. Such intercourse has been frequent in historical times, through 

 the wandering Arabs and the slave-traders, who have overrun nearly 

 all of Africa. 



III. There is some grammatical affinity between the Yoruban, 

 Egyptian, and Coptic languages. 



IV. While there are well-understood natural causes for such changes 

 as are instanced by the reviewer, from cjjislwpos to bisliop and eieque, 

 and from jilhis to hijo, it is difficult to attribute to any other cause 

 than a common origin, such analogues as Latin dare, fere, finis, 

 ibi, circum, curtun, rudimetdu'm, samis, teres, totus, and their Yoru- 

 ban synonyms da, fere, fing, ibe, kiri, Iciiru, rudi, sang, tere, tofo* 

 all of which appear to be derived from roots which are still retained 

 in the latter language. 



V. Among the aboriginal dialects of America, the Algonquin and 

 the Iroquois furnish a remarkable instance of the persistence of gram- 



* Loc. cit. pp. 45—6. 



