68 PROCEEDINGS OF THE CANADIAN INSTITUTE. 



mean. Dr. Ellis would derive it " in common with the other 

 American tongues from Central Asia " (Peruvia Scythica, 1875). 

 The last two writers note its resemblance to the Basque. 



What the relations of the Araucanian family are has not yet been 

 satisfactorily settled. Prof. Campbell claims the Fuegians (a branch 

 of the Araucanian stock) as members of his " Khitan" family and 

 immigrants into America from Asia. Mr. Hyde Clarke claims to 

 have discovered a relation between the Fuegian and certain languages 

 of South Africa. He says : — " It is sufficient to state that these 

 gelations enter into a definite group, No. YIII. of Dr. Koelle's 

 Poly-Metta Africana, and bear a notable resemblance to the 

 Ngoten and Ekamtufulu languages. How such relations can exist 

 between Tierra del Fuego and South and West Africa, others may 

 explain. My own conclusions have often been made public " (Journ. 

 Anthrop. Inst, of Gt. Brit, and Ire., Nov., 1885). The existence of 

 a primitive Negroid people at the southern extremity of South 

 America is not very improbable. If Mr. Hyde Clarke's views be 

 correct, the relation of the Araucanian proper and the Pampean 

 families to the Fuegian will be of great help in solving the problems 

 of South American ethnology. 



But it is in Brazil and Guiana that the philologist and the ethno- 

 logist alike are puzzled. The multitude of languages in these regions 

 bewilders the investigator, but it is quite likely that proper study 

 and application will prove (as has indeed been tacitly assumed by 

 most writers) that they have all (or nearly all) sprung from the 

 primitive Cai'ib-Guarani stock. Lang (Polynesian Origins and 

 Migrations), comparing certain languages of Polynesia and Guiana, 

 would bring the Polynesians to Amei-ica, and the contrary theory 

 ias been held by not a few. 



A recent writer, M. Peixoto (Archivo do Museo nacional do Rio 

 de Janeiro, Yol. YI., 1885), arrives at the conclusion that "no type 

 vet examined in Brazil presents the essential characteristics of a 

 race ; a great mixing has been long going on among South American 

 peoples ; the plastic forms of the primoi'dial factors in the mixture 

 have long ago disappeared in the general fusion." The resemblance 

 in manners, customs and mythology of the Botocudos, Suyas, Yaruras 

 and other Brazilian tribes to those of certain African peoples has 

 often been noticed. It is possible that the ancestors of the Mosquito 

 Indians of Central America, of the Caribs of the West Indies and 



