Secrion II., 1900 [ 205 ] Trans. R. $. C. 
VI.— Mexican Colonies from the Canary Islands traced by Language. 
By the REVEREND JoHN CAMPBELL, LL.D. 
(Read May 29, 1900.) 
Many years ago, in some contributions to the “ Canadian Naturalist” 
of Montreal, I pointed out a large Celtic element in the dialects of Peru, 
and notably in that of the Aymaras. At the same time, I drew attention 
to the Celtic features of the Berber speech of Northern Africa, and of the 
related extinct dialect of the Canary Islands, without suggesting any 
connection between these and the.Peruvian. My philological inductions 
relative to the ancient peopling of this continent, both north and south, 
led me to reject any and every theory of an oriental derivation prior to 
the visits of the Northmen, including that of the Eskimos from Northern 
Europe, and that of the late Horatio Hale, which brought the Huron- 
Iroquois-Cherokee family from Biscay. The latter I have completely 
disproved ; first, by the coincidence of Siberian expressions in use among 
the Koriaks and Tchuktchis, thoroughly differentiated from the Basque, 
with similar expressions in Iroquois; and secondly, and more definitely, 
by at least two inscriptions, the one Qatoghie Huron, the other Cayuga 
Iroquois, which are pure Japanese, the classical or ecclesiastical language 
of these peoples, which, while of the same origin as the Basque, is of 
vastly simpler construction. The relations of the Peruvian vocabulary, 
or vocabularies, I was able to illustrate by partial comparison with the 
Japanese on the one hand and the Basque on the other, but a large num- 
ber of words failed to adapt themselves to the Procrustean bed of either. 
The Peruvian grammatical system is essentially Turanian, agreeing in 
general terms with that of the Basque and the Japanese in the matter of 
postpositional order, common to both of these languages. But I failed 
to give due weight to certain important features by which Peruvian 
grammar is differentiated from Japanese, and even from Basque. Peru- 
vian grammar has a genuine plural in cuna, embracing the Basque plural 
in ac and the old Khitan or Hittite plural in na; while the Japanese has 
no real plural. On the other hand, there is no polysynthetic incorpora- 
tion of the verb substantive with pronouns subjective and objective, both 
direct and indirect, as in Basque ; but the Peruvian employs pronominal 
suffixes, as do the Celtic dialects, which neither the Basque nor the Japa- 
nese do. 
So recently as last May my scientific position was set down in the 
following words, contained in a paper on “The Kootenay and Tshimsian 
Languages of British Columbia,” read before this section of the R. S. C. : 
‘“ East of India and Bokhara, the only people that is supposed to connect 
