526 
APPENDIX. 
affinities of the substantives themselves : since the personal 
pronouns are parts of speech which languages are slow to 
borrow from each other. 
To apply the observation to the details of the vocabularies 
under notice_, we find that the Juri and Javita prefixes are 
difierent from each other_, and different from those of the Tari- 
ana^ etc. throughout_, — the pronominal syllables being tch, and 
wa, respectively. Neither of these elements lead to anything 
very clear and patent in the way of affinity. On the other 
hand it is equally clear that in the Tariana^ Isanna_, Barre^ and 
Tomo-Maroa lists the prefix [-n~) is constant_, however much 
the roots which follow it may differ; so that the inference_, 
in favour of the possessive pronouns (at least) being the same 
throughout those four tongues^ is legitimate. In the Uainam- 
beu the case is slightly different. It is only in some of the 
words that the prefix is -n : in others it is -eri. This however 
is not very material_, since the two forms, in all probability, 
represent two persons, — nu = my, and eri = tJiy (or vice versa) 
respectively. 
Without, then, taking cognizance of the roots at all, the 
classification of the languages before us, according to the simi- 
larity of their pronominal prefixes, is as follows : — 
a. Allied. — Uainambeu, Tariana, Isanna, Barr^ Tomo- 
Maroa. 
b. Disconnected. — Juri, Javita. 
c. Uncertain. — Lingoa Geral, Coretu, Cobeu, Tucano. 
The geographical localities of these tongues coincide with 
the natm^e of their pronominal prefix, and favour [pro tanto) 
the notion that they all belong to one and the same class, — a 
class of which the value is at present wholly conjectural. 
How far are the roots themselves similar ? Upon the whole 
it may be said that where the geography and where the pro- 
nominal prefixes indicate affinity, the roots themselves do the 
same ; though not so clearly and patently as the investigator 
unpractised in American philology has a tendency to expect. 
