232 CLASSIFICATION OF ALGONQUIAN TRIBES [ETH. ANN. 28 
gest Passamaquoddy k—lpen, which might be taken for k—ulpen, but 
as a matter of fact the w has nothing to do with the termination; 
owing to the phonetics of the language if a vowel following 7 is elimi- 
nated, thereby causing the J to become final or immediately to pre- 
cede a consonant, the preceding vowel takes an o or a u tinge (see 
the discussion of Eastern Algonquian, p. 283). Now isit not possible 
that there is a similar phenomenon in Piegan and that the termina- 
tion should really be given as ki—apinnan®, in which the x represents 
a secondary change of original n, as does the 1 of the Passamaquoddy 
form?! The same query would apply to certain other forms not 
dealt with here. 
To judge from Tims, the termination for we(excl.)—him agrees in 
formation with Cree and Ojibwa. The agreement with the latter is 
no doubt purely fortuitous. 
Forms like nestoa (Tims) 1 show agreement with Cree. 
According to the writer’s information some demonstrative pronouns 
have reference to the state of the object designated, that is, whether 
at rest or in motion; but some informants contradict this. It is a 
matter that deserves special attention. 
Summing up, we may say that though Blackfoot must be classed 
apart from Eastern-Central Algonquian, it has the closest affinities 
to Fox, Eastern Algonquian, and Cree. 
CHEYENNE - 
Cheyenne possesses consonantic clusters, though not in so great 
profusion as Piegan. By consulting the various tables it will be seen 
that some of the clusters are peculiar to the language. As is men- 
tioned more than once in this paper, the fact that such Algonquian 
languages as have numerous clusters differ with respect to the types 
of clusters tends to show that most of these are unoriginal. 
The following clusters of two consonants were noted in three of 
the writer’s Cheyenne texts: 
er, Second member of cluster 
Tnitial 
con- |—— T = 
SOnaay k t n s c ts v 
= | 
t tk tts 
n ns ne 
8 sk st Sts 
c ck ct : cv 
ts | _—otsk isn 
x xk at rs rts 
v vs ve 
1 Or it may be that the original sound is lost and that the z is an accretion, as x in ki—irpuwadwa yE— 
ME (cf. Fox ke—ipwa), 
