260 



The Upenmwik Greenlanders, then, have carried this change 

 fartherst since the time when it began. 



Of great interest is the fact that tlie Labrador dialect has 

 taken part in the change of J to s, namely in a number of 

 those words where j was preceded by t (N Al. fj). From that 

 I conclude that this change has begun in the words which 

 contained tj. Al this point the Labrador dialect came to a 

 standstill without having completely carried out the change. 

 Therefore this dialect has sometimes ts. and sometimes tj in 

 those words where Greenlandic has got ts^ while the more 

 western dialects in Alaska have tj and, just like Labrador, j 

 in those words, where WGr. has got s. Where the western 

 dialects have ts^ these sounds have been preserved in the L. 

 dialect, while Greenlandic through more advanced assimilation 

 has generally got s- (ors), cf. the examples cited in g 32. This 

 change of sound in the East Eskimo dialects (tj^ts) must ac- 

 cordingly have begun shortly before the separation of the L. 

 and S Gr. dialects. In other words , the tendency to change 

 the sound j to *•, a change which the present Greenlanders 

 have extensively carried out, had already begun in the districts 

 on the other side of Smith Sound while the ancestors of the 

 South Greenlanders were still closely connected with the Labra- 

 dor tribe. It may be dated as the time just before the invasion 

 into Greenland in the IS**" С 



The change of ^ to s in Greenlandic is probably some- 

 what later. 



The next phonetical phenomenon to be taken up is of just 

 as great importance for the establishment of boundaries between 

 the dialects as the one we have just left. It is the metathesis 

 or retrogressive uvularization that has been treated in g 33. It 

 seems as if the boundary -line drawn by this phenomenon 

 must lie between the Mackenzie River and Labrador; for the 

 present I assume it to be the west side of Hudson Bay. The 



