82 STELLER'S JOURNAL 



afterwards saw, were equally bare and without any woods. ^^s 

 I have not been able to find any explanation beyond the follow- 

 ing: 



(i) These islands have a twofold position. Those from here 

 on towards America bear northeast and southwest, while on the 

 other hand, those in the Channel and nearer Kamchatka^^* trend 

 northwest and southeast, and I have noticed both with regard 

 to the large and the small rocks that they strike in the same 

 direction. 



(2) In addition, all have the peculiarity that they are very 

 long and at the same time their breadth is quite out of propor- 

 tion, for example: Shumagin's Island ^*^ is from twenty to thirty 

 versts long [and] two to three broad, Bering's Island is thirty 

 miles long and only four, or at most seven, versts broad. All the 

 islands, of which we noticed seven ^^^ between here and Bering's 

 Island, were quite similarly formed. From this it follows that, 

 as they lie exposed to the north and south and consequently 

 suffer the most rapid changes of heat and cold, and on account of 

 such a slight breadth are moreover swept freely by the exceed- 

 ingly severe storms of these regions, neither tree nor shrub can 

 grow or get rooted. ^^y Even the smallest shrubs grow so crooked 

 and interwoven that it is impossible to find in the entire region a 



183 It is interesting to note this observation of Steller's on the western 

 Hmit of tree growth on the offshore islands of the Alaska Peninsula and 

 its continuation, the Aleutian Islands. The general features of the vege- 

 tation according to present knowledge may be studied on two maps by 

 the late Dr. A. H. Brooks (U. S. Geol. Survey Prof. Paper No. 43, 1906. 

 PI. 12, and Geogr. Rev., Vol. 15, 1925, PI. I, land classification map). (J) 



i8< By "those in the Channel and nearer Kamchatka" Steller means 

 Bering and Copper Islands and the westernmost of the Aleutian Islands 

 proper, ending in Attu. On "the Channel" see. above, footnote 149. 



18 5 Now Nagai Island (see Vol. 1, p. 142, footnote 86). 



186 Possibly the following are meant according to the identifications 

 made in the present work: Atka. Adak, Kiska, Buldir, two of the Semichi 

 Islands, Copper Island. (J) 



18' Steller is probably right in his theory that the lack of shelter from 

 severe winds accounts for the absence of timber on narrow islands w^hose 

 axes are parallel to the prevailing high winds. (B) 



