and Old World Birches. 189 



the coast of Alaska, the islands of Behring Sea, and adjacent 

 Kamtschatka this depressed forni with small orbicular or reni- 

 form leaves retains its characteristics in a marked degree, and 

 were the plant known only from that district it would stand as 

 an undoubted species. 



Mr. Coville has called the attention of the writer to a shrub 

 four to six feet high which is associated with the depressed 

 Betula glandidosa^ var. rotundifolia^ and the Cook's Inlet 

 tree which is identified with B. i?endida^ var. jcqyonica. This 

 intermediate shrub is well represented in the xs'ational Herba- 

 rium, and it is possible, as Mr. Coville suggests, that it is of 

 hybrid origin, since similar hybrids of trees and dwarf shrubs 

 have before been noted." The material presents a strong super- 

 ficial resemblance to Greenland specimens of B. atba^ var. 

 rniiior, although the strongly resiniferous branchlets hardly 

 place it with that shrub. 



In conclusion, it should be emphasized that the specific lines 

 in Betula^ as in Alnics^ Querciis and Salix, are often too vague. 

 It is quite possible to trace by a series of specimens a direct 

 connection between the dwarf Betula nana or B. glandidosa 

 and the tall B. alba. Thus B. nana in its larger development 

 is separated with difficulty from the Scandinavian B. alpestris. 

 This shrub, in turn, is quite like glabrate states of the Ameri- 

 can B. pumila., which, through its var. glandidifera, passes to 

 B. glandidosa^ the larger developments of whicli pass in the 

 Cascade Mts. to B. microjpliylla^ and in the Saskatchewan 

 region to B. alha, var. minor. The latter shrub is often 

 inseparal)le on the New England mountains from B. alha^ var. 

 cordifolia^ which on the lower slopes becomes a large tree and 

 passes gradually to the broad-leaved form figured by Michaux 

 as B. papyracea. A very similar series is I'eadily made to 

 include B. pendida and B. humilis. But since it is obvi- 

 ously impracticable to regard all these forms as one species, it 

 seems wiser to recognize the more marked centers of variation 

 as species which are admitted to pass by exceptional tendencies 

 to other forms ordinarily distinguished by marked charac- 

 teristics. 



The American representatives of § Costatae, Betula nigra, 

 B. lenta, and B. hitea, are represented by related species in 

 Asia, but none of these trees are of very boreal range, and 

 they appear well distinguished as endemic species. 



* B. inihescens [albci] x humilis, Warnst. Yerh. Bot. Yer. Brandenb. xi, 

 129 (1870). 



B. nana x verrucosa [pendula], Sael. Medd. Soc. Faun. et. Fl. Fenn. xiii, 

 256 (1886). 



B. nana x piibescens [alba], Koehne, Deutsclie Dendr. 112 (1893). 



B. pumila x lenta, Jack, Gard. and For. viii, 243, fig. 36 (1895). 



