170 JOURNAL OF THE ARNOLD ARBORETUM voL.t 



distinctly indicated, but this variety ohscura certainly does not seem to 

 belong to S. pyrifolia sensu stricto which is the same as Andersson's var, 

 lacfa in 1868. 



Bebb, who in 1879 dealt with S. balsamifcra and took up this name, dis- 

 cussed it again in 1888 when he proposed besides a var. typica three more 

 varieties. Of these var. vegeta, too, coincides with the type while according 

 to Fernald (in Rhodora xvi. IIC [1914]) var. lanceolata can be retained as 

 S, pyriJoUa var, lanceolata Fernald because it '' is found in some of the 

 swamps of eastern New Foundland to be so uniform and to occur in such 

 extensive areas that it seems well worth recognition." There is no type 

 in Bebb's herbarium in the Field Museum, and his figure (t. 81, fig. 7) 

 shows a leaf as it is often found in the typical form. Bebb's var. alpeslris 

 seems to be nothing but a lower small leaved form of which the leaves are 

 only partly green on both sides, as is shown by the lyi)e which was col- 

 lected by E. Faxon on Mt, Lafayette, White Mts., Eagle Lake, 1400 m,, 

 July 20, 1885 (sheet No, 276971 in C). 



As to the relationship of S. pyrifolia Bebb, in closing his remarks on S. 

 bahamifera (in 1888) says: *' A well defined species, not shading off into any 

 other, not even into Its nearest congener, S. cordata, from which it is always 

 distinguished by the peculiar texture and veining of the leaves, absence of 

 stipules and verj'^ loosely flowered fertile aments." As I have already said, 

 I believe that the most distinct character by which S. pyrisoUa may be 

 separated from the Conlafae is the short thick and blunt gland. All the 

 species of sect. Cordaiae known to me have narrow elongated glands which 

 look very different. Only one species, S. myrtillijolia^ has similar glands, 

 but as is the case with the other Cordatae its bracts are fuscous (at least 

 at the tip), and it docs not have the distinct nervation or reticulation of the 

 mature leaves so conspicuous in S. pyrifolia^ which besideshas the remark- 

 able balsamic odor of the unfolding leaves which it does not lose even in 

 the herbarium. 



Salix pyrifolia is principally found in the region from Newfoundland and 

 southern Labrador westward to the Winnipeg region; it also occurs (but 

 apparently sporadically) in Saskatchewan, British Columbia and the 

 southwestern part of the Northwest Territories. The exact limits of its 

 known distribution may be given as follows: Northwest Territories (be- 

 tween Lake Athabasca and Chesterfield Inlet, J. W. Tyull, No. 1715; 

 Hill Island Lake, halfway between Lake Athabasca and Great Slave Lake, 

 F. TlarpcTy No, 99021, O.), Saskatchewan (specimens of Drummond's and 

 Bourgeau's mentioned above), Manitoba (Lake Winnipeg region, Richard- 

 S072 (t Bourgeau; Bcrcns River, J. M. Macouriy No, 24252, 0.)> Ontario 

 (Kenora, Thunder Bay, and Carleton Districts), Quebec (Abitibi, Terre- 

 bonne, Chateaugay, Chambly, Stanstead, Champlain, Portneuf, Temis- 

 couata, Illmouski Districts, Magdalen Islands), Labrador (Sandwich Bay 

 Chdteau, Caribou Isl.), Ne^v^oundland (from the eastern drainage area of 

 the Iluniber River system to the Avalon Peninsula), New Brunswick 

 (Northumberland, Kent, Westmoreland, Sunbury and St. John Counties), 



