70 JOURNAL OF THE ARNOLD ARBORETUM [vol. ii 



Herbarium is one which Andersson himself named " Salix vagans cinerascens 

 lalijoliar In 1867 and 1868 he, however, states that the type of S. vagans 

 cinerascens rostrata latifolia was collected by Lyall " in ins. Vancouver." 

 The description runs: "foliis 2i 3 poll, longls, supra medium 1-H poll, 

 latis, longius acuminatis, tenuibus, subintegris, subtus molliter tomentosis; 

 stipulis parvis acutissimis." I have not yet seen a specimen of S. Bcb- 

 hiana from Vancouver Island, and Henry (Fl. South. Brit. Col. 98 [1915]) 

 says of >S. Bchhiana: "common east of the Cascades, Kamloops; Arm- 

 strong; Crows Nest Pass." It ai)parently docs not grow on Vancouver 



Island. 



S. Scoulcriana. 



i folia 



Of Lyall's specimens from the lower Frazer River before me none agrees 

 with /Vndersson's var. or f. lanaia which he, in 1867, characterizes as fol- 

 lows: "fuliis ex ovato-subcordata basi ovali-oblongis, crassioribus, li poll, 

 longis, integris, breve acuminatis, utrinque cinerco-sublanatis; capsulis 

 magnis a basi latissima ovatis; ramis nudiusculis, junioribus fusco-tomen- 

 tosis." In 1868 Andersson only says: "fol. ovato-subcordatis crassiusculis 

 utrinque dense tomentosis." S. Behhiana rarely has lea^'cs Avith an ovate- 

 subcordate base, and I have never seen young branches with a fuscous 

 pubescence. On the other hand it seems hardly probable that Andersson 

 misinterpreted this plant because he had seen the fruits. Otherwise one 

 might believe that var. lanata should be referred to S. Scoulcriana or some 

 other species. After all Andcrsson's varieties and forms mostly are very 

 uncertain. None of them can be regarded as really identical with one of the 

 varieties discussed later. 



As far as I can judge by the material I have seen there are two main 

 variations: the typical S. Behhiana {S. rostrata Richardson sensu stricto) 

 which is found in the north and east, and a western-southwestern variety 

 agreeing with Rydberg's S. pcrrostrata, and chiefly differing from the type 

 by its smaller, smoother leaves glabrescent below. In his original descrip- 

 tion of S. pcrrostrata Rydberg does not state the differences between it and 

 S. Behhiana. In 1906 (and 1917) Rydberg says of S. pcrrostrata: mature 

 leaves "thin, glabrous, faintly nerved," and of S. Behhiana: "mature 

 leaves firm, i)ul>escent or tomentose beneath, more strongly nerved." These 

 statements do not prove correct. The texture and nervation are often 

 much alike in both varieties, and only by the glabrousness and by the gen- 

 eral impression of the plant is it possible to decide whether it should be re- 

 ferred to the type or to var. pcrrostrata. 



The type of S. pcrrostrata came from the Black Hills in South Dakota 

 where it had been collected by Rydberg (No. 1018) near Hermosa in 1892. 

 Rydberg also referred to it a specimen collected by R. S, Williams at Daw- 

 son, Yukon Territory. Coville (1901) said: "Wiilc his bibliographical 



references indicate that the species is a segregate of hehhiana, the author 

 gives no comparison of distinguishing characters, I am unable to find in 

 his description anything to distinguish our Alaskan specimens from what 

 I take to be typical behhiana, whatever may prove to be the relation of 



