1918] SCHNEIDER—AMERICAN WILLOWS 125 
at least in the Chilliwack Valley. This fact explains the confusion 
of the two species, and there is to me no great ‘“‘mystery how they 
came to be placed together on the same Kew sheet.” It is, how- 
ever, “more inexplicable how so critical a salicologist as ANDERS- 
SON should have been misled into combining the characters of the 
two in his S. subcordata.”’ I have also seen the “corresponding num- 
bers of the Hooker, BARRATT, and Torrey distribution in the 
Torrey Herbarium” mentioned by Bess. There are 3 sheets 
before me. One contains a large leaf and 2 sterile young branch- 
lets, and all 3 pieces belong to var. subcordata. This sheet bears the. 
following 2 labels: “No. 90. Herb. H.B. [&T., crossed out; instead 
of it is written beneath “‘fig.’’] S. obovata var. glabra,’ and “88 
Barratt, Rocky Mts. Ament leafy at the base about 4 leaves— 
Smooth and paler beneath.” Underneath the big leaf BEBB, in 
1887, has written “S. crassijulis Trev.?, S. subcordata And. in 
part.” The second sheet contains two flowering branchlets and 
bears the label ‘“No. 89 Herb. H.B. & T. Rocky Mts.” as 
well as the statement in BEBB’s handwriting “.S. swbcordata And. 
in part.” These flowering branchlets are identical with those in 
Herb. Kew. The third sheet bears the Barclayi form previously 
mentioned. 
ANDERSSON apparently had no clear idea of his S. subcordata; 
in 1858 he stated, “Quoad habitum quasi hybrida a S. cordata (cujus 
folia habet sed breviora) et S. glauca (amenta!).” In 1868 he 
referred to it some more specimens collected by BouRGEAU and 
DE La Pytatr, which I have not yet seen. The material before me 
looks very much like other robust specimens of S. arctica, but 
the leaves possess some stomata in the upper surface, at least 
along the main nerves. I think, therefore, that it is best to keep 
these forms as a variety of S. arctica, and I use ANDERSSON’S 
name. 
S. ARCTICA var. subcordata (And.) nov. var. seems to differ 
from typical arctica chiefly by the following characters: foliis 
maximis obovato-ellipticis ad 8:6 cm. vel obovali-oblongis ad 
722.5 cm. vel ellipticis ovali-ellipticisve ad 8.5:5 cm. magnis 
> RypBErG (Fl. Rocky Mts. 167. 1917) uses the name in a different sense. 
I am not yet quite sure what form is meant by him. 
