136 BOTANICAL GAZETTE [AUGUST 
1858’; and upon them is written “Salix arctica R. Br. subal pestris 
And. (forte n. sp.).’’, ANDERSSON apparently changed the varietal 
name later to petraea. ‘The Kew sheet also bears, in the lower left 
corner, the inscription ‘‘Salix herbacea. Montagnes rocheuses 
Palouse prés les Glaciers, 18 aoit 1858.’’ According to MAcouN 
(Cat. Canad. Pl. preface, p. viii. 1883), BourGEAu “spent some 
time, in August 1858, in the Bow River Pass and the adjacent . 
mountains’? in Alberta. S. petrophila differs from S. anglorum 
chiefly in the color of the rather pale or grayish green leaves, which 
are not distinctly paler and never whitish beneath. The differences 
indicated by RypBERG between the two species are of no value, 
because his S. anglorum is mostly S. groenlandica. As I have 
already said, there are some forms in the northern habitat of 
petrophila which I have not yet been able to interpret properly. 
So far as I can judge by the specimens before me, the species 
ranges from about 52° N. latitude in southwestern Alberta and 
southeastern British Columbia through western Montana, north- 
eastern Wyoming, and central Colorado to the Truchas Peak in 
northern New Mexico. I have not seen specimens from Washing- 
ton and it is not mentioned in Pirper’s Flora. In eastern Oregon 
I know only of two localities. From Utah and Nevada I have 
seen very little material, and in California it is found in the Sierra 
Nevada from Sierra County to Tulare County. 
In western Nevada and the Californian Sierra, S. petrophila is 
mostly represented by a form which has been described as S. caespi- 
tosa by KENNEDY (Muhlenbergia 7:135, pl. 9. 1912). Through 
the kindness of Professor C. W. Lantz I have seen the type, which 
is preserved in the herbarium of the Agricultural Experiment 
Station at Reno. It was collected by the author on Mount Rose, 
Washoe County, Nevada, August 17, 1905 (no. 1173, fr.). It 
differs from typical petrophila in the more copious pubescence of 
the upper leaf surface, the acuter leaves, and the very short style. 
The last character seems very variable, and the type material before 
me consists only of fruits with withered styles and stigmas. Never- 
theless, I am inclined to use the name caespitosa for a variety which 
seems to be the prevailing form in the western part of the range of 
petrophila, and this var. caespitosa (Kennedy), nov. var., may be 
