cent within, the lobes either purplish or pale, « 
long, 2 min. fwide, the apex rounded, pubescent — 
glabrous inside, veined, apparently rotately Betis 
living state: ~ white edged with brown, the | 
broader than long: tities vari lanceolate, 1 mm. ke rar 
half as picks the short stout filaments half the noe of 
erect, iecioe paeenine beyond the petals and ane ‘sun | 
cent on the lower half, the stigmatic apex not lobed: ovary to- — Ri 
mentose and densely glandular with stalked glands. REARS 
The type, in my herbarium, is S. B. Parish’s no. 5564, col- 
lected April 14, 1906, on the northern slope of the San Bernar- 
dino mountains, San Bernardino county, California. Mr. Par | 
ish writes as follows concerning it: 
“All the specimens I have sent you came from the same 
individual shrub. It is a shrub of some four feet in height, 
branching from the base in four slender stems, and grows on the 
northern slope of the San Bernardino mountains at about 2500 
feet altitude, on a hill slope in what may be called an open 
growth of shrubs. 
“J consider it as identical with the shrubs on this side, 
which may attain a height of eight feet, with numerous rather 
slender stems, which occurs in the chaparral growth at from 
2500 to probably 5000, possibly 6000 feet altitude. ‘The shrubs 
are not numerous.” 
R. purpurascens is perhaps more related to R. viredifolium 
than to any other species, notwithstanding the great difference 
in general appearance. The latter species is about twice larger 
in all its parts, and in addition has a differently shaped leaf. I 
have typical R. vrzdifolium in my herbarium, and have also 
examined the type, in the herbarium of Stanford University. 
