42 



lobes when present entire," a feature apparently not found in 

 my type of H. saxicola. 



Apparently referable to this species is my No. 591 1, col- 

 lected on the southern slope of Mt. Sanhedrin, Lake county, 

 California, July 19, 1902, distributed as Holodiscus dumosa^ 

 although at the time I suspected it would prove to be an unde- 

 scribed species. These Lake county specimens have rather 

 longer, more cuneate leaves, which are m.ore pubescent, being 

 quite silvery underneath; the panicles are somewhat denser, and 

 the flowers deeper colored. It grew at an elevation of 5000 feet 

 or more, in a region which has yielded a number of Sierran spe- 

 cies, though separated by the broad reaches of the Sacramento 

 Valley. 



Boisduvalia sparsiliora 



Simple or branched, 2-5dm. high, the branches when pres- 

 ent beginning from the base, more or less outcurved, then as- 

 cending, nearly 3dm. long in large plants, but always shorter 

 than the central stem, pubescent with short appressed hairs, or 

 above somewhat woolly: leaves on the lower parts of the main 

 stem and branches linear-lanceolate, entire, 3-4cm. long, 5mm. 

 wide; those of the upper parts shorter, with an ovate base, all 

 acuminate-pointed, shortly appressed pubescent on both sides: 

 floriferous throughout, a flower usuall}^ in the axil of each leaf: 

 petals rose-purple, 5mm. long, twice the length of the calyx: 

 capsules oblong, about icm. long, 2-3mm. wide, slightly curved> 

 a little narrowed at both base and apex, appressed pubescent- 

 seeds pale greyish-browTi, smooth and glabrous. 



No. 7021, collected July 25, 1903, in moist ground at the 

 lower end of Donner Lake, Nevada county, California. A spe- 

 cies apparently related to B. siricta.. 

 Boisdiivalia iiiabricata (Greene) 



Boisduvalia densifiora var. imbt^icaia Greene, Fl. Francisc. 

 225. 1891. 



As Professor Greene says, this is indeed "a plant of very 

 different aspect" from B. detisifiora^ which has been stretched 



