[Vol. 8 
180 ANNALS OF THE MISSOURI BOTANICAL GARDEN 
North Dakota: on gravelly, hilly plains, Dunseith, Rolette 
County, June 3, 1911, Lunell (Rky. Mt. Herb.); stony barren 
summits of hills, Butte, Benson County, May 14-June 4, 1905, 
Lunell (U. S. Nat. Herb. and Rky. Mt. Herb.); Leeds, Benson 
County, June 5, 1909, Lunell (Rky. Mt. Herb.); Leeds, Benson 
County, May 28, 1901, Lunell (Rky. Mt. Herb.) ; railroad banks, 
Leeds, June 10, 1899, Lunell (U. S. Nat. Herb.); dry hills, Butte, 
Benson County, May 20-June 10, 1906, Lunell (Geo. Osterhout 
Herb., Deam Herb., U. S. Nat. Herb., and Rky. Mt. Herb.); dry 
hills, Towner, McHenry County, May 29, 1908, Lunell (Geo. Os- 
terhout Herb., U. S. Nat. Herb., Deam Herb., and Rky. Mt. 
Herb.); on gravelly hills, Minot, Ward County, June5and 6, 1909, 
Lunell (Rky. Mt. Herb.); on sunny slopes, Williston, May 2, 
1906, Lunell (Rky. Mt. Herb.); Dickinson, June 30, 1912, Wal- 
dron 128 (Deam Herb.). 
South Dakota: Date, Perkins County, June 15, 1912, Visher 
571 (Rky. Mt. Herb.); slopes of Cedar Pass, Stanley County, 
June 6, 1914, Over 6268 (U. S. Nat. Herb.); prairie, Newell, 
May 12, 1913, Carr 8 (Mo. Bot. Gard. Herb. and U. S. Nat. 
Herb.); Hot Springs, June 11, 1892, Rydberg 533 (U. S. Nat. 
Herb.). 
L. arenosa is likely to be confused with no other species except 
L. argentea, and from that it is distinguished chiefly by the very 
slender stems and the much smaller size. The leaves in arenosa 
are broader in proportion to their length and more frequently 
toothed than those of argentea. The pubescence in the smaller 
species is less dense than in the larger one. Finally, the ranges 
of the two, although apparently overlapping to some extent, 
are in general rather definitely separated. It may be found im- 
possible to retain L. arenosa in specific rank and the treatment 
accorded it by Watson may be ultimately reinstated. No true 
intermediates have been seen. 
The phylogenetic position of this species with relation to L. 
argentea is not clear. The characters separating the two are 
slight but L. arenosa resembles L. recurvata, the parental type 
of this group, more than does L. argentea. The geographical 
position of the two, however, would indicate that L. arenosa is 
derived from L. argentea rather than the reverse. Several 
gregates have been proposed but the differences they represent 
