476 Rydberg: Studies on the Rocky Mountain flora 



linearis suhulata. Dr. Brand did not give any reason for this 

 change. Furthermore, he did not cite any specimens of his C. 

 iinctoria from CaHfornia and I have seen no specimens of C. 

 aristella from that state. Of the variety suhulata, on the contrary, 

 there are several collections from CaHfornia in our herbaria. 

 There is nothing either in Kellogg's description or in his figure 

 which would indicate that Dr. Gray had made a misinterpretation. 

 Kellogg's figure is drawn from a young, simple, undeveloped 

 plant, and the peculiar branching of the var. suhulata in age does 

 not show. Whether C. tinctoria and C. aristella should be united 

 into one species is another question, but in such a case the variety 

 suhulata should have been made the species, viz., C. tinctoria 

 Kellogg, and C. aristella a variety thereof ; and this for the follow- 

 ing reason: The variety suhulata is certainly found in the type 

 region of Collomia tinctoria, v/hile C. aristella apparently is not. 

 Seen from another standpoint, the local and more specialized C. 

 aristella must be regarded as the derivative of the more common 

 and less specialized C. tinctoria (i. e,, the var. suhulata). 



Brand transferred Gilia sinister M. E. Jones to Collomia 

 without having seen the plant. This was probably because Mr. 

 Jones placed it in the Collomia section and compared it with G. 

 aristella. But Jones also made the following statement: "This has 

 the general appearance of G. inconspicua, but without the basal 

 leaves." The relationship is also with G. inconspicua. Several 

 of the species of that group have the calyx enlarged somewhat in 

 fruit; this is true in G. sinister, but it is at last ruptured by the 

 capsule and does not have the structure of the calyx in Collomia. 

 It is in my opinion a true Gilia. 



Dr. Brand included a number of forms, in my judgment several 

 good species, under Polemonium pulcherrimum Hook. He divides 

 it in three subspecies, tricolor, delicatum, and parvifolium. The 

 first is separated by its tricolored flowers and equals P. tricolor 

 Eastw. The other two subspecies he separated only by the length 

 of the leaflets, a very poor character to use for separating sub- 

 species.* He overlooked the fact that in all these forms included 



* Under var. Haydenii Dr. Brand made the following remarks: "The forms from 

 the southern Rocky Mountains, which could be counted to this, are better to be 

 regarded as depauperate forms of subsp. delicatum." 



