io3 



somes) to subsequent gene mutation. This implies parallel muta- 

 tion, which is known to occur. Vavilov has shown in the most strik- 

 ing manner how parallel variations, through long series, arise in 

 related genera of plants, where no question of hybridization is in- 

 volved. On this basis a tetraploid AABB rose may have evolved 

 from a diploid AA, and the B septets it contains may only simulate 

 those of the diploid BB, without having any genetic connection. 

 This leaves the irregular septet forms as probably of hybrid origin, 

 and as Hurst remarks, several of them are actually known to be 

 hybrids. With regard to the fossil roses, it must be remembered 

 that the Florissant species date from the Aliocene, and occur with a 

 flora which is not more primitive than that of the present day. 

 Chaney'* has identified one of the Florissant species (Rosa hilliae 

 Lesqx.) in the Bridge Creek (Tertiary) beds of Oregon, his ma- 

 terial consisting of detached leaflets and a piece of stem with 

 prickles. There is no more basis for referring this material to R. 

 hilliae than to any one of several living species; nor, I think, any 

 reasonable probability that it belongs to Lesquereux's species. 

 Chaney speculates on the possibility that all the Florissant rose 

 leaves belong to one species, and suggests that R. rnskiniana Ckll. 

 represents the fruit of this species. As a matter of fact, R. rnskini- 

 ana was based on a bud and its fruit is unknown. I think there is 

 no doubt that R. hilliae and R. wilmattae Ckll. are quite distinct, 

 but R. scudderi Knowlton may be a variation of R. wilmattae and 

 the other two, based on a bud and immature fruit respectively, 

 probably belong with some of the leaves. 

 University of Colorado 

 Boulder, Colo. 



* Geology and Paleontology of the Crooked River Basin with special 

 reference to the Bridge Creek Flora. Publ. 346, Carnegie Inst, of Wash- 

 ington (1927), p. 123. 



