86 Rydberg: Notes on Rosaceae 



the type nor any authentic material. The habit, judging from 

 Lehmann's plate, suggests, however, this group and is most like 

 that of P. arizonica. 



Multifidae 



This group, as treated in the North American Flora, contains 

 1 6 species, of which P. paucijuga and P. lasiodonta are new. 

 The former is probably most nearly related to P. pseudosericea, 

 but the stem and petioles are silky-villous with spreading hairs, 

 as in P. pnlchella, and the inflorescence is open. I have seen 

 no specimen except the type, which in habit resembles somewhat 

 P. rubripes, but the styles are quite different and place the plant 

 in this group. The type specimen is mounted on the same sheet 

 with four specimens of P. propinqna. 



Potentilla lasiodonta was based on material distributed as 

 P. pennsylvanica and P. strigosa. It has the broad leaflets of the 

 former and the pubescence of the latter, but differs from both in 

 the dense inflorescence, similar to that of certain species of Dry- 

 mocallis, and in the numerous lanceolate divisions of the leaflets. 

 These number 15-21, while in P. pennsylvanica and P. strigosa 

 there are 9-13. The following specimens belong here: 



Alberta: Calgary, July 21, 1897, Macoun 167 16. 



Saskatchewan: Tramping Lake, Aug. 4, 1906, Macoun & 

 Her riot 6q8io. 



Manitoba: Rapid City, 1896, Macoun 12576. 



The Multifidae are one of the most difficult groups of the 

 North American species of Potentilla. The difficulty is to draw 

 lines between species. A conservative botanist might admit 

 but four or five species, viz., P. pulchella, P. bipinnatifida, P. mul- 

 tifida, and P. pennsylvanica, and maybe P. virgulata. Nelson, 

 in the New Manual of the Central Rocky Mountains, admits 

 P. bipinnatifida Dougl. (which he, however, has renamed P. pin- 

 natifida Dougl.), P. pennsylvanica with two varieties, P. atro- 

 virens, and P. virgulata. P. pseudosericea, he makes a synonym 

 of "P. pinnatifida." P. pulchella, P. multifida, and their allies 

 are not found in the Rocky Mountain region. Dr. Wolf admits 

 P. pulchella, P. pseudosericea, P. bipinnatifida, P. multifida, P. lit- 

 oralis, P. pennsylvanica, and P. glabrella. Note the differences of 

 opinion. Nelson regards P. atrovirens and P. virgulata as good 



