MEMOIRS FROM THE DEPARTMENT OF BOTANY OP COLUMBIA UNIVERSITY 17 



4. The genera here proposed would, under all circumstances, be treated at least as 

 subgenera under Potentilla, just as some of them were treated by Torrey and Gray in 

 their Flora. If they can be characterized and distinguished from each other as sub- 

 genera, there is no reason for withholding from them generic rank, and the classifica- 

 tion will be much simpler. 



5. As far as our American species are concerned the genera proposed here are well 

 defined and can be readily separated by their general habit as well as by their floral 

 characters.-^ They can be distinguished from the others much more readily than Ivesia 

 can from either Horkelia or Potentilla. 



Potentilla arguta and its allies differ from Potentilla proper by the nearly basal style, 

 ascending and orthotropous seeds, the flat disk-like anthers, the arrangement of the 

 stamens in five festoons, and in general habit. 



Potentilla tridentata differs by the lateral style, hairy achenes, ascending and amphi- 

 tropous seeds, shrubby habit, and leaflets that are jointed to the rachis and at last decid- 

 uous. 



Potentilla fruticosa and some Asiatic species differ by the same characters as the 

 last (except the leaves), by the arrangement of the stamens in five festoons, and the 

 scarious more or less sheathing stipules. 



Potentilla Aaserina, Egedii and ansciinoides by the lateral style, thick corky achenes, 

 ascending and amphitropous seeds, and (at least from all American Potentillae) by the 

 fact that they propagate like Fragaria by true runners. 



Ivesia depauperata and I. sabulosa differ from all the other Ivesiae by their narrowly 

 linear petals, the two pistils and the peculiar structure of the hypanthium, described 

 later. 



Ivesia santolinoides differs by its anthers, which are obcordate and open by a sub- 

 terminal pore, by the Potentilla-Yike flowers, solitary style and the general habit and pe- 

 culiar leaves. 



When comparing the original species of Horkelia and Ivesia, viz., H. Californica 

 Cham. & Schlecht, and I. Gordonii (Hook.) T. & G. with each other and with a typical 

 Potentilla, no one would hesitate to call them good genera. There are, however, species 

 so intermediate between Ivesia and the other two genera, that the differences are 

 nearly obliterated. This is especially the case with the distinctive lines between Ivesia 

 and Horkelia. Horkelia tridentata, H. congesta and H Howellii have only 4-15 pistils 

 and the two first only slightly dilated filaments. This is also the case with my H. ar- 

 gyrocoma, which is in all respects a good Ivesia, belonging to the same group as I. 



> Three of them were recognized hy Torrey and Gray as subgenera. 



