1903.] NATURAL SCIENCES OF PHILADELPHIA. 657 



many cases, a number of slightly differentiated local forms, sometimes 

 susceptible of being banded together as races of one specific group or 

 sometimes so intricately interrelated as to involve several previously 

 well-established "species," and to render any arrangement in groups a 

 matter of the closest study and more or less arbitrary decision. 



No one description will accurately cover all of these related races, 

 and as the specific descriptions of older writers are often based entirely 

 upon one race or form, there is obviously no course but to recognize 

 the others on an equal basis. 



Whether or not our present system of nomenclature will prove 

 inadequate for the purpose remains to be seen, but under any circum- 

 stances the recognition of these forms which nature has differentiated 

 in a greater or lesser degree will be inevitable, as it is becoming obvious 

 that they, instead of the clumsy specific aggregate, are the funda- 

 mental units of systematic work. 



The existence of these "variants" among vertebrates was first 

 clearly accepted in the Check List of North American Birds, issued by 

 the American Ornithologists' Union in 1886, and a system of trinomials 

 was adopted by which they could be independently designated and 

 at the same time their relation to their specific aggregate denoted. 

 This plan has been largely followed in vertebrate zoology since this 

 time, but as material and knowledge has increased the difficulty of 

 grouping forms in specific aggregates, in such a manner as not to do 

 violence to the proper function of a name on the one hand and to a 

 fact of evolution on the other, has so increased that by some writers 

 trinomials have been all but abandoned, and binomials employed to 

 designate every form, no matter how slightly differentiated. In botany 

 where the trinomial system has never been so widely adopted as in 

 vertebrates, almost all the recently named forms are designated as 

 species. 



AsH have elsewhere stated,^ I am convinced that the use of tri- 

 nomials is still our best method for denoting these races, though not 

 on exactly the basis originally proposed in the American Ornitholo- 

 gists' Union Code of Nomenclature. 



The results of my study of variation in the genus Viola I have 

 considered under three headings: Racial Variation, as exhibited in 

 plants; ^'ariation in the Genus Viola; and a Synopsis of the Violets of 

 Philadelphia and vicinity. For the sake of comparison I have pre- 

 ceded these with a brief resume of racial variation among terrestrial 

 vertebrates. 



' Condor, 1903, p. 43. 

 42 



