172 Bulletin American Museum of Natural History. [Vol. XXXVI, 



regard to what constitutes a generic character. At present we are without 

 such a standard. The genus of one author is the subgenus of another, and 

 is not recognized at all by a third. In many instances, therefore, a genus 

 becomes merely a personal expression of opinion concerning the taxonomic 

 value of certain admitted characters. The same species may be referred 

 to a dozen or more different genera by as many writers, all of whom may 

 agree on the details of structure and relationships involved, and disagree 

 only on questions of nomenclature. 



If the ultimate object of systematic zoology were classification based 

 only on analysis, we might be warranted in carrying dissection and descrip- 

 tion to their utmost limit and applying to the results as many names as the 

 most minute differences discovered seemed to require. It is, however, 

 commonly agreed among biologists that the primary object of systematic 

 zoology is to provide a nomenclature which can be used with some degree 

 of precision, and which shall be based not wholly upon analysis but to a 

 degree upon synthesis as well. 



So far as species and their geographical races are concerned, the trinomial 

 system of nomenclature permits the systematist to recognize but slightly 

 differentiated forms by name without obscuring their more essential rela- 

 tionships. He may not express lines of descent, even if they are known, he 

 may indeed name first the most recent offshoot of a certain stock, but in 

 the naming of a subspecies he does not disguise its group relationships. 



In our treatment of species, intergradation, known or probable (see 

 beyond under Treatment of Subspecies) is the test which determines sub- 

 specific status. But in our treatment of genera, this test is largely ignored. 

 We have, it is true, subgenera, but so far as practical nomenclature is 

 concerned we are trinomial with species and binomial with genera. In 

 consequence, many of our generic terms are just as false, just as misleading, 

 just as far from conveying an idea of actual relationships, as though we 

 were to use a binomial for every subspecies. They are, indeed, more 

 misleading since in the latter case the name employed would indicate at 

 least generic relationship, while in the former all suggestion of relationship 

 may be lost. 



I am aware that the two cases are not wholly comparable, and that 

 strictly to apply the test of intergradation to generic groups, would, in 

 some cases, place in the same genus species, which in the light of our current 

 understanding of what constitutes a genus, could not be considered as 

 generically related; and thereby lead to nomenclatural results as undesirable 

 as those based on excessive analysis. But I am also aware of how erroneous 

 an impression may be given by unduly emphasizing differences which are 

 obviously of less importance than resemblances, and then, chiefly for con- 



