VOl iS 1VI ] General Notes. 441 



more or less fundamental differences of structure. Moreover, the Troth i- 

 lidae bear to other birds much the same relation in this respect that birds 

 in general bear to other classes of vertebrates, for no other family of birds 

 is at the same time so numerous in species and so varied in the details of 

 external structure, yet, notwithstanding the extraordinary range of varia- 

 tion among the more than five hundred species composing the family, so 

 uniform in fundamental structural characters that no one has yet been 

 able to satisfactorily divide it into groups of supergeneric rank. Usually 

 there is little difficulty in segregating the Trochilidse into generic groups, 

 complying in all respects with the requirements of a genus according to 

 the generally accepted definition; and certainly Trochilus, Calypte, and 

 Selasphorus are groups which can be defined, this being really the best 

 test. Trochilus and Calypte both differ from Selasphorus in the forked 

 instead of rounded or graduated tail, and in entire absence of rufous from 

 the plumage, all the species of Selasphorus presenting, in both sexes, more 

 or less of rufous in the plumage and the tail of reverse form from that of 

 Trochilus and Calypte; while the two last named differ from one another 

 in the exclusive possession by Trochilus of abruptly reduced inner (proxi- 

 mal) primaries, with a subterminal angular projection to the inner web, 

 while the adult males have the pileum concolor with the back, the lateral 

 feather of the "gorget" short, and the lateral rectrices pointed; those of 

 Calypte having the pileum brilliantly metallic reddish purple or violet (like 

 the "gorget"), the lateral feather of the gorget elongated, and the lateral 

 rectrices rounded terminally and otherwise different in form. The very 

 natural and well-circumscribed group of nine species constituting the 

 genus Selasphorus contains no two species more closely allied than S. rufus 

 and S. alleni, except two of the Costa Rican forms ; hence, while everyone 

 (including myself) will agree that it would "be as reasonable to put S. 

 alleni in one genus and S. rufus in another, as to split up Trochilus [i. e., 

 the supergeneric group comprising Trochilus, Calypte, and Selasphorus] on 

 the basis of characters of no more weight than those separating these two 

 species," J I do not believe that anyone can be found who will claim that 

 S. alleni and S. rufus are as distinct from one another as either of them is 

 from species of Trochilus or Calypte. 



What is known concerning hybrids among birds, instead of supporting 

 Mr. Taylor's view that Trochilus, Calypte, and Selasphorus are not good 

 genera indicates, if anything, exactly the contrary. In the first place, it 

 may be fairly questioned whether hybrids are relatively more frequent 

 among the Trochilidse than in other families. Again, hybrids between 

 congeneric species are, so far as I am aware, invariably fertile (e. g., Hel- 

 minthophila pinus + H. chrysoptera, Colaptes auratus + C. cafer, Anas 

 platyrhynchos + A. rubripes, etc. 2 ) while those between distinct genera are 



1 Italics mine. — R. R. 



2 The list might be considerably extended, but this would open the way for a 

 controversy concerning specific characters! 



