BIRDS OP NORTH AND MIDDLE AMERICA. 25 



in warmer countries. In the latest special treatise on the family" one 

 hundred and nine species, belonging to twelve genera, are recognized, 

 of which thirty-one species and six genera* are American — all but one 

 of the former and two of the latter being peculiar to the Western 

 Hemisphere. 



The number of American genera of Hirundinidse is a question very 

 difficult to determine. As already stated, the most recent authority 

 allows but six — Progne, PetrocheUdo?i, Atticora, Stelgidopteryx^ CUvi- 

 cola, and Hirimdo — and these are- the genera which are generally 

 accepted. The second, fifth, and sixth of these are cosmopolitan in 

 range; the iirst, third, and foui'th being peculiar to America. It is 

 chiefly with respect to the first, third, and sixth that the question of 

 generic homogenity is concerned; and I am convinced that each of 

 these should be subdivided if we are to have generic groups which 

 are naturally circumscribed. Certainly Atticora, as usually under- 

 stood, is a most heterogeneous group, and cannot be defined by any 

 character or combination of characters, while most of the eight or 

 nine species composing the artificial group in question differ from' 

 one another so much in details of form that they may easily be segre- 

 gated into seven lesser groups (mostly monotypic), each of which 

 differs from any other in structural characters quite as tangible as 

 those which distinguish other recognized genera. On the other hand, 

 if left together, the group thus formed has nothing to .hold it together, 

 the species having little in common beyond the rouudish vertical 

 nostrils (shared also by Progne, PTiaeoprogne, Petrochelidon, and 

 Stdgidopteryx ') and the relatively long tarsus. 



The other group concerned is that included by Sharpe and others 

 in Ilirundo. This genus, properly restricted, is, in the main, an 

 exceedingly natural one, though its limits as to Old World species 

 (many of which I have not been able to examine) are uncertain; but 

 there is no question in m.j mind that the purely American types, 

 comprising species which, while differing much among themselves in 

 details of external structure, are all different in this respect as well as 

 style of coloration from all Old World species of Ilirundo, should be 

 removed from that genus. 



« A Monograph | of the | Hirundinidse | or | Family of Swallows. | By R. Bowdler 

 Sharpe, LL. D., F. L. S., F. Z. S., Etc., | . . [=6 lines of additional titles] | and | 

 Claude W. Wyatt, | Member of the British Ornithologists' Union. | —Volume I 

 [-II] I — I London: | Henry Sotheran & Co., | 37 Piccadilly, W. 1 140 Strand, W. 

 C. I 1885-1894. 2 vols., 4to, Vol. I, pp. i-lxx -l 1-356, pU. i-Hv; Vol. II, pp. 

 i-viii + 357-67S, pll. Iv-cxxix. 



^Both genera and species are, however, given wider limits by Sharpe and Wyatt 

 than the facts seem to warrant. 



c In this connection it should be remembered that the roughened edge of the outer- 

 most primary, so distinctive of adult males of Stdgidopteryx, often does not exist in 

 adult females. 



