368 ANALYSIS OF GENERA AND SPECIES 



Wbite-bellied and Violet-green Swallows, which I ranged 

 together under Tachycineta. Though this may seem exces- 

 sive subdivision, it is difficult to get along with fewer genera, 

 if we are to accept even such as Cotyle and Petrochelidon ; for 

 the ultimate modifications of structure and details of form are 

 as appreciable here as in the cases in which, in other families, 

 generic groups are established. Respecting the extralimital 

 forms, I may remark, that CallicJielidon includes two beautiful 

 velvety or lustrous greenish and golden species, white below, 

 G. cyaneiviridis and C. euchrysea, allied to our T. hicolor and 

 T. thalassina; Atticora and its subdivisions embrace a number 

 of diminutive and very plainly colored species, somewhat re- 

 sembling Swifts ; while Phceoprogne includes some large South 

 American Martins, like Cotyle in dullness of coloration, but 

 near Progne in form. 



The seven established North American species all occur in 

 the Colorado Basin. They may readily be determined by the 



following 



Analysis of North American Genera and Species 



1. Tail deeply forficate, witli linear lateral feathers; lustrous steel-blue 



above, rufous below Hirundo eryihrogasira. 



2. Tail simply emarginate ; lustrous green ; beneath white. 



Tachycineta hicolor. 



3. Tail simply emarginate ; opaque velvety-green ; beneath white. 



Tachycineta thalassina. 



4. Tail nearly even; lustrous steel-blue ; rump rufous. 



Petrochelidon lunifrons. 



5. Tarsus with tuft of feathers below; lustreless gray; below white. 



Cotyle riparia. 



6. Outer edge of first primary serrate ; lustreless brownish ; paler below. 



Stelgidopteryx serripennis. 



7. Bill very stout, curved ; male entirely lustrous blue-black. 



Progne purpurea. 



Now that the thoroughly unnatural order " Fissirostres " 

 has been abolished by nearly universal consent, after enduring 

 long in the teeth of frequent protests from scientific ornithol- 

 ogists, it is unnecessary to more than allude to the prepos- 

 terous notions of classification which caused these strictly 

 Oscine Passeres to be classed with the Swifts and Goatsuckers 

 (Cypselidce and Caprimulgidw). What little resemblance may 

 be traced between the strictly Passerine Hirundinidce and the 

 Picarian families just named, results from purely adaptive 

 modification, the respective types of structure being radically 

 diverse. 



