EVOLUTION OF THE COLORS OF BIRDS. 251 



latter are brown. In the latter species, moreover, the 

 wing coverts, scapulars and back are streaked with white 

 which is wanting in the former. In spite of these con- 

 stant specific distinctions, certain characters vary from 

 one species to the other in an unbroken succession. 

 Thus Plate VIII shows the outer tail feathers of three 

 forms, and the uniformity of the reduction in the 

 white from the eastern species through the two varie- 

 ties of the western is very noticeable. This reduction 

 in white is most pronounced on the northwest coast and 

 is undoubtedly due to the direct action of the environ- 

 ment, being paralleled to a certain extent by the Florida 

 race (P. erythrophthalmus alleni). Reference to the. 

 lower map on Plate IX will show how the two species 

 were probably marked off by the Rocky Mountain ice 

 cap, so that the white streaks on the back of the western 

 species may be due solely to the influences of geographi- 

 cal isolation and the indirect effects of a different en- 

 vironment. They would be of use as recognition marks 

 and so be encouraged by natural selection. 



The different races of the two species have been mod- 

 ified principally by climatic influences, P. maculatus 

 ardicus (B 2) of the Great Plains being the form with 

 the maximum of white markings, P. maculatus megal- 

 onyx (3) having the white markings much restricted, 

 and P. maculatus oregonus (4) of the Northwest Coast 

 region with the Avhite markings lacking or much re- 

 stricted. The fact that the Mexican form of the genus 

 is so light colored is not easy to explain. The fol- 

 lowing hypothesis may throw some light upon it, how- 

 ever. The Pipilos were a northerly genus probably with 

 no white on the back, and simply with the Avhite recogr 

 nition marks at the tip of the tail. They were crowded 

 south by the ice into two districts, one to the east, the 

 other to the west of the Rocky Mountains. Those on 



