192 SEXUAL SELECTION : BIRDS. Part II. 



they will almost always have been exposed will cause 

 them to undergo, judging from a widely-spread analog^y, 

 a certain amount of fluctuatino^ variability. In this 

 case sexual selection, which depends on an element 

 eminently liable to change — namely the taste or admi- 

 ration of tlie female — will have had new shades of colour 

 or other differences to act on and accumulate ; and as 

 sexual selection is always at work, it would (judging 

 from what we know of the results on domestic animals 

 of man's unintentional selection), be a surprising fact if 

 animals inhabiting separate districts, which can never 

 cross and thus blend their newly-acquired characters, 

 were not, after a sufficient lapse of time, differently 

 modified. These remarks likewise apply to the nuptial 

 or summer plumage, whether confined to the males or 

 common to both sexes. 



Although the females of the above closely-allied 

 species, together with their young, differ hardly at all 

 from each other, so that the males alone can be distin- 

 guished, yet in most cases the females of the species 

 within the same genus obviously differ from each other. 

 The differences, however, are rarely as great as between 

 the males. We see this clearly in the whole family of 

 the Gallinacese : the females, for instance, of the com- 

 mon and Japan pheasant, and especially of the gold and 

 Amherst pheasant — of the silver pheasant and the wild 

 fowl — resemble each other very closely in colour, whilst 

 the males differ to an extraordinary degree. So it is 

 with the females of most of the Cotingidse, Fringillidse, 

 and many other families. There can indeed be no doubt 

 that, as a general rule, the females have been modi- 

 fied to a less extent than the males. Some few birds, 

 however, offer a singular and inexplicable exception ; 

 thus the females of Paradisea ajpoda and P. papuana 

 differ from each other more than do their respective 



