306 THE ANIMALS OF NEW ZEALAND 



objects, especially as in P. chalconoius the changes have been 

 in the opposite direction; and not only have the white alar and 

 doi-sal bars been lost, bnt the whole of the under surface has also 

 turned black. P. chalconotus and P. steivarti live together in 

 Stewart. Island, and even inhaliit the same shaggeries. It is, 

 therefore, impossible for these opposite variations to have been 

 caused by external conditions. 



The origin of the white alar bar in P. cami^hclU, or the white 

 dorsal patch in P. steivarti, being new characters, cannot be due 

 to reversion. Can they be accounted for by sexual selection? 

 What right have we to assume that a preference was shown by 

 individuals of the opposite sex to one that had some white 

 feathers on its wing or back? Even if this was the case, what 

 guarantee is there that successive generations would all show 

 the same preference? Unless they did do so, the selection would 

 he destroyed, and the variation would not accumulate. The 

 bright colour of the skin on each side of the face in cormorants 

 will be looked upon as a typical example of sexual selection ; but 

 the same difficulty occurs here also. Why should both sexes 

 prefer the same colours in their partners as they have themselves, 

 although they cannot see their own colours? Why should the 

 preference for one particular colour, as for crimson in P. onslowi, 

 go on for generation after generation in one species, while another 

 species has a similarly constant predilection for another colour? 

 Why did P. ranfurhji, on Bounty Islands, and P. traversi, on 

 Macquarie Island, lose their white dorsal patches? Why did 

 P. carimculatus, in New Zealand, lose its crest? Can the loss of 

 a white patch, or a crest, as well as the acquisition of those 

 characters, be put down to sexual selection? 



The loss of the white bar and of the crest is also probably due 

 to reversion, and other specific characters may have the same 

 origin. New characters, however, whidi Avere not due to 

 reversion, occasionally arose. When this took place on 

 continental areas, we may, perhaps, attribute their preservation 

 to their usefulnass as recognition marks. But this will not help 

 us with those species found only on a single island, or in a single 

 locality far away from other species. The difference in the 



