Vu ! 9 o 1 - 11 '] HUTTON, The Cormorants of Neiv Zealand. 3 



rostris is descended from P. mclanoleucus , individuals of this 

 latter species must have come to New Zealand from Australia 

 or New Caledonia, and have given rise to P. brevirostris in New 

 Zealand. The latter seems to be the more probable, because 

 P. melanoleucus is also related to P. pygmaius of the Mediterranean 

 and Central Asia. So that geographical distribution would lead 

 us to suppose that P. brevirostris was derived from P. melano- 

 leucus and the latter from P. pygmceus. There is confirmatory 

 evidence of this in the fact that P. brevirostris is a very variable 

 species, apparently not yet capable of breeding truly, while 

 P. melanoleucus is not variable, but an old and well-established 

 species. 



P. pygmceus is black, with scattered white plumes on the head, 

 back, and abdomen. P. melanoleucus is black above and white 

 below, with a white frill on each side of the neck. P. brevirostris 

 is black, with a white chin and throat. So that if the evidence 

 from geographical distribution and from variability can be 

 trusted, the descendants of P. pygmceus, as they travelled south- 

 wards, acquired a pure white breast and abdomen. But when 

 some individuals reached New Zealand a new change took place, 

 and the breast and abdomen became black. Why was this ? 

 Can any explanation be found ? 



Some ornithologists would explain the variations in P. brevi- 

 rostris as cases of partial albinoism, or of intercrossing with 

 P. mclanoleucus. But if the variations were due to albinoism 

 we should expect the white to appear on the upper as well as 

 on the lower surface, and to be unsymmetrical ; while P. melano- 

 leucus is not sufficiently common in New Zealand to allow us 

 to suppose that the abnormal birds are hybrids. I think, there- 

 fore, that we find a simpler explanation of these variations in 

 occasional reversion to ancestral characters. 



Evidently the origin of the variation in P. brevirostris is not 

 to be found in the surrounding conditions. For if this were the 

 case other Cormorants with white breasts and the same habits, 

 like P. varius, would have been similarly affected. It seems 

 more probable that the variation was due to a partial reversion 

 to P. pygma?us, and that the reversion is not yet complete, so 

 that the plumage of the immediate ancestor — P. melanoleucus — 

 often appears. 



Then, supposing the variation to have arisen by reversion, 

 how has it been preserved ? Natural selection is out of the 

 question, for the character is not a useful one. Cormorants 

 have no enemies in New Zealand, and a black abdomen cannot 

 be better than a white one when fishing, for so many Cormorants 

 are white on the lower surface. Again, the idea of sexual selection 

 does not satisfy me. Supposing a. dark variety arose and two 



