256 J. L. Bonhote: 



V. COLOUE. 



It lias seemed best, for reasons which will shortly be 

 apparent, to deal with colour under a separate heading, 

 although it has usually been treated as the most obvious and 

 easily distinguishable character by which reversion might 

 be recognised. 



The first point to notice is that among the Pintail-Spotbill- 

 Mallard trigens, as well as among the New Zealaiid-Spotbill- 

 Mallards — the only two trigens of which a sufficient number 

 have been bred — we find two and only two forms occurring. 

 As already stated, under the description of these forms, we 

 have in the Pintail trigens the light and dark varieties, and 

 in the New Zealand trigens the form in which the Mallard is 

 dominant, and the form in which the Spotbill is dominant, the 

 New Zealand characters, although present, being swanked by 

 those of the other species. Furthermore these dimorphic forms 

 resemble typical wild dimorphic species {e.g., the Arctic Skua), 

 from the fact that, although there is never any doubt as to 

 which form any particular individual belongs, yet it is rare 

 to find two individuals that are precisely similar. 



One of the most striking results of hybridisation, especially 

 when carried beyond the first generation, is that it seems to 

 produce a flood of variation, a fact that was, I believe, first 

 noted by Pallas, and when we come to examine these results 

 critically we find that such hybrids may show either : 



(i.) Resemblances to one or other of their parents. 



(ii.) New variations 



(a) Resembling species other than their parents, 

 or (6) Resembling no known species. 



(iii.) White coloration. 

 At the present time it is part of our creed that u species " 

 represent varieties Avhich have become fixed and perfected by 

 natural selection, and that those species existing to-day are 

 of the type most suitable to their particular surroundings and 

 needSj unsuitable varieties having been immediately elimi- 

 nated by the relentless action of natural selection. 



But although natural selection has so far overcome varia- 

 tion as to cause species to breed to all intents and purposes 



