On the Hybridising of Buchs. 261 



In the New Zealand trigen we find these same tracts still 

 more marked, and by carefully reading- the descriptions in the 

 earlier portion of this paper many other examples of varia- 

 tions occurring on poeciloineres will be found, which it is 

 unnecessary to recapitulate here. 



We trust, therefore, to have made it clear that the varia- 

 tions among- these hybrids, although at first sight they may 

 seem to appear sporadically all over the birds, do in reality 

 make their appearance along very definite lines and in definite 

 tracts and that these lines and tracts are those which occur 

 throughout the Mammalian and Avian kingdoms as 

 " poeciloineres." 



Now some of these patches, which have just been shown to 

 be variations following the lines of the poeciloineres, might 

 possibly be also considered (as in fact has been previously 

 suggested) as instances of reversion to the Teal and Pintail. 



And which view are we to take ? Are they to be considered 

 as "reversions " — doubtful term taken to imply an older form 

 and more ancestral type — or as "variations " showing resem- 

 blances to two well-known species, such resemblances owing 

 their origin to variation following the well-defined lines which 

 I have above and more fully on a former occasion shown to 

 hold constant and good for the whole of the higher vertebrates? 

 If we accept the hypothesis of reversion, in the concrete 

 instances noted above, we have then to answer the further 

 questions : (1) Are the Pintail, Teal, Gadwall and Wigeon, 

 to which resemblances are shown all of a more ancestral 

 type than the Mallard group, which is the most .widely spread 

 genus ? 



(2) If from dealing with hybrids of five species "rever- 

 sion " is obtained to four other genera, is it not likely that it 

 is only a matter of hybridising more species to get resem- 

 blances to more genera, and that instead of coming back to one 

 type as reversion would lead us to exjsect, we seem to be 

 " reverting " to many types ? 



And furthermore, if we hold to the reversion theory how- 

 are we to account for those varieties that resemble no known 

 species, which occur pari passu with the others ? 



If, however, these resemblances to other species are con- 

 sidered merely in the light of variations that are following 



