248 GEOEGE JOHN ROMANES 1888- 



Now if one mare among many has the slightest 

 degree of insterility with the second variety, she will 

 have a larger representation in A, and a less represen- 

 tation in that intermediate portion of species which 

 is ready to vanish away. 



Her unsuccessful attempts with B come into the 

 ordinary number of failures to fertilise which the 

 supply of ova provides for. Her nourishing and 

 bearing powers are spent on producing A. 



And this is so in whatever smallest degree she is 

 infertile with B, that is in whatever smallest degree 

 she tends not to bear intermediates. We want no 

 special segregating variation, only all possible varia- 

 tion in this respect as in others and the smallest 

 will be accumulated by ' new natural selection ' pre- 

 cisely as the smallest variations in hoof or hair are, 

 i.e. by less or more representation within the group. 



The mares who bear no intermediate or centrally 

 typed forms will be in each generation more and 

 more preponderatingly represented. Those which 

 are equally fertile in all directions will become rarer 

 and rarer till at last they are counted abnormal. 



That explains sterility of two adjacent species 

 inter se. 



But not the infertility of hybrids. 



I think that acts in a similar way. Two mares 

 bear first crosses, which are fertile either wiuh original 

 species on either side or with other first crosses. 



If one of these mares bears first crosses which 

 are less fertile, she is less represented, i.e. succeeded 

 in the central portion of the stock, and more in her 

 own part of it, by her other children, and so in the 



