180 SELECTION. 



stitution. We can calculate the chance which any individual of 

 this group has, to be amongst those who are parents of the next 

 generation. Other things being equal, this chance depends 

 upon the rate of reproduction of the group, upon the stabiUty 

 of conditions favouring development of organisms of this kind, 

 and greatly also upon chance combinations of circumstances 

 around every young individual. If for some reason, the large 

 individuals have some additional advantage over the smaller 

 ones, the chance for a large individual to be amongst the par- 

 ents of the next generation will be accordingly greater. If 

 originally both large and small individuals happened to exist 

 in approximately equal proportions, it is evident that the bet- 

 ter type, the larger type would tend to become the specific 

 tj^e, the type of the species, if the variability of the whole 

 group tended to reduce itself for any reason or combination 

 of reasons. 



But if the large individuals are far in the minority, this low 

 proportion of their ntunber to the total number has to be 

 reckoned with. The chance for a large individual to be one of 

 the pro-creating individuals depends for a good deal upon the 

 proportion of large ones to small ones, and if this chance is 

 materially increased by a greater fitness, we are still concerned 

 with the modified, low proportion. If now we consider artific- 

 ial selection, we must see that if a breeder considers great size 

 of some advantage, he will isolate a group of large individuals. 

 Some of these may be impure, of course, and as the choice may 

 happen to be wholly on phenotypic merit, many individuals 

 which are potentially small ones may be included, but it is 

 evident that by this isolation of a number of large individuals, 

 the chance of the eventual tjTpe of the selected group to be 

 large is very greatly enhanced, out of all proportion of what 

 can happen along this line by natural selection without 

 isolation. 



And this consideration of the fact that selection under domes- 

 tication is effective through the fact of being isolation as 

 much as selection shows us, where and when we have to look 



