ACCUMULATION OF POTENTIAL SEGREGATION. 169 
shall have prepotency of each group within the circle of its own group. 
This process may take place when a group is protected by complete 
isolation, however produced. Let us next consider a case in which a 
‘small group partially protected from mixture with the original type 
by incomplete local and industrial segregation produces a variation 
whose ovules are more readily fertilized by pollen from the same 
group than by pollen from the original type. Is it not evident that 
this variation will gain with each generation an increasing prominence 
in the new group that maintains somewhat new methods of dealing 
with the environment in its partially isolated habitat? This will be 
so, first, because variations possessing but little or no prepotency with 
their own group will eventually coalesce with the original stock, and 
especially will this be the case if the new group becomes somewhat 
numerous and passes beyond the limits of its narrow habitat into 
districts where the original type abounds; and, second, because vari- 
ations possessing the prepotency with their own group in a superior 
degree will remain distinct, breeding with each other, and their de- 
scendants will become still more segregate and still more perma- 
nently divergent. Of the law of accumulation of segregative endow- 
ments, we may say that as the descendants of the best fitted necessarily 
generate with each other and produce those stull better fitted, so the de- 
scendants of those possessing the most segregative endowments necessarily 
generate with each other and produce those that are still more segregate. 
It will, however, soon be shown that unless the reproduction and 
power of survival is greater for the pure segregate forms than for 
the mixed forms, the proportion of pure forms to mixed forms will 
decrease in each generation. 
It is evident that when either segregate potency or segregate pre- 
potency is associated with the free distribution of the fertilizing ele- 
ment by wind or water, the combined effect must be in the former case 
complete, and in the latter case partial, positive segregation, for the 
breeding together of compatible forms is thereby secured. 
It may at first appear that a slight degree of segregate prepotence 
will prevent crossing as effectually as a higher degree, but further 
reflection will show that the efficiency of the prevention will vary in 
direct proportion with the length of time over which the prepotent 
pollen is able to show its prepotence, and this will allow of innumer- 
able grades. If, in the case of certain individuals, the prepotency is 
measured by about twenty minutes, while with other individuals it 
enables the pollen of the same variety to prevail though reaching the 
stigma an hour after the pollen of another variety has been applied, 
the difference in the degree of segregation will be sufficient to make 
the persistence of the latter much more probable than that of the 
