DIVERGENT EVOLUTION THROUGH SEGREGATION. 293 
Artificial segregation is caused by the relations in which the organ- 
ism stands to the rational environment, that is, to the purposes of man. 
In other words, artifieial segregation is the rational form of environal 
segregation. Though the bearing of segregation on the evolution of 
species in a state of nature has been for the most part overlooked, its 
effects have been quite familiar to the breeder of domestic races. 
As a convenient method of illustration, let us consider the different 
results that will be gained according as we subject the same ten paus 
of wild rock pigeons to one or the other of the following methods of 
treatment: 
In the first experiment let the treatment be as follows: Let ten avia- 
ries be prepared, and in each aviary put one male with the female that 
most nearly resembles it. When the young of each aviary arrive at 
maturity, let them be inspected, and if any individual resembles the 
inmates of one of the other aviaries more than the inmates of the aviary 
in which it was produced, let it be placed with those it most closely 
resembles. If any unusual variation arises, let it be placed in a new 
aviary, and let the one of the other sex that most closely resembles it 
in that respect be placed with it. When the crowding in any aviary 
becomes injurious to the health ef the birds let the numbers be indis- 
criminately reduced. Let this process be continued many generations, 
the inmates in all the aviaries being fed on the same food, and in every 
respect treated alike, and what will be the result? 
No experienced breeder will hesitate in assuring us that under such 
treatment a multitude of varieties will be formed, many of which will 
be very widely divergent from the original wild stock. In other words, 
cumulative segregation will produce accumulated divergence, though there 
is no selection in the sense in which natural selection is selection. 
Again, let us take the same ten pairs, and putting them into one 
large aviary, let them breed freely together without any segregative 
influence coming in to affect the result, and who does not know that 
the type would remain essentially one, though a considerable range of 
individual variation might arise. That is, without segregation no diver- 
gence of type will arise. 
THE NATURAL LAW OF CUMULATIVE SEGREGATION. 
I shall now show that there is in nature a law of cumulative segrega- 
tion. There are large classes of activities in the organism and in the 
environment that conspire to produce segregate breeding; and to pro- 
duce it in such a way that in a vast multitude of cases it becomes a 
permanent fact, which no cause that we are acquainted with can ever 
obliterate. Moreover, when one form of segregation has become fully 
established, we find that the different branches are liable to be again 
subjected to segregative influences, by which each branch is subdi- 
vided, and in time differentiated into divergent forms that are not 
liable to inter-cross in a state of nature. 
