1897] REPRODUCTIVE DIVERGENCE 407 
of the race into two distinct races more and more obvious. Again, 
I had supposed that the more widely any two individuals differed, 
the less on an average would be their mutual fertility. Hence this 
principle of reproductive divergence is a cumulative one, and the 
later stages of divergence will be very much more rapid than the 
initial ones. 
Again, Dr Jordan objects to my statement that as the race 
diverges, shorter and longer individuals than those originally present 
will gradually be evolved. This seems to me to be so obvious as to 
scarcely need demonstration. Thus, let us suppose for the moment 
that our 900 individuals of each sex are split up into 450 short 
individuals and 450 tall ones. The members of each of these 
groups will deviate in either direction from the average size in the 
same proportion as the members of the original single group did. 
For instance, in the 928 adult offspring obtained in the above- 
mentioned anthropometric data, the mid-stature or median of the 
whole group was about 68:2 inches, and 10 per cent. of the group 
were below 64°5 inches in height. Supposing now the median of 
the new group of short individuals be .65°8 inches, it follows that 
10 per cent. of this group will be below 62:2 inches in height. 
That is to say, individuals shorter than any of those originally 
present will have arisen. The tall individuals will, of course, 
deviate in a similar manner in the opposite direction. 
This fresh attempt to demonstrate the correctness of the prin- 
ciple of reproductive divergence will not, I am afraid, appear much 
more easy to understand than the former one, but it at least has 
the merit of being roughly founded on actual data, so that fewer 
preliminary assumptions are necessary, and the result obtained is a 
more absolute one, and is moreover independent of the Law of 
Regression towards mediocrity.t. To me it seems the principle is 
sufficiently obvious without the help of any mathematics at all, if 
it be looked at in the following manner. Let any number of 
individuals in a species be divided up into two groups—the larger 
ones and the smaller. Then if there be a correlation between size 
and fertility, it follows that those larger individuals which happen 
to breed with the smaller ones will give rise to fewer individuals of 
intermediate size than they would have done if there had been no 
such correlation. That is to say, the race will begin to diverge, 
and as this divergence is cumulative, it will ultimately split up into 
two or more new races, H. M. Vernon. 
1 &* Natural Inheritance,” p. 95. 
