258 APPENDIX III—LETTERS PUBLISHED IN NATURE. 
papers on “‘Divergent Evolution through Cumulative Segregation”’ 
and ‘‘Intensive Segregation’’ I have endeavored to show that there 
must be several principles somewhat similar to sexual selection, which 
I have grouped with it under the names reflexive segregation and 
reflexive selection. In the former of these papers, pages 212-214, I 
have pointed out that of freely crossing forms of any species it is 
only those that are most successful that are perpetuated; while of 
forms that have by isolation escaped from competition with the 
original stock and are not crossing with it every variation is perpet- 
uated that is not fatally deficient in its adaptations to the environ- 
ment; and this will be the case whether the forms are held apart by 
reflexive or environal segregation. 
2. A Difference in Use that is not a Useful Difference. 
Let us consider the case of two allied species occupying the same 
area, and differing from each other in what Dr. Wallace has so appro- 
priately called their recognition marks, and in the segregating sexual 
and social instincts correlated with these marks. If investigation 
justifies the belief that an early stage of divergence, due, perhaps, to 
local segregation, resulted not only in sexual and social segregation, 
but also in what I have called divergent social selection (or what Dr. 
Wallace prefers to call selective association), then we are warranted 
in the belief that this segregative and selective principle was sufficient 
to perpetuate and intensify the new character, although the section of 
the species possessing the new character had not migrated into any 
new environment, and had not been exposed to any change in the old 
environment, and although it had not gained any new adaptation to 
the common environment of the two sections and, therefore, while 
both sections of the species were equally subject to identical forms of 
natural selection. 
Now, seeing that the individuals of the segregated sections are able 
to find and keep company with associates, and in the season to pair 
with suitable mates, as effectually, but no more effectually, than be- 
fore they were segregated, what shall we say of the usefulness of the 
distinctive characters that produce the segregation? It is plain that 
these divergent characters are in constant use; but does that prove 
that the divergence is a useful divergence? Is at not possible that there 
should be a difference in use which 1s not a useful difference? And if 
nothing has been gained by the difference either in maintaining the con- 
ditions of individual life, or in propagating the species, how can we call 
it a useful difference? And how can we attribute the divergence to 
natural selection, seeing that natural selection is the superior mainte- 
