196 APPENDIX II—INTENSIVE SEGREGATION. 
race is usually the result. In other words, balanced selection produces 
stability of type,and unbalanced selection produces transformation of type.* 
In the light of this twofold law we see how there may be stringent 
selection without transforming effect. In nearly every species there is 
a constant struggle between the different forms of variation; and as it 
never happens that all the forms are equally successful, the process of 
natural selection is always bearing in full force upon the species. If, 
then, it could be shown that natural selection, wherever it exists, must 
necessarily produce transformation, it would be impossible to resist 
the conclusion that nearly every species is undergoing transformation 
through this cause. But it is unbalanced rather than balanced selec- 
tion that produces transformation. We also see that heredity tends 
to make the most successful form the average form, and thus to con- 
vert unbalanced into balanced selection. From this it follows that in 
order that selection should produce continuous transformation, cov- 
ering a wide range, it is necessary that the form of variation selected 
should from time to time be changed. This may be expressed as the 
principle of continuous transformation through successive changes in the 
character of the selection. 
Though selection produces transformation only when it involves 
the survival of other than typical forms, it is still very possible that 
there are only a few species in which completely balanced selection 
prevails for very many generations in succession. It is still certain 
that long-continued independent selection gradually passes into diver- 
sity of selection producing divergent evolution. 
(3) Though in more than one passage Darwin maintains that uni- 
formity of external conditions involves uniformity of natural selec- 
tion, and that isolation can have no effect in transforming a species if 
physical conditions and surrounding organisms remain the same, 
still I think that if the question had been distinctly brought before 
him he would have admitted that exposure to a new or changed 
environment was not a necessary condition for change in the char- 
acter of sexual selection. Now, I think it can be shown that, besides ~ 
sexual selection, there are several forms of selection that depend upon 
the relations of the members of one species to each other and that 
may undergo change without the organism being exposed toa different 
environment. 
Selection depending on the relations of the organism to the envi- 
ronment I call environal selection, of which I find two kinds, namely, 
* The general law is here stated, without any attempt to explain why selection, 
securing the exclusion of those falling below a certain standard, is necessary for 
the preservation of that standard in the case of many characters (see par. (6) below). 
