V. 



The Problem of Variation. 



IT is certain that, in order to get a deeper insight into the causes and 

 modes of organic evolution, we must investigate the problem of 

 variation. The first requisite for such an investigation is a definite 

 grasp of the problem, an analysis of its complications. Much that is 

 written at the present time on ultimate biological theories is more or 

 less barren from the absence of a definite and comprehensive concep- 

 tion of the phenomena which have to be explained. It is easy to 

 expand the proposition that, given all kinds of hereditary variations 

 in individual organisms, selection will accumulate those which lie in a 

 particular direction and result in special adaptations of structure to 

 function. But what causes the variations ? It is more important and 

 more difficult to construct a theory of heredity as due to the continuity 

 and conservatism of the germ-plasm, but we have to explain, not 

 merely conservative, but progressive and retrogressive heredity. 



It is necessary always to bear in mind that the phenomena to be 

 dealt with are presented by individual organisms. There is no such 

 thing as the origin of species apart from the origin of individuals. It 

 has been maintained by philosophers that a horse can be locked in a 

 stable, but the species horse cannot, being a mental abstraction 

 formed from the perceptions of individuals ; but it seems to me 

 more correct to say that we recognise a number of individuals with 

 such a degree of similarity to one another that we class them together 

 and give them all a specific name. It would not be impossible to put 

 all existing horses together into one enclosure. 



Variations then, of course, must be variations of individuals; 

 but a most important distinction, which is constantly ignored, exists 

 between contemporary individual differences and the differences 

 between consecutive generations, which form, so to speak, the units 

 of evolution. The nature and degree of individual differences in a 

 single species, the relations of varieties and doubtful species, are 

 important enough as observed facts ; but Darwin, Wallace, and 

 others present them as the material on which selection acts, without 



juately discussing the question of the relation they bear to the 

 difference between one generation and its successor. Galton and 

 Weldon have recently applied the higher mathematics to the study 

 of individual differences, but it does not seem to me that in this way 



