APPENDIX B 363 



the accumulation of small fluctuations, and that stability results 

 from the fixing of originally unstable variations by continued 

 selection, and the mutation theory which supposes that evolution 

 depends on mutations which are stable from the beginning. 

 Which of these theories is true ? Have the selectionists failed 

 to note that the variations on which evolution is founded are 

 absolutely stable from the moment of their first appearance, or 

 have the mutationists, in their general neglect of the doctrine of 

 recapitulation, failed to note that the variations on which evolu- 

 tion is founded are invariably fluctuating at first, but become 

 more and more stable under the influence of selection? 



We may dismiss at once, as obviously in conflict with notorious 

 and indisputable facts, the hypothesis that mutations are stable, 

 and therefore are the materials of evolution, because their transmis- 

 sion is Mendelian. The human race offers peculiarly favourable 

 opportunities for the study of heredity opportunities that have 

 been woefully neglected. Of necessity, we know it better than 

 we can possibly know any other species. We constantly note 

 the occurrence of variations amongst our relations and friends, and 

 we cannot possibly miss the occurrence of mutations, at any rate 

 considerable mutations among them. Owing to the cessation 

 of stringent selection as regards many particulars, Man is 

 probably the most variable of all natural species. 1 The 

 varieties of mankind scattered as they are over the whole 

 habitable world, have diverged so greatly from one another that 

 they differ, not in one or a few characters, but practically in all 

 observable characters. Great migrations to distant lands, and 

 therefore to novel conditions of life, have occurred. All human 

 races appear to be inter-fertile, and hardly a race exists which 

 has not crossed occasionally with members of almost all other 

 races. In some instances this inter-crossing has been so extensive 

 that over large stretches of country, as in South America, the 

 half-breeds outnumber the pure varieties. In other instances 

 several varieties have united to form races of compound hybrids. 

 Now, how often do we observe a mutation the transmission of 

 which is Mendelian 1 In pure or intra- varietal breeding only 

 in very rare cases of useless or injurious mutations albinism, 

 polydactylism, and the like. The most striking evidence against 

 the Mendelian-mutation hypothesis is afforded, however, by 

 crossed varieties. If human evolution has been on lines of Men- 

 delian mutations, then, of course, crossed human varieties should 

 display Mendelian inheritance. Not a solitary example has been 

 recorded. The most diverse races e. g. Anglo-Saxon and Negro 

 blend perfectly, and the blend is transmitted unfailingly to the 

 latest descendants. The Negro, it is true, transmits his eye colour 

 to the vast majority of his descendants, but this is prepotency in 

 the Darwinian, not dominance in the Mendelian sense. Pre- 



1 See 141. 



