366 APPENDIX B 



explanation lies in the fact that their variations are not swamped 

 by bi-parental reproduction. The swamping, or rather averaging 

 effect of sexual reproduction is naturally greatest in the case of 

 those animals whose superior powers of locomotion enable them 

 to commingle most. It is least in the case of plants that are self- 

 fertilized, and especially in the case of those which reproduce 

 parthenogenetically. Species of animals, of which the powers of 

 locomotion are limited, possess almost as many varieties as 

 plants. 1 



We are told that selection (i. e. the selection of small fluctuat- 

 ing variations) cannot create. But if it cannot create it can 

 alter, and so in the course of ages produce what is in effect a 

 new creation. The essence of the selection theory is that nothing 

 that is useful, and therefore permanent, is ever " created," but 

 that all such characters arise by the gradual alteration of pre- 

 existing structures. In this way the human hand and brain, to 

 take extreme example, have arisen from ancestral structures 

 which were immensely different, but which in every stage were 

 useful to their possessors. Thus also have arisen the scales of 

 fishes and reptiles, the feathers of birds and the hairs of 

 mammals. 



The mutation theory of evolution is quite impossible. If, 

 instead of groping in the comparative obscurity which surrounds 

 wild nature, or amid the confusion which has resulted from 

 the unrecorded crosses and the generally abnormal conditions 

 under which domesticated species exist, men had turned their 

 attention to the clear, voluminous, decisive evidence, much of it 

 statistical, afforded by their own species, it would never have 

 been propounded. 



Bi-parental inheritance, as we have seen, may be blended, par- 

 ticulate, or exclusive. Since characters may be inherited inde- 

 pendently of one another, particulate inheritance may be regarded 

 as a variety of the exclusive type. Exclusive inheritance may be 

 alternative, as in typical Mendelian cases, or non-alternative, as 

 when one parent is prepotent over the other in the Darwinian 

 sense. 



Since many species reproduce parthenogenetically, the union 

 of germ-cells derived from different individuals is clearly not a 

 necessary condition of reproduction. Nevertheless this method 

 of reproduction, notwithstanding its obvious disadvantages, is so 

 nearly universal in nature, that it must possess some function of 

 immense importance. Weismann has suggested that this func- 

 tion is the causation of variations, both progressive and regres- 

 sive. But variations have been proved to occur in abundance in 

 parthenogenesis, and, as we see, parthenogenetic and self -fertilized 

 species are particularly rich in varieties. It has been proved, 



1 For example, snails. See Romanes, Darwin and after Darwin, vol. 

 iii., p. 16. 



