BI- PARENTAL REPRODUCTION 85 



by experience. Thus, though we are able to detect vast 

 differences between the people of our own race, Chinamen 

 seem much alike to us. The ordinary man hardly knows 

 one sheep from another ; the shepherd knows every member 

 of his flock. A famous old Dutch gardener was said to have 

 been able to distinguish all the hundreds of varieties of 

 hyacinth by their dried bulbs. 1 Moreover the germ-cells of 

 the same individual vary among themselves, and thus may 

 give rise to regression in this or that character in the off- 

 spring of individuals who apparently are not unlike. 



146. Bi-parental reproduction occurs invariably in all large 

 and complex organisms ; that is, in all organisms that are apt 

 to vary uselessly in a great number of ways. 2 It occurs 

 periodically at any rate in most parthenogenetic species ; and 

 is periodically prevalent also in many if not all unicellular 

 forms. It is an exceedingly effective device to assist and direct the 

 regression which is the necessary result of development In/ recapit- 

 ulation ; for, unlike the latter, it is selective in action. It elimi- 

 nates as a rule only useless variations. It plays exactly the 

 same part in selecting regressive variations that Natural 

 Selection plays in selecting progressive variations. Bi- 

 parental reproduction, indeed, is only another name for 

 bi-parental selection. It is almost as important a factor in 

 evolution as Natural Selection. 3 



1 Animals and Plants, vol. 1, p. 384. 



2 Animals are more complex than plants, and, therefore, have a 

 greater need for regressive variations. Though the germs of the same 

 individuals differ among themselves, the germs of separate individuals 

 differ more. Consequently plants are sometimes self-fertilized, animals 

 are not. " As yet I have not found a single terrestrial animal which 

 fertilizes itself. This remarkable fact which offers so strong a contrast 

 with terrestrial plants . . ." (The Origin of Species, p. 123.) Self-fer- 

 tilization occurs, however, among certain more or less vegetative marine 

 animals. 



3 It must be noted, however, that bi-parental reproduction itself must 

 have arisen through the survival of those organisms that were capable 

 of this mode of reproduction. Ultimately, therefore, all the phenomena 

 of evolution must be referred to Natural Selection. Even reversion 

 due to failure to recapitulate the life-history must be so referred, for 

 reversion depends on antecedent progression, and there can be no pro- 

 gression in the absence of Natural Selection. Since bi-parental repro- 

 duction planes away the vast multitude of useless variations, it is evident 

 that it must contribute, on the whole, to the stability of the species. 

 Weismann is mistaken, therefore, in supposing that parthenogenetic 

 species are particularly stable. No doubt bi-parental reproduction con- 

 tributes to the instability of species by aiding in the elimination of 

 organs which have become useless, but its major function is the elimina- 

 tion of those useless variations which arise in one generation and are 

 lost in the next. 



