THE CAUSE OF SPONTANEOUS VARIATIONS 99 



herds, e. g., the males and females are so like each other that 

 the offspring, were it possible, might almost as well be 

 raised by buds or cuttings." 1 If this be so, crossing such 

 types would cause a reversion to the ancestral condition of 

 greater variability. 2 



164. Varieties of the same species, when evolved by artificial 

 selection, are generally very fertile when crossed, no matter 

 how divergent they may be in structure. But varieties 

 evolved under Natural Selection are generally much less 

 fertile even when the divergence in structure is apparently 

 slight. 3 Distinct species tend to be inter-sterile; distinct 

 genera are nearly always so. Natural varieties are evolved 

 much more slowly than artificial varieties. It is evident, 

 therefore, that time rather than apparent divergence of 

 structure is the principal factor in the causation of specific and 

 varietal inter- sterility, a hypothesis which is confirmed by 

 the fact that species inhabiting distinct regions tend to be 

 sterile when crossed. 4 The explanation seems to lie in the fact 

 that, as regards domesticated varieties man has sought to 

 cause differences, not in the embryo, but in the adult forms. 

 His varieties develop from egg to foetus along parallel lines, 

 and on that account are inter-fertile. But, as we know, 

 variation, regressive and progressive, is at work on the embryo 

 as well as on the adult. It slowly brings about changes in the 

 embryonic development, so that in time the embryonic develop- 

 ment of two long-separated forms proceeds on lines more or 

 less divergent, or at any rate not parallel. When two domestic- 

 ated varieties mate we tend to get reversion ; that is, the last , 

 the divergent, links in the chain of development are dropped. 

 Presumably the same thing occurs when the divergence 

 occurs early in embryonic life, but then the links are dropped 

 in the beginning or middle of the chain and further develop- 

 ment becomes impossible. Hence, probably, the reason why 

 long-separated species are inter-sterile. Of course changes 

 may occur so early as in the germs. Thus the chromosomes 

 may not be of an equal number. When the germs are in- 

 compatible we get complete sterility. When the develop- 

 mental tendencies are incompatible the embryo perishes at 



1 The Penicuik Experiments, p. xlvi. 2 See 142-3. 



3 " With respect to sterility from the crossing of domestic races, I know 

 of no well-ascertained cases with animals. This fact, seeing the great 

 difference in structure between some breeds of pigeons, fowls, pigs, dogs, 

 etc., is extraordinarily in contrast with the sterility of many closely- 

 allied natural species when crossed." (Animals and Plants, vol. ii., 

 p. 81.) 



4 Origin of Species, p. 384. 



