210 SUMMARY 



this occurs the child inherits the characters of both parents but 

 develops those of only one. Blending then occurs between the 

 patent traits of one parent and the latent traits of the other. 

 Therefore, speaking practically, 1 all inheritance is blended in bi- 

 parental reproduction. Unit segregation, gametic purity, and 

 independent inheritance of characters (in the Mendelian sense) are 

 all myths that have been founded on experiment but have not 

 been tested by it or in any other way. They are mere guesses. 

 The evidence from which they have been inferred is very restricted 

 and fragmentary, and, especially when considered in connection 

 with other larger bodies of fact, really points to quite other 

 conclusions. 



346. In sexually dimorphic species every individual possesses 

 one set of non-sexual characters, all of which are patent, and two 

 sets of sexual characters of which only one set is patent. But 

 if parents differ greatly in any non-sexual character, the reproduc- 

 tion of it tends to follow the sexual mode. This rarely happens 

 in the normal intra-varietal breeding of natural varieties, for under 

 such conditions individuals who differ greatly from the norm 

 seldom survive and have offspring, but it is quite a common 

 occurrence under artificial selection, which is distinguished from 

 natural selection by the choosing of mutations rather than of 

 fluctuations, the crowding together of newly formed varieties in 

 the same locality, and their constant intercrossing through design 

 or accident. All the evidence indicates that in nature varieties 

 arise solely or almost solely under conditions of geographical 

 separation. Artificial Selection and artificial varieties there- 

 fore differ profoundly from Natural Selection and natural 

 varieties. In artificial varieties are found dormant traits which 

 are practically unknown in natural varieties. Judging from the 

 analogy of patent characters, it is an error to suppose that dormant 

 traits persist for ever. Having no utility, they are not maintained 

 by Natural Selection, and therefore tend to disappear in time. 

 Thus when the horse is crossed with the Burchell zebra the repro- 

 duction of the long dormant ancestral stripes is very faint as 

 compared to the vivid reproductions of ancestral coloration which 

 may appear in offspring when two dull-plumaged breeds of poultry 

 or two white varieties of garden flowers are crossed. 



347. The progress of science depends first on the collection of 

 verified facts, and second on their systematic arrangement. The 

 facts must be verifiable or we cannot know them to be true. They 



1 See 322. 



