WHERE DORMANT TRAITS ARE ABSENT 175 



type, but proof of blending is furnished by the fact that continued 

 infusion of white blood in succeeding generations results in a 

 gradual change of type till at last the hair loses all traces of 

 woolliness. In these later generations also the colour of the 

 eye may suddenly be changed to a lighter hue. This last fact 

 affords another crucial instance which demonstrates that Mendelian 

 inheritance implies, not segregation, but latency. The inheritance 

 of eye-colour is alternative in ordinary intra-varietal breeding. 

 But when a white race is crossed with a black the eye-colour of 

 the former becomes permanently latent, or, at least, as long as 

 the mongrels breed together. But the latency is weakened by 

 successive infusions of white blood, till at length the descendants 

 may have blue or hazel eyes. If there were segregation of units 

 accompanied by the destruction or the non-reproduction of those 

 for light colour, the latter could never reappear, no matter how 

 often the infusion of white blood were repeated. 



291. In addition to manifesting nearly complete general blend- 

 ing, human races when crossed differ from crossed varieties of 

 domestic plants and animals in another, an equally remarkable, 

 and a very important way. Never in human crossings is a latent 

 ancestral character brought to light, whereas the reappearance of such 

 characters, dormant perhaps for many generations, is one of the 

 most conspicuous results of the crossing of domestic varieties. 

 Now what is the significance of these very remarkable differences 

 of inheritance ? How is it that in the one case blended inheritance 

 is the rule, whilst in the other alternative reproduction combined 

 with a frequent development of hitherto latent characters is 

 general ? If we are able to unravel this mystery, I think we shall be 

 able to solve those problems of heredity which, of recent years, have 

 especially occupied the attention of experimental workers, and indeed 

 of most biologists. 



292. The majority of experimental observers claim that the 

 inheritance of mutations is alternative. If we substitute 'develop- 

 ment ' for ' inheritance,' the evidence that they are right seems 

 clear, at least as regards many mutations. In man occur polydacty- 

 lism, hypophylangia, alkaptonuria (a dark-coloured condition of 



I the urine), haemophilia, colour-blindness, deaf-mutism, and other 

 large variations, 1 all of which may be classed as mutations, 

 and the development of which is alternative. Again, the majority 

 of experimental observers declare that a mutation is properly 

 defined as a large discontinuous variation, one which is not 

 1 See 769-71- 



