DURATION OF RACES. 341 



As offspring tend to resemble grandparents almost as 

 much as parents, and as a line of close-bred ancestry 

 is generally prepotent, so newly-originated varieties 

 have always a tendency to reversion. This is pretty 

 sure to show itself in some of the progeny of the ear- 

 lier generations, and the breeder has to guard against 

 it by rigid selection. But the older the variety is 

 that is, the longer the series of generations in which 

 it has come true from seed the less the chance of re- 

 version : for now, to be like the immediate parents, is 

 also to be like a long line of ancestry; and so all the 

 influences concerned that is, both parental and an- 

 cestral heritability act in one and the same direction. 

 So, since the older a race is the more reason it has to 

 continue true, the presumption of the unlimited per- 

 manence of old races is very strong. 



Of course the race itself may give off new varie- 

 ties ; but that is no interference with the vitality of 

 the original stock. If some of the new varieties sup- 

 plant the old, that will not be because the unvaried 

 stock is worn out or decrepit with age, but because in 

 wild Nature the newer forms are better adapted to the 

 surroundings, or, under man's care, better adapted to 

 his wants or fancies. 



The second question, and one upon which the discus- 

 sion about the wearing out of varieties generally turns, 

 is, Will varieties propagated from buds, i. e., by divis- 

 ion, grafts, Iwlbs, tubers, and the like, necessarily dete- 

 riorate and die out f First, Do they die out as a matter 

 of fact ? Upon this, the testimony has all along been 

 conflicting. Andrew Knight was sure that they do, 

 and there could hardly be a more trustworthy witness. 



