98 INHERITANCE IN POULTRY. 



THE MrTATION THEORY IN ITS RELATION TO THE ORIGIN OF DOMESTICATED ANIMALS. 



While the mutation theorj' of de Vries has received widespread adherence 

 among botanists, many students of animals, and especially of domesticated 

 races, have appeared as its opponents. Foremost among these are Professors 

 Keller, of Zurich (1905), and Plate (1905), of Berlin. I think that the essence 

 of the mutation theory is too little apprehended. It rests on the funda- 

 mental theory of heritable unit characters and assumes their very limited 

 mutability. It recognizes the important results wrought by artificial selec- 

 tion, but considers them as arising from two processes — first, the selection of 

 minute favorable variations of the fluctuating sort, and, secondlj^ the preser- 

 vation of new unit characters suddenly appearing. Such unit characters can 

 usually be not only maintained but much improved by subsequent selective 

 breeding. 



Now, it is true that breeders nowadays do not regularly wait for favorable 

 qualities to crop out. The process is too slow, uncertain, and expensive. If 

 one had scores of thousands of individuals, desired mutations would come 

 more frequently ; but even then they would rarely be of a desirable sort. 

 Every breeder can, on the other hand, improve any characteristic by selec- 

 tion, and that is for the most part the only method of improving a quality 

 that is open to him. Of course he can make new combinations of qualities 

 by crossing, but this does not, typically, result in new qualities. 



The question of the permanence of the improvement wrought by selection 

 of minute variations is the first point of difference between de Vries and 

 Keller. De Vries asserts that such improvement persists only so long as 

 selection is maintained. Keller adduces some interesting cases on the other 

 side, and the cogency of some of his evidence must be admitted. He traces 

 the gradual evolution in Egypt of long lop-eared hounds from straight-eared 

 ones. Ear length in rabbits, as Castle (1905, pp. 125-126) has shown, is not a 

 unit character ; at least, it blends in hybridization and con.sequently exhibits 

 any desired intermediate condition. The same is probably due for dogs ; 

 consequently this character may well have arisen by summation of minute 

 variations. Yet Keller goes on to show the long-eared condition has per- 

 sisted in central Africa, where selective breeding no longer occurs. Hence 

 one characteristic originated by selection of fluctuations has not retrogressed 

 on removal of selection. 



The preceding method of proof is not, however, of general validity. 

 Evidence that a characteristic arisen in domestication does not disappear 

 when the race becomes feral again is not evidence against the permanence 

 of fluctuations unless it is also proven that the characteristic arose by selec- 

 tion of fluctuations. This is usually not the case. The instance of long 

 ears would seem to be peculiar. Some of the other examples offered by 

 Keller of persistence of characteristics despite discontinuance of selection 

 avail little, since the precise origin of the unit characters concerned is un- 



