OTHER THEORIES OF SPECIES-FORMING. 3 2 9 



with the parent type of the variable organ or individual. 

 Darwinians have always been interested in such variations, 

 for if they do occur in any considerable numbers they might 

 offer a possible solution of that difficulty in the selection 

 theory of explaining the origin of new structures and the 

 needed degree of size and development sufficient to make 

 these beginnings useful and hence available as handles for 

 natural selection. But it has long been recognised that such 

 sports or discontinuous variations are too few and occur too 

 rarely to furnish the basis for a comprehensive theory of 

 species-forming. Like the extremes of individuals in the 

 series of fluctuating variations, their characters would be 

 lost or swamped by crossing. Darwin himself made as full 

 a list of such sports as any post-Darwinian writer, ex- 

 cepting Bateson, has been able to do, and he recognised 

 the fact that certain species, or races at least, of domesticated 

 animals and cultivated plants undoubtedly had their begin- 

 nings in these sports. His examples of such discontinuous 

 or saltatory evolution as the Ancon and Mauchamp sheep, 

 the Paraguay cattle, 4 etc., are the classic examples in ani- 

 mal evolution, and to this day nearly the only ones ! Bate- 

 son 5 has, to be sure, gathered together in his "Materials, 

 for the Study of Variation" a much larger list of sports or 

 discontinuous variations than Darwin included in his knowl- 

 edge (it should be borne in mind in referring to Bateson's 

 list that several, probably, indeed, many of his alleged ex- 

 amples are cases of teratogenic regeneration) but he has 

 been able to add almost no new examples of the origin of 

 a new species from such discontinuous variations. A few 

 cases are known of the inheritance through a number of 

 generations of suddenly appearing sports or discontinuous 

 variations in human beings (cases of polydactyly, etc.) and 

 cats. (Kennel's stump-tailed cat, which produced in six 

 litters four stump-tailed, twelve tailless, and twelve normal 

 young, is an example of several similar cases which have 



