328 DARWINISM TO-DAY. 



nection, to observations and experiments carried on nearly 

 simultaneously with those of de Vries by Correns 2 and 

 Tschermak. 3 As a matter of fact the data on which de 

 Vries bases his theory of species-forming by heterogenesis 

 are, in part, nearly identical with those obtained by Mendel, 

 Tschermak, and Correns, which, however, is concerned 

 primarily with explaining the "laws" of heredity. 



The meaning of heterogenesis in connection with species- 

 forming and descent is essentially this : whereas by the 

 Darwinian theory species are transformed slowly and by 

 slight changes in at first one or two or a few and only 

 later in more parts, and all new species are derived from the 

 old ones (which usually disappear as the new ones appear) 

 by the gradual selection of the advantageous ones among the 

 regular slight, fluctuating, individual variations (known 

 commonly as Darwinian variations and which mostly occur 

 according to the law of error), by the theory of hetero- 

 genesis new species appear suddenly, not by a selective 

 choosing among the slight fluctuating Darwinian variations, 

 but independently of selection, and largely independently of 

 the so-called Darwinian variations, by the appearance in 

 fixed definitive form of several to many slight to consider- 

 able variations, which give the new species definite char- 

 acteristics differentiating it often in many particulars from 

 the old species, which differentiating characteristics are fully 

 and faithfully transmitted to the succeeding generations of 

 individuals derived from this suddenly born new species. 



In some theories or views of heterogenesis the suddenly 



appearing new variations and none of these theories gives 



a satisfactory explanation of the cause of these 



Discontinuous suc jden variations which give rise to new 

 variation, 



species, are those large, sometimes monstrous, 

 variations known as "sports," "monsters," etc. ; or, in Bate- 

 son's words, are "discontinuous variations," i. e., considerable 

 variations not connected by a continuous series of gradations. 



