372 EXPERIMENTAL STUDY OF INHERITANCE 



to fungoid attacks than those with dentate margins, so that in 

 the course of time in certain conditions the former race would 

 tend to be ehminated by natural selection ; but it is also handi- 

 capped by the hereditary conditions, since three dominants 

 are always being produced to one recessive. 



Swamping Effects of Intercrossing. — A well-known objection 

 to Darwinism, first clearly stated by Prof. Fleeming Jenkin, 

 is that variations of small amount and sparse occurrence would 

 tend to be swamped by intercrossing before they had time 

 to accumulate and gain stability. In artificial selection the 

 breeder takes measures to prevent this " swamping-out," by 

 deliberately pairing similar or suitable forms together, or by 

 deliberately removing undesirable forms ; but what, in nature, 

 corresponds to the breeder ? 



Various answers are possible: — (i) It may be that similar 

 variations occur in many individuals at once and many times 

 over. (2) It may be that the variations which really count 

 in evolution are not small individual fluctuations, but discon- 

 tinuous variations. (3) It may be that many variations are 

 not from the first unstable, but express changes of organic 

 equilibrium which come to stay if they get a chance at all. (4) 

 There are numerous conditions in nature — summed up in the 

 concept ''isolation" — e.g. geographical barriers, differences in 

 habit, psychical likes and dislikes — ^which tend to prevent free 

 intercrossing between sections of a species. Similar forms 

 may pair, and, in various ways, assortative mating may come 

 about naturally. And whenever inbreeding sets in prepotency 

 develops — i.e. pecuUarities, even if trivial, gain great staying- 

 power in inheritance. (5) But even more important are the 

 facts disclosed by Mendel and his school, that crossing does not 

 tend to swamp new features, for if the hybrids be inbred there is 

 a persistent segregation of the parental type. A new mutant 

 crossed with a related form of contrasted character may be 

 dominant or recessive in the immediate hybrid (F^), but in 



