134 REDUCTION OF VARIABILITY. 



Let US for the moment assmne that the two species of Lychnis 

 differ in 64 genes, some present in diuma and not in vespertina, 

 and some common to the vespertina plants only. If a hyjsrid is 

 produced, it either crosses into the vespertina or into the diuma 

 papulation. If there is nothing especially valuable about the 

 characters of the hybrid, in other words if its chances for sur- 

 viving are the same as for those of the individuals of the spe- 

 cies into which the hybrid merges, in six generations the last 

 traces of the hybridization will have disappeared. If any quali- 

 ty, resulting from a chance combination of genes given in the 

 potential variabiUty of the hybrid and its descendants confers 

 an advantage, it may take a little longer. But only if a new char- 

 acter, or at least a new genotype makes a number of plants 

 grow where they are not crossed back into the mother-species, is 

 there any chance of perpetuation. Isolation of some sort is nec- 

 essary, without isolation even selection can not work against 

 the nivellating effect of the factors tending to reduce the po- 

 tential variability. In the chapter on selection we will further 

 discuss the possibility of change in species under selection with- 

 out isolation. 



The very fact that Lychnis vespertina and diurna pltmts are 

 infinitely more nmnerous than plants of hybrid origin tends to 

 keep the two species pure and stable. Each of them has its 

 own potential viariability, which is normally very small, which 

 may be temporarily heightened by crossing, but always again 

 automatically reduces itself. In a region where the two species 

 exist at the same time, and where plants with a new genotype 

 must occasionally be produced, there is no room for a third spe- 

 cies to establish itself, simply because of the fact that no group 

 of plants can effectively become isolated from the "swamping" 

 multitude of individuals of the two species. 



If a new form is prevented by any circumstance from inter- 

 breeding with the parent-species, its chances to persist are very 

 much greater. And it is evident that these chances are better in 

 those casses, where the new group fills a niche different from 

 that of the parent-species. For in that case it does not need to be 



