TJie Evidence from Hybrids. 123 



has precisely the same character, I think we cannot re- 

 gard it as due to unnatural disturbance. 



According to our theory, variation is due to the action 

 of changed or unnatural conditions upon certain cells of 

 a preceding generation. Now, as characteristics of both 

 parents are mingled in a hybrid, it must nearly always 

 happen that certain cells with peculiarities of one parent 

 will be in contact with, or will depend in some way upon, 

 cells with peculiarities inherited from the other species. 

 There will therefore be a lack of the perfect adjustment 

 between each cell and its neighbors, which has been 

 brought about in each parent by natural selection, and 

 this imperfect adjustment will cause the cell which is 

 unfavorably placed to throw off gemmules. The cells of 

 the body of a hybrid will therefore be unusually prolific 

 of gemmules, and will transmit variability to later gen- 

 erations. 



According to our hypothesis, a hybrid is more likely 

 to transmit variability than a pure species, because more 

 of its cells are placed under circumstances favorable to 

 the production of gemmules. 



For the same reason a hybrid between two domesti- 

 cated or cultivated forms must have more tendency to 

 vary than one produced by crossing two wild species, for 

 the domestic or cultivated parents live under unnatural 

 conditions, and therefore have more tendency than wild 

 species to transmit gemmules, and thus cause variabil- 

 ity. 



The Sex of the Parent affects the Variability of Hy- 

 brids. 



I have shown that the body of a hybrid is peculiarly 

 favorable for the production of gemmules, and that, 

 for this reason, the descendants of hvbrids are variable 



