62 EXPERIMENTAL EVOLUTION LECT. 



influence, which determines no variation at all in the 

 maternal organism, has, however, effected a deep 

 change on the germ-cells, and has operated on an 

 important part of the organism. Instances of the 

 same sort may be met with among plants. It 

 happens, for instance, that after a branch of a varie- 

 gated variety has been grafted on a non-variegated 

 plant, some variegated branches sprout from the 

 latter. And this shows, as M. Armand Gautier con- 

 tends, and as all naturalists must admit, that race- 

 variation, or, generally speaking, variation of any 

 sort, or of any importance, is not a mere external 

 fact a mere external modification but that the 

 modification makes itself felt in the utmost depths 

 and intimacy of the cells. Otherwise stated, there 

 is not only a mere difference of form ; underlying 

 the formal or external difference, there are modifi- 

 cations of much greater importance in the chemistry 

 or physiology of the cell-plasma. These differ- 

 ences may be localized instead of being general. 

 For instance, the same orange-tree may bear oranges 

 and lemons, if some flowers have been impreg- 

 nated with pollen from lemon-trees ; and it may 

 happen even that the variation is more localized 

 still when one half of the fruit is orange and the 

 other lemon when, as Naudin has seen, one half is 

 Datura stramonium, the other D. lacvis. Of course, 



