NATURE AND CAUSE OF VARIATION. 307 



body in symbiotic relation, must be thought of as a unit or separate entitv. 

 That is, a germ cell is not germ plasm plus protoplasm and stored food, but 

 the germ cell, with its initial endowment of plasma, which it derived by 

 division from its ancestors, and which it always retains and hands on from 

 generation to generation in each species, is all germ plasm. Each germ cell 

 later, by the accumulation of stored food or the elaboration of other materials, 

 becomes differentiated, but the original chromatic and plasmic endowment is 

 singularly constant. In Lcptinotarsa each germ cell comes to the beginning 

 of its period of differentiation (growth period) with an almost exactly similar 

 endowment of dense homogeneous plasmic material. 



While we can not, I think, at the present time form any adequate picture 

 or frame any conception of the architecture of the germ plasm, we can not, 

 however, escape the conviction that the germ cell is an entity, and is the 

 product of a series of developmental stages which always recur in the same 

 order and are relatively invariable. It is able to repeat these stages in devel- 

 opment, to overcome serious difficulties, and in the end attain the same gen- 

 eral result, which differs only from its specific type in relatively minor points. 

 Beyond all doubt, this sequence of stages is conditioned in the germ plasm, 

 but not predetermined in the germinal material as material particles, each 

 the bearer of particular attributes of one or more cells or cell groups. Rather 

 we must think of the germ plasm as containing a few general characters 

 from which, in some wholly unknown way, there come the sequence of stages 

 in development of the diverse organs and the apparent isolation and unity of 

 characters which we find. We have good reason, therefore, for the belief 

 that development is an epigenetic process, that one state is conditioned 

 by preexisting stages, and that a variation is the result of responses in the 

 plastic developing organism ; that is, variation is to be interpreted upon the 

 basis of responses to stimuli directed by the stage of development reached 

 and the nature of the preexisting stages. J'ariation is also epigenetic and 

 not a predetermined character in organisms. 



Mutation. 



De Vries and his followers maintain that variations standing far apart 

 from the general population, and differing often in many characters, differ 

 fundamentally from the normal inheritable fluctuating variations. De Vries's 

 "mutants" bred true, but with intense selection and in-and-in breeding ; and 

 in my experiments with Leptinotarsa there have been found and reared, with- 

 out intense selection, examples of these same extreme variations. De Vries 

 maintains that the "mutants" of Oenothera are in all directions, which may be 

 true; but, after examining his work and seeing the various "mutants" grow- 

 ing, I should say that they were not in all directions, but in two chief direc- 

 tions, and that the "mutation" consists in plus and minus changes of the 



