Method of Origin of New Species and Structures 423 



First, exact studies on variation seem to show that the fortu- 

 itous variations of Darwin are of two distinct Idnds or classes. 

 One class, now called fluctuating variations, includes those caused 

 by the immediate environment acting either forcibl}^ (producing 

 injury, &c.), or through stimuli calling out irritable responses; 

 they are not hereditary, and therefore have no influence in evolu- 

 tion. They are variations of the body plasm only, in Weismann's 

 sense. The reader will find good examples of variations of this 

 type in the differences between the individuals within each of the 

 mutation groups shown in figure 173. The other class includes 

 those that are inborn, and hereditarily transmitted from genera- 

 tion to generation, — variations of the germ plasm, in Weismann's 

 sense. These are variations upon which natural selection works. 

 Their origin is unknown, but they are related if not identical with 

 mutations, and with permutations and combinations of Mendelian 

 unit characters. The two classes of coinse, are indistinguishable 

 by the eye, and only determinable by experiment. 



Second, the very newest studies, announced during the writing 

 of this book, appear to be demonstrating that the mutants of 

 Evening Primrose discovered by de Vries, are simply in large part 

 the separation or segregation out of original elementary species 

 which hybridized together to form the original Oenothera La- 

 marckiana. This case of mutation, therefore, is not an instance 

 of the appearance of new species, but simply of the reappearance 

 of old ones temporarily obscured in a combination; and it leaves 

 unsolved the question of the origin of elementary species. 



Third, all the recent work is confirming the reality of the exist- 

 ence of elementary species or biotypes, though it is throwing very 

 little light on their origin. Moreover, and here is a most important 

 point, it is showing that these biotypes, though apparently homo- 

 geneous (and therefore forming a single phwnotype), are in reality 

 composite, since they embrace a good many Mendelian com- 

 binations (or genotypes). But it is not worth while to follow these 

 matters further at present, since we now verge close to the firing 



