THE DETERMINATION OF DOMINANCE. 32 1 



better than I the complexity of experiments of this kind, the dif- 

 ficulties involved in the analysis of the results, and the caution 

 that should be exercised in making statements from them. It 

 seems certain from these experiments, as far as they have been 

 carried out, and they are by no means complete, that we may defi- 

 nitely conclude that when like materials are combined under differ- 

 ent natural environments, differences in the products, depending 

 upon the conditions under which the combination take place, 

 result. It is certain that the type which came out of the culture 

 in the Balsas Valley was quite different from that which resulted 

 from the cultures at Orizaba, and these are different from the 

 dominant type which arose at Tucson. 



Whether or not the dominant types resulting in these experi- 

 ments differ sufficiently to be called species is a matter of opinion. 

 To judge by analogy, I suspect that if a systematist had found 

 these materials in nature they would have been classed as species, 

 or at least would have been given the value of a variety. Since 

 their history is known, I presume that they lose all claim to any 

 specific or varietal distinction, from a systematic standpoint. 

 The experiments, however, clearly indicate that the process of 

 hybridization carried out under divergent conditions in nature 

 gives identical results, as far as principle is concerned, with the 

 crosses carried out under divergent conditions in experiment. 

 It is true that these cultures are group cultures and are synthetic 

 and not analytic, but a series of experiments is being conducted 

 under diverse natural environments, which gives promise of con 

 firming the general point, that conditions incident or external to 

 the germ cell at the time of fertilization may profoundly modify 

 the behavior and the relationships of the characters entering into 

 the crosses. 



One point of very considerable interest is the behavior of these 

 dominant types in exactly the way in which DeVries's CEnothera 

 Lamarckiana behaves, giving in each generation a greater or 

 less number of rather divergent individuals, which, when inbred, 

 are found to be stable germinal variations. I shall report upon 

 this more extensively in another paper. 



Bateson ( 1 902 , p . 1 5 3 ) has suggested that the mutations observed 

 by DeVries in CEnothera Lamarckiana are in reality due to some 



