120 



• THE AMERICAN NATURALIST [Vol. XL VII 



tion divisions each with 14 chromosomes. From such nuclei 

 gametes would be formed with 14 chromosomes. 



The position of Stomps is clear. He believes that the Dutch 

 biennis and cruciata have identical germinal constitutions except 

 for the factors that determine floral structure and therefore 

 with respect to other characters can be crossed as though they 

 were homozygous. Since the cross gave two marked variants 

 which differed from the parents in other respects than those of 

 floral structure these two plants are mutants. These conclusions 

 are then applied by Stomps to the problem of the status of 

 Lamarckiana in the following line of reasoning. Since biennis 

 mutates and since biennis is probably an older species than 

 Lamarckiana it follows that mutations among the Oenotheras are 

 older than Lamarckiana and that consequently the mutations 

 of this species can not be the result of hybridization. 



The line of argument rests primarily on the assumption that 

 biennis and cruciata have exactly the same germinal constitu- 

 tion except for floral characters. This I can not believe possible 

 considering the long evolution back of the two lines. Why did 

 Stomps find it necessary to cross bie nnis with cruciata to obtain 

 his "mutants" ? If homozygous in all respects except for 

 flower structure why should not biennis alone or cruciata alone 

 give the same mutants ? From my point of view Stomps really 

 made a cross between two rather closely related species and 

 obtained first the segregation of flower types to be expected in 

 the F 2 generation among which from my experience I should 

 expect a wide range of variation, and second Stomps obtained 

 two marked variants due to some germinal modification as the 

 result of the cross. 



This sort of phenomenon I am obtaining frequently in crosses 

 of my races of American biennis and grandiflora. The nanella 

 condition of dwarfed growth is very common in P 2 generations. 

 And, last summer in an F 8 generation a large plant appeared 

 with leaves so thick and stems, buds and flowers so stocky that 

 I have hardly a doubt but that the cytological examination will 

 show an increase in the number of chromosomes. 



In so far as the observations of Stomps bear upon the problem 

 of mutation my interpretation would be exactly the reverse of 

 his. To me they further illustrate the same phenomenon that I 

 am obtaining through my hybrids of biennis and grandiflora, 

 namely, that behavior by which these hybrids in the F 2 generation 



