TWIN HYBRIDS OF OENOTHERA HOOKERI T. AND G. 273 



empty seeds. The splitting is, therefore, to be considered as unilateral. 



The phenomena afforded by these twins are exactly the same as 

 in the case of Oe. grandiflora and therefore require the same expla- 

 nation. The difference in the percentage figures does not seem to 

 be essential, since in both cases the same figures were found for the 

 first generation ana for the offspring of the self-fertilized laeta. 

 The figures for Oe. grandiflora indicate equality of the two groups, 

 but in the case of Oe. Lamarckiana there is a loss on the side of the 

 laeta. If we assume for Oe. Lamarckiana a mass mutation into velu- 

 tina, analogous to that of Oe. grandiflora, but hidden by a linkage with 

 the lethal factor which produces the empty seeds (de Vries 1918 b), 

 exactly the same explanation of the twins holds good for both cases. 

 In both, moreovei, the laeta do not split after self-fertilization into 

 three or more groups, according to the laws of Mendel, but only 

 into two. This fact, evidently, demands a special explanation. 



All our suppositions are deduced from one point of view, a perfect 

 analogy between Oe. grandiflora and Oe. Lamarckiana in the pro- 

 duction of their twins, due to the same principal cause. Oe. grandi- 

 flora mut. ochracea and Oe. Lamarckiana mut. velutina do not differ 

 from their parent species in one character only but in a whole set 

 of such, one of which is the individual weakness in one case and the 

 early death in the other. The absence of the parental type is com- 

 mon to both, but other properties differ. So, e.g., the leaves, which 

 are broad in ochracea but nanow in velutina, smooth in the first 

 but hairy in the second, and so on. 



From this discussion it is easily seen that our hypothesis of a 

 mass mutation and lethal factors is apt to simplify our conception 

 of the twin hybrids and to bring them into causal relation with the 

 phenomena of mutability in general. 



For this reason it seems desirable to give here a review of the 

 instances of constant and of splitting laeta. Constant laeta are pro- 

 duced by Oe. Lamarckiana and some of its derivatives (e.g., Oe. 

 brevistylis and Oe. nanella) as well as by Oe. grandiflora after crosses 

 with the female gametes of Oe. biennis and Oe. syrticola with the 

 male gametes of Oe. biennis Chicago, with both kinds of sexual cells 

 of Oe. Cockerelli, and in some other instances. If we assume the 

 gametes of Oe. Lamarckiana and of Oe. grandiflora to consist of two 

 types, one of which produces the laeta hybrids and the other the 

 velutina, we have to consider each of these combinations as a uni- 

 sexual cross, the laeta-velutina character finding no antagonist in 

 the other species. According to the conceptions, set forth in my book 



18 



