290 TWIN HYBRIDS OF OENOTHERA HOOKERI T. AND G. 



xOe. biennis. Its first generation consists uniformly of rubiennis, 

 but this splits off, in the succeeding generations, specimens of the 

 type of Hookeri, which are constant in their progeny. Besides these, 

 however, a noticeable number of different mutants appear. 



12. Corresponding mutants were observed among the progeny of 

 the cross Oe. suaveolensxOe. Cockerelli, the hybrids of which do not 

 show any regular splitting. 



13. In their crosses with the pollen of Oe. biennis, the new types, 

 Oe. franciscana and Oe. Lamarckiana mut. velutina, repeat the split- 

 ting produced by Oe. Hookeri. The first generation is uniform and 

 shows a predominance of the characters of the pollen parent. The 

 second splits into this same type and another, which is almost like 

 the second grandparent. In the numerical proportions these latter 

 prevail, about 20—30 percent being rubiennis and about 70—80 

 percent belonging to the other type. 



14. The crosses of Oe. grandi flora and Oe. Hookeri, both of which 

 are large-flowered species, produce hybrids which split off small- 

 flowered specimens. The size of these flowers is about that of Oe. 

 biennis L. The large flowers dominate, the small are found in a pro- 

 portion (27 percent) which corresponds to the formula of Mendel for 

 monohybrids. 



15. In the hybrids between the large-flowered Oe. suaveolens and 

 Oe. Hookeri, a splitting as to the size of the flowers was also observed, 

 but here the small size, corresponding to that of Oe. biennis, was 

 dominant over the large flowers of both the grandparents. 



16. From this same cross a new and beautiful mutant arose, being 

 a pure aurea, with bright golden leaves. It appeared in the third 

 generation in eleven specimens, without intergrades or transitions, 

 and yielded a uniformly golden second generation. 



17. Summarizing all these facts we find that splitting laeta and 

 splitting rubiennis are produced by three large-flowered races, and 

 by these only, as yet. These races are two Californian species: Oe. 

 Hookeri T. and G. and Oe. franciscana Bartlett, and a mutant which 

 originated in my garden, Oe. Lamarckiana mut. velutina (syn: Oe. 

 mut. blandina). In the large number of other cases the hybrids of 

 the Oenotheras, whether twins or monotypic, are constant in their 

 progeny, apart from the rare cases of Mendelian splitting. These 

 laeta and rubiennis always split into two types, one of which repeats 

 their marks, whereas the second resembles the other grandparent. 

 No constant specimens of laeta or rubiennis have been found among 

 them. 



