272 TWIN HYBRIDS OF OENOTHERA HOOKERI T. AND G. 



and others. The offspring of the laeta, however, split into two main 

 types. I counted among 60 flowering plants, 52 percent laeta and 

 48 percent velutina, both repeating the characters of the first gene- 

 ration. I have also made the reciprocal cross, Oe. grandiflora x Oe. 

 franciscana, but had only eleven plants of the first generation, 

 which, however, showed exactly the same types as the first mentioned 

 instance. There were 4 laeta and 7 velutina. 



From these experiments we see that Oe. franciscana behaves in 

 these crosses in the same manner as Oe. Hookcri. 



Crosses of Oe. Lamarckiana with Oe. Hookeri 



Both reciprocal crosses between these two forms yield the twins 

 laeta and velutina, but whereas these twins are constant in their 

 progeny in other cases, the laeta from these crosses splits in its 

 succeeding generations into laeta, which repeat the splitting, and 

 into velutina with a constant progeny (de Vries 1913, p. 131 — 132). 

 Moreover, Oe. Hookeri as well as Oe. Lamarckiana are isogamic, 

 carrying in their pollen and in their ovules the same hereditary 

 qualities. The dimorphic groups to not contain their types in equal 

 numbers, the average proportion being in the first as well as in the 

 following generations, about 25 percent laeta and 75 percent velutina 

 (idem, pp. 227—228 and 245—256). No constant laeta have been 

 found. In consideration of some results with Oe. mut. blandina, 

 which will be described elsewhere, I was especially interested in 

 the question, whether the first generation would perhaps contain 

 such, and would thus in reality embrace three types instead of 

 two. Therefore I sowed in 1915 the seeds of a cross Oe. Hookeri x 

 Oe. Lamarckiana made in 1909, got a culture of 81 specimens, among 

 which 9 were laeta, retained only these and self-fertilized all of them. 

 The next year (1916) I sowed the seeds, planted 60—70 specimens 

 from each parent and counted their twins in June, when the plants 

 were large rosettes of radical leaves showing their differences sharply. 

 All of the nine groups were dimorphic with 5—11 laeta among a 

 far larger number of velutina. The average percentage was 13. Some 

 boxes were kept during the whole summer, in order to control the 

 countings of June. 



In the same lots of seeds I counted the germs in one hundred 

 grains for each of the nine laeta of 1915. I found the following figures: 

 88, 90, 91, 91, 91, 92, 93, 94, 98, with a mean of 92 percent. From 

 these numbers I conclude that these laeta do not contain hereditary 



