286 



TWIN HYBRIDS OF OENOTHERA HOOKERI T. AND G. 



and to an origin from a mutated sexual cell producing in the second 

 generation a half mutant, analogous to those of Oe. gigas nanella. 

 If we combine all these data into a pedigree, we get the following 

 summary: 



Table 14 



Pedigree of Oe. suaveolens x Oe. Hooker i. 

 Cross of 1913. 



Leaving the narrow leaved, small-leaved, aurea and lutescens out 

 of consideration, as probably due to mutations, this pedigree coin- 

 cides with those given in my book (191 3) for the three splitting types of 

 hybrids of Oe. Hooker ix biennis, Oe. biennis Chicago x Hooker i and 

 Oe. cruciata (atrovirens) x Hookeri. 



For the two latter it was assumed that the pollen lacks the hetero- 

 gamous characters of the pistil parent and carries mainly those of 

 Oe. Hookeri. If it is allowed to apply this conception to our cross, 

 it must be possible to replace the pollen of the rubiennis in fecun- 

 dation by the pollen of pure Oe. Hookeri or of the Hookeri-Wkt 

 hybrids of our pedigree without changing the result. Therefore I 

 made these crosses in 1915. In both cases the progeny was the same 

 as that of self-fertilized rubiennis, with the exception that deviating 

 (mutated) forms failed and that the percentage for the Hookeri-\\ke 

 was rather high. It was 73 percent among 70 offspring of Oe. rubien- 

 nis x (suaveolens x Hookeri) Hooker /-like and 60 percent among a 

 progeny of 68 inaividuals from the cross Oe. (suaveolens x Hookeri) 

 rubiennis x Oe. Hookeri T. and G. These results justify the assumption 

 made concerning the pollen of the rubiennis hybrids. 



iMany special features of these experiments require a further in- 

 vestigation, but the main result, viz., the splitting in succeeding 

 generations, seems above all doubts. In combination with my pre- 

 vious researches it indicated a latent quality of Oe. Hookeri, different 



