OENOTHERA RUBRINERVIS, A HALF MUTANT. 



369 



of course will be more complicated. I fertilized a strong biennial 

 specimen of 0. scintillans with the pollen of a plant of 0. rubrinervis 

 and had in 1916 a culture of only 23 plants, all of which flowered 

 in August. There were 5 types: 11 Lamarckiana, 2 lucida, 1 sub- 



robusta, 9 scintillans, and 8 oblonga. The first three were the same 

 as in previous crosses and confirm their result; the last two named 

 types are the same as are always seen in the first generation of 

 crosses of 0. scintillans when this is used as the seed parent. 



Moreover, in 1915 I fertilized 4 plants of my race of 0. lata with 

 0. rubrinervis, counted the lata and albida in their progeny in May 

 1916, and for want of space planted out only a part of the others, 

 in order to distinguish the types, but without trying to determine 

 percentage figures. Altogether I had 434 seedlings, among which 

 7 per cent were lata and 6 per cent albida. At the time of flowering 



I counted 23 Lamarckiana, 1 lucida, 20 brittle rubrinervis, besides 



II mutants (5 oblonga, 5 obovata, and 1 scintillans). No subrobusta 

 occurred in these cultures. These results confirm those previously 

 given. 



Second and later generations. — Brittleness and dwarfish stature 

 are recessive characters, and as such may be expected to be split 

 off in the succeeding generations. For the crosses between 0. rubri- 

 nervis and 0. nanella this splitting has already been dealt with. 

 For the other crosses our analytical formula for 0. Lamarckianax 

 rubrinervis shows that the types lucida and subrobusta may be ex- 

 pected to produce a splitting, whereas the LamarckianaAike hybrids 

 cannot contain the necessary factors. The production of brittle 

 plants from subrobusta had been observed in the case of the dwarfs, 



24 



