280 



TWIN HYBRIDS OF OENOTHERA HOOKERI T. AND G. 



Of course, it is not necessary to use Oe. Hookeri as a male parent 

 for these experiments. The pollen of other species may be chosen as 

 well. Therefore I made some further crosses in 1913, using the same 

 hybrid races, and cultivated next year 60 plants from each cross, 

 in the same way as for the crosses with Oe. Hookeri. Here also the 

 hybrids could be compared with the binary ones and the result was 

 a complete likeness. Moreover, the cultures were uniform, as was so 

 be expected. I got the following results: 



Table 10 



Crosses 



Type of hybrids 



Oe. (biennis x syrticola) xOe. biennis Chicago 

 Oe. {biennis x syrticola) x Oe. Cocker elli 

 Oe. (syrticola x biennis) xOe. biennis Chicago 

 Oe. (syrticola x biennis) x Oe.Cockerelli 



Oe. biennis x Oe. biennis Chicago 

 Oe. biennis x Oe.Cockerelli 

 Oe. syrticolaxOe. bien. Ch. 

 Oe. syrticolaxOe. Cockerelli 



These experiments confirm the results of those with Oe. Hookeri' 

 Excepting the hypothesis of merogony of R. Goldschmidt (1912), 

 which has been refuted by Renner (1913), no explanation of the 

 heterogamous condition of Oe. biennis, Oe. syrticola and some other 

 species has been proposed as yet. If it were allowed to apply our 

 conception of the secondary mutability of hybi id mutants in this case, 

 we might assume the initial mutations, by which these species arose, 

 to have embraced changes which differentiated the sets of paternal 

 and maternal characters. If then a hybrid mutant were produced, 

 it would make two kinds of sexual cells in every generation. It would 

 then be necessary to assume further that the gametes of one set 

 were eliminated in the production of the pollen, and those of the 

 other set in that of the egg cells. This would give constant hetero- 

 gamous species just as our experiments show them to be constituted. 

 Lethal factors would then have to be made responsible for these 

 eliminations, but they would have to do their work before, and not, 

 as in other cases, after fecundation. But it must be left to further 

 researches to give an answer to these problems. 



Crosses of Oe. franciscana with Oe. biennis 



Even as in the case of the crosses of Oe. franciscana with Oe. La- 

 marckiana the strong affinity of the former species with Oe. Hookeri 

 must lead us to expect a similar behavior in its crosses with Oe. 

 biennis. These have been studied already by Davis (1916, p. 217), 

 who found a high degree of variability among their offspring and 

 described a number of types. Among them a large majority were 



