OENOTHERA RUBRINERVIS, A HALF MUTANT. 363 



the corresponding cross with 0. Lamar ckiana, and they were com- 

 pared during all the time of their development. I could not find any 

 differences. The descriptions for the deserens laeta are exactly the 

 same as those given previously for the Lamarckiana laeta. Although 

 I have not made the cross 0. syrticolaxO. rubrinervis, I have added 

 the second experiment of the table. From this and the result of 0. 

 syrticolaxO. blandina (mut. velutina) described in my former article 

 the result of the cross 0. syrticolax rubrinervis may be predicted, 

 and so it would be in other cases also. 



Summing up the results of these experiments, we see that in pro- 

 ducing twin hybrids 0. rubrinervis is split in exactly the same way as 

 an artificial mixture of about equal parts of gametes of 0. deserens 

 and 0. mut. velutina would be. The conclusion that its gametes really 

 possess this dimorphy is thereby as clearly proven as might be 

 expected. 



Crosses of 0. rubrinervis with 0. Lamarckiana and its derivatives. — 

 In Gruppenweise Artbildung I have described the first generation 

 of these crosses as consisting of two types, 0. Lamarckiana and 0. 

 hybr. subrobusta. The 'atter is a rubrinervis in which the brittleness 

 fails, and thereby very similar to our new mut. erythrina 1 ); but this 

 similarity is only an external one, since after self-fertilization the 

 hybrid subrobusta splits off, as a rule, brittle rubrinervis plants, 

 whereas the erythrina produces the decipiens, which is not brittle. 

 Shortly after publishing my book, however, I discovered in the 

 summer of 1913, among the progeny of a cross of 0. rubrinervis and 

 0. Lamarckiana, a slight difference among the Lamar ckiana-Wkt 

 plants. Some of them were stouter and had broader and less crinkled 

 leaves than the others. I self-fertilized them and got a culture, which, 

 although not uniform, repeated the deviating marks of the parental 

 type in the majority of the individuals. I shall call this hybrid type 

 lucida. Moreover, in making a large number of crosses of individuals 

 of the same family of rubrinervis with Lamarckiana plants from 

 various sources, as well as different mutant strains, I discovered 

 that the second hybrid type is not always the solid subrobusta, but 

 sometimes the brittle rubrinervis. I have not as yet discovered why 

 this should be so. We should expect the brittleness to be recessive 

 to the production of strong fibers, and as a rule it is so, but not always. 

 The two contrasting cases have occurred mainly in strains derived 

 from different initial plants, and some hidden mutation might be 



x ) Zeitschr. f. Ind. Abst. 16:262. 1916. Opera VII, p. 131. See also p. 386. 



