TWIN HYBRIDS OF OENOTHERA HOOKERI T. AND G. 287 



from allied species, which must be made responsible for the special 

 splittings observed after so many crosses of this species. 



In 1917 I have continued the cultuie of the new mutant aurea, 

 in order to see whether this character was constant. I found it to 

 be so. I sowed the seeds of three self-fertilized specimens of 1916; 

 they yielded about 600 seedling plants, all of which were pure aurea. 

 No green leaf was seen among them. I planted 180 individuals, 20 

 of which stayed in the condition of rosettes, whereas most of the 

 others flowered in August. There was a high degree of fluctuating 

 variability in their development. Some were weak and died early, 

 others were strong and flowered richly on the main spike and a 

 number of side branches. All intermediate degrees were seen. But 

 the color was golden without exception and remained so as long as 

 the weather was bright and sunny. In July and August they slowly 

 assumed a pale green color, evidently under the influence of the 

 diminishing number of sunny days, the season being very rainy this 

 year. In this respect they behaved like so many horticultural aurea 

 varieties. The whole beds were uniform, apart from the size of the 

 flowers. There were 73 percent of specimens with small flowers, of 

 the size of those of Oe. biennis or smaller (petals about 1.5 cm.) and 

 27 percent large-flowered ones (petals 3.5 cm.). This points to a 

 Mendelian proportion with the small petals as the dominant character 

 and justifies the expectation that the large-flowered will constitute 

 a constant race. This would be one of the most showy among all the 

 mutants of the Oenotheras, apart from its origin from crossed seed. 

 But since neither of the parent species has golden foliage it is evident 

 that this new character must be due to a mutation. In this connec- 

 tion it should here be remembered that it arose only in the third 

 generation, and in eleven specimens. 



Crosses of Oe. suaveolens with Oe. Cockerelli 



In order to control some of the features of the previous cross I re- 

 peated it, replacing Oe. Hookeri by Oe. Cockerelli, a species which is 

 isogamic like the other, but is not known to produce splitting hybrids. 



The combination Oe. Cockerelli x Oe. suaveolens gave a uniform 

 progeny like the corresponding one, described above. It consisted 

 mostly of weak, yellow plants. No second generation has been 

 cultivated. 



The combination Oe. suaveolens x Cockerelli gave in the first genera- 

 tion hybrids of an intermediate type and some aberrant plants with 



