170 OENOTHERA LAMARCKIANA MUT. VELUTINA. 



Cockerelli) velutina, but in 1915 those young plants which had not 

 been planted out in order to flower were a little too crowded. I 

 therefore pulled out those which were indubitably velutina, and 

 planted all the dubious ones on a separate bed, giving them just 

 as much space as in ordinary cultures. About one-half of these 

 (11 plants) flowered in September, but all of them displayed, at 

 that time, the characters of 0. (Lamar ckiana x Cockerelli) velutina, 

 so as to leave no room for doubt. No laeta has appeared among 

 the offspring of this cross. 



0. Hooker i x 0. bland ina. — Only one Hooker i was crossed in 

 1913 with one specimen of blandina of my race. The seeds gave 

 60 offspring in 1914 and 85 in 1915, of which 25 and 10 flowered. 

 In the crosses of 0. Lamarckiana and its derivatives with this Cali- 

 fornian species, the velutina have almost the features of 0. Hookeri 

 itself, showing only a small influence of the other parent. I com- 

 pared my hybrids in both years with first generation hybrids, using 

 for the comparison both the reciprocal crosses of 0. cana of 1913 

 in one year and the hybrids of a cross 0. Hookeri x Lamarckiana 

 of 1909 in the next year. Although the specimens, which were not 

 allowed to flower, made stems of only a few centimeters, or stayed 

 in the conditions of rosettes of radical leaves, the type of 0. Hookeri 

 was strongly pronounced in them. All of them had the long narrow 

 leaves of velutina, and no laeta occurred in the whole culture of 

 145 plants. 



Resuming these details, we see that 565 hybrids of 0. blandina 

 with splitting species have been studied, and that all of them bore 

 the unquestionable features of the velutina of the corresponding 

 crosses of 0. Lamarckiana and of 0. cana. 



Crosses of 0. blandina with 0. Lamarckiana. — If the laeta qualities 

 have become latent and inactive in 0. blandina, we should expect 

 that this mutant would have acquired the property of splitting 

 itself these qualities in 0. Lamarckiana. The confirmation of this 

 expectation must obviously strengthen the conclusion from which 

 it started. And since 0. nanella and other mutants may give rise 

 to the same hybrid twins as the parent species, we may expect 0. 

 blandina to split them also. I made both the reciprocal crosses with 

 the parent species and one with the dwarf, and all 3 cases have 

 corroborated my conception. I made the crosses in 1913 in the third 

 generation of my race, and cultivated the first generation in 1914, 

 repeating it in 1915. The splitting occurred in all 3 cases as expected, 

 giving nearly equal groups of the two types. One of these types 



