168 OENOTHERA LAMARCKIANA MUT. VELUTINA. 



the percentage of laeta quite sensibly. These deviations, however, 

 have been amply studied in my work on Gruppemveise Artbildung; 

 they show what size the cultures must be in order to prove the 

 absence of laeta. I cultivated 60—80 offspring for each culture, 

 and repeated the same cultures during 2 years, making on the average 

 140 specimens for each cross. The first year I compared them with 

 the hybrids laeta and velutina derived from the corresponding 

 crosses of my new dimorphic mutant 0. cana 1 ), which happened to 

 be at hand in a complete set; but the second year I have sown the 

 twin hybrids issued from crosses of 0. Lamarckiana itself with the 

 same species as were used for the crosses of 0. bland ina. The material 

 is externally the same in both cases, and quite as good for the com 

 parison, but in the latter instance the proof is a more direct one. 

 I grew the hybrids of 0. blandina and the control cultures side by 

 side, and compared them from the beginning of germination in 

 February until the ripening of the first fruits in September. The 

 differences between the two twins of a cross are large and striking 2 ), 

 from their very first leaves, and it was therefore impossible that 

 a laeta among the hybrids of 0. blandina could have escaped ob- 

 servation. As a matter of fact, no single laeta appeared, although 

 altogether over 500 hybrids were cultivated. 



For these crosses I chose 2 heterogamic and 2 isogamic species, 

 and made the combinations in both groups in the opposite direc- 

 tions, so as to use 0. blandina twice as a pistil and twice as a pollen 

 parent. I made these pollinations in 1913, in the third generation 

 of my race, choosing the most vigorous individuals which had had 

 a normal green color from their very youth. The crosses with the. 

 heterogamic species were 0. syrticola Bartlett 3 ) x blandina and 

 0. blandina x 0. biennis Chicago. Those with the isogamic forms 

 were 0. Hookeri x 0. blandina and 0. blandina x 0. Cockerelli*). 

 These combinations would have given the twins laeta and velutina 

 if 0. Lamarckiana had been used instead of 0. blandina. Now, 

 however, they gave only the hybrid velutina, showing that the 

 splitting capacity of the parent species is absent in the mutant. 



From these facts it follows that in 0. blandina the laeta qualities 



») New dimorphic mutants of the Oenotheras. Bot. Gaz. 62 : 249-280. figs. 

 5. 1916. Opera VII, p. 83. 



2 ) For descriptions and photographs as well as for colored plates of the twins, 

 see Gruppenweise Artbildung. 



3 ) This is the 0. muricata L. of my Gruppenweise Artbildung. 



4 ) For description and figures of these species see Gruppenweise Artbildung. 



