lO 



BOTANICAL GAZETTE 



[JANUARY 



O. ruhrinervis is far stronger, since its two constituents are both 

 represented in my cultures. This fact makes a complete analysis 

 possible, as I have already pointed out. If O. ruhrinervis is split 

 by some cross into laeta and velutina on account of its composition 

 of gametes of deserens and velutina, then the corresponding cross 

 with O. deserens must evidently give the same laeta and that with 

 O. Lamarckiana mut. velutina the same velutina. Thus the split 

 progeny can be duplicated by the addition of its components. 



I have described the splitting crosses in Gruppenweise Artbild- 

 ung (pp. 122, 196-200, 1913) and repeated some of them so as to 

 have the dimorphic progeny together with the cultures of the pre- 

 sumed constituents, in order to be able to identify their characters 

 during the whole time of their development. The percentage 

 figures given in my book are as follows: 



In the second and third generation of the two latter crosses the 

 laeta have split off brittle ruhrinervis in about one-third of the cul- 

 tures, whereas the velutina remained constant. 



I repeated the two first named crosses in 1915, but not the third 

 one. On the other hand, I have repeated the cross with O. Hookeri, 

 in the progeny of which I had previously not been able to distin- 

 guish the twin types. I had the following cultures in 191 6. Most 

 of these plants flowered in August. 



