272 



BOTANICAL GAZETTE 



[OCTOBER 



was counted in July, during the beginning of the flpwering period, 

 but about one-half of some of the cultures were kept in the boxes 

 so as to reach only the stage of rosettes at that time. No second 

 generations have been cultivated. I will first describe the results, 

 and afterwards give the necessary details concerning the several 

 cultures. 



The main purpose of these crosses was to decide the question 

 whether the special characters of 0. cana are handed down by the 

 ovules only, as inO. scintillans and 0. lata, or by the pollen also. 

 Table IX gives the result of the crosses with wild species other 

 than 0. Lamarckiana. Here a splitting occurs into laeta and 

 velutina or into densa and laxa, exactly analogous to that produced 

 by 0. Lamarckiana itself. The only exception is the pollen of 

 0. biennis, which does not split; it gives with 0. cana the same 

 type as with the parent species. The result was very striking on 

 the beds. No specimens of the cana type occurred in the cases 

 where this mutant had been the father, whereas such individuals 

 abounded in the results of the reciprocal crosses. 



TABLE IX 

 Crosses or 0. cana with other species 



In this table, 0. syrticola Bartlett is the 0. muricata L. of my 

 Gruppenweise Artbildung, and 0. airovirens Bartlett has been 



