206 Ostenfeld. 



■in being coarser and somewhat lower in the corymb-scape and with 

 shghtly larger heads, all characters from H. pilosella. The two others 

 (No. 5O4 and 5O5) were also much alike and near H. pilosella, but 

 still greatly differing by the corymb-scapes having several (not solitary), 

 long-stalked heads, whose size was a great deal under that of the 

 heads of H. pilosella. They were hermaphrodite, while the four other 

 individuals [siib-txcelle?is) were female. 



All 6 individuals have in common that their power of fruiting 

 is very slight. Notwithstanding repeated experiments in different 

 ways I have not yet succeeded in getting more than a few apparently 

 full fruits in any experiment and in most of them no full fruits at 

 all. Thus, after isolation I have got only a single Fj-individual of 

 50 2, of 50 3 and of 50 g each, — that is of three individuals of the 

 sub-exce/lens-group; these F2 individuals have all been like their F^. 



An experiment of crossing 5O5 with the mother parent (H. pilosella) 

 gave rise to only three individuals which were different from one 

 another and all three different from their F^-parent. One (No. 4061) 

 was, at least apparently, pure H. pilosella, and consequently a 

 complete segregation must have taken place. The other two showed 

 hybrid characters; one of them (No. 4063) was rather near the mother 

 (the primary hybrid), stiU somewhat nearer //. pilosella, and the third 

 (No. 4062) was verj' peculiar; it had leaves narrower than those of 

 both parents, long thin stolons, mostly approaching the excellens-type; 

 the corymb bore long branches with few flower-heads, much as the 

 inflorescences on the stolons of H. excellens. Unfortunately both the 

 last named plants were weak and died in the winter after the first 

 flowering; they seemed to be quite sterile. 



The cross between H. excellens and H. pilosella has thus given 

 only the following result: the primary hybrid is heterogeneous, 

 but less than in the other crosses, being nearly dimorphous^). All 

 the individuals of F^ are nearly sterile, the few individuals of 

 F2 point to constancy, while the poor experiment of back-crossing 

 may be said to show segregation. 



On the whole, this hybrid combination seems to be unfavourable 

 as a subject of experiment. It has therefore not been used so much 

 for experiments, as interest in its peculiar behaviour would merit. 



If we sum up the result of all the hybridization ex- 

 periments hitherto made, we get the following conclusions: 



1) Such cases H. de Vries (1907, 1908) has named Twin- hybri ds. 



