5f 



Ostenfeld. 



the pronounced disposition of this species to hybridize, as will be 

 mentioned later. 



All the other species of Pilose/la examined in 1906 — 1909 hav'e 

 been able to give fruit after castration. My investigations, unfortuna- 

 tely, do not comprehend so many species as is desirable — altogether 

 23 sets belonging to 14 species. The species of Pilosella have com- 

 paratively small heads and are therefore, technically, rather difficult 

 to castrate, the heads easily being spoilt during the operation, 

 especially in the very small-headed forms. The small fruits are also 

 difficult to count from their outer appearance, and in several species 

 each head gives few fruits. All these circumstances impede the 

 investigation. The countings tabulated in the Table III are therefore 

 much more incomplete than in the case of Arc/iicraciuvi, and the 

 numbers given are admittedly too small. Still, I think they point 

 to the general conclusion that the number of full fruits is somewhat 

 diminished by castration — in other words, that the apogamic 

 species of Pilosella are not absolutely apogamic, but that 

 some of the flowers of each head require fertilization. 



There are altogether 7 sets that can be used for comparison, 

 and, of these, five distinctly point in the direction named. Of the 

 two pointing the opposite way, one {H. florentviuiii All.) at least is 

 too insufficiently investigated, as the numbers are small. The other 

 is a set of H. aiirantiaann and is hardly very convincing. Just about 

 this species Mendel (Correns 1905) has written, that it is absolutely 

 unable to be used for crossing e.xperiments (that is to say as mother 

 plant, but its pollen can very well fertilize other .species), and this 

 would agree well with its being absolutely apogamic. Further, another 

 set of H. aiirantiacujn gives the opposite result. The general conclusion 

 from these cases is, therefore, in my opinion, that the countings given 

 are without value, so far as H. aurandacinii is concerned. 



Although the experiments with species of Pilosella are not so 

 comprehensive as they ought to be, we may sum them up in the 

 following general sentences: Within the subgenus Pilosella the 

 species H. auricula is absolutely sterile after castration'), 

 while the other species examined are capable to fruit apo- 

 gamically, yet apparently in such a way that, at least in 

 most species, a small part of the flowers require fertili- 

 zation. This result agrees with Rosenberg's cytological investigations 

 (1907) and with the great number of hybrids known in this subgenus. 



') Several experiments seem to indicate that it is even self-sterile. 



