FiirtlKT Studies on tlie Apogamy and Hybridization of the Hieracia. 253 



which in outer appearance does not differ from the other 

 forms of H. unibellatum, which all require fertilization. The 

 differences in percentage of full fruits in the apogamic form between 

 castrated and non-castrated heads seem to indicate that the apogamy 

 is not absolute (79:91, 82:79, 4^:78, 56:86, 61:74, 63:52. and 

 51 : 84). — All the other examined sets of E. iimbellatum, including 

 all those that proceed from spontaneous forms, and further the nearly 

 allied species E. lactaris Bert., have not given any fruit capable of 

 germination after castration and thus agree with the cytological 

 investigations by O. JUEL (1905). We have thus in E. umbellatuvi 

 the peculiarity that most of the forms, "races", are normally 

 sexual, while a single form is apogamic'). 



Looking through Table II it is curious to see how unhke the races 

 of H. imibellatimi are in regard to fruiting and how small the percentage 

 of full fruits is in most cases. It seems to be a species whose fruiting 

 at the present time is little stable and in which we might hope to find 

 something to help us to an understanding of the mysterious pheno- 

 menon of apogamy. Still, I have as yet no hints of this understanding, 

 but I hope that further investigations on this species will clear up 

 the matter. 



3. Subgenus Pilo»ella. 



In my paper of 1906 I could record that E. auriaita L. did not 

 fruit after castration — which Rosenberg's cytological investigations 

 (1907) also confirmed — while the other five examined species of the 

 subgenus Pilosella were able to do so. I have later repeated my 

 experiments with E. auricula, also using other sets of this species, 

 and with the same result. Connected with this fact is, no doubt, 



') In 7?osn something similar seems to be the case. Dinglek (i9°7). "'ho 

 previously supposed apogamy in som? cases, doubts that there is apogamy. and 

 Str.\sburger (1904) has shown that in the species e.xaniined, among others in Ro^a 

 caniiia, the development of the embryo-sac is quite typical and that fertilization is 

 necessary. On the other hand O. Rosenberg (1909, p. I55 — 158) has e.xaniined a 

 form, R. canina, subsp. pcrsaticifolia Almq., in which the development of the embryosac 

 points to apogamy. consequently differing from the "race" of R. canina studied by 

 STR.\SBt;RGER. 



Besides E. Lundström (1909, p. (16)) in a preliminary note has recorded, that 

 he has had fruits developed after castration in R. virentiformis Matss. and in R. f;laiica 

 Vill., subsp. Alzdiana, var. dilatans Almq., which latter according to Rosenberg 

 (1. c, p. 156) must also be supposed apogamic. 



It is remarkable that the number of chromosomes in the apogamic race of 

 R. canina is about double what it is in the normally sexual race of the same species. 



