292 GOTE TURESSON 
Before summarizing the facts with regard to the different forms of 
H. umbellatum a few words should be said of the castration experi- 
ments and the crosses made. It is known that nearly all the species 
belonging to the subgenus Archieracium are apogamic; H. umbellatum 
belongs to the very few sexual ones (OSTENFELD, 1906, 1921). The 
castrations I have made in order to find out whether this rule holds for 
the different forms of the species in my cultures includes trials with 
two individuals (5—10 flower-heads on each plant) each from the list 
Fig. 56. H. umbellatum. Fı-plants from a cross within the Falkenberg dune type. 
nos. 3, 10, 11, 16, 18, 26 and 33. The upper half of the unopened 
flower-heads was cut off with a razor in the way recommended by 
ÖSTENFELD. The results were in all cases negative; no fruits developed. 
Other experiments with the same plants in the same year (1921) 
showed that these plants were self-sterile when isolated with pergamin 
bags. In a few heads (list nos. 26 and 33) the seeds seemed at first to 
begin to ripen but none gave full fruits capable of germination. Crosses 
between some of the varieties were made during the summer of 1922, 
and these will be followed up later on. A few crosses between different 
plants of the same variety were made in 1921. Although the number 
