2^.2 üstentclci. 



The two sets of H. canadense Michx. give a tendency in fa\'our 

 of castration (95 : 86 and 56 : 45), while the form named //. canadense, 

 aff. gives no tendency (56 : 56). Here more investigations must be 

 made before an opinion can be expressed. 



As mentioned above, H. virga aiirca Coss. is not apogamic; 

 I have tried castration several times, but always got the same result, 

 that the heads are quite sterile if castrated. When left to itself, it 

 fruits rather well, and from these fruits come a very heterogeneous 

 offspring, which indicates that crossing must have gone on. It flowers 

 too late, however, in our latitude to be a good experiment plant. — 



Among the Arcliieracia is still left the group H. umbellatiun, which 

 I have studied more thoroughly, as it seems to me to be particularh' 

 interesting. When, in 1905, it had become evident that four forms of 

 this group required fertilization in order to fruit, or more correctly that 

 they did not fruit after castration, I procured from the seed cata- 

 logues as many forms of this group as possible. I have in the years 

 since then examined altogether twelve forms (sets), see Table II. From 

 the castration experiments made in 1906 I got the result that two sets 

 of H. umbellatum were apogamic, the other sets not. I had, however, 

 to go to work with caution, for the seeds received from the Botanical 

 Gardens were so untrustworthily determined that not even the five 

 experiment plants selected from a sowing were similar. For example, 

 in 1906 in the one experiment, with a H. unibellatuui from Lille, 

 I got seeds after castration, while the two plants of the same set 

 that were examined in 1907 and 1909 gave a negative result. The 

 offspring of the apogamic individual — that evidently has been among 

 the three of the five original experimental plants which died in the 

 winter 1906 — 1907 — gave also fruit after castration, just as the 

 parent had done. This dissimilarity is easily explained when we 

 learn that the apogamic individual belongs, not to H. mnbcllatttm, 

 but to H. boreale, the seeds of which must have been intermingled 

 with those of H. unibellatuin ; while on the other hand the two plants 

 examined in 1907 and 1909 are true H. umbellatiun. In consequence, 

 the one case of apogamy in H. umbellatum has to drop. 



The other set of H. umbellatum, which gave fruit after castration, 

 is from St. Petersburg and was named H. umbellatum, var. linear i- 

 foliiim, but looks like a typical H. umbellatum. Dr. Dahlstedt con- 

 firms that it is a true H. umbellatum. There is then a form of 

 H. umbellatum which is capable to fruit apogamically , but 



