244 Ostenfeld. 



more easily than I am, to procure rich material of seeds. It will be 

 specially interesting, if he could get seeds of the peculiar Andine 

 species. 



2. Subgenus ArcMeraciutn. 



The castration experiments with species of the subgenus Archie- 

 racium have been continued in 1906 — 1909, always with new species, 

 in order to examine as many as possible of the immense number of 

 species of this subgenus. In this way it was hoped to ascertain how 

 common apogamy is in this subgenus and whether certain sections 

 perhaps differed from the others. 



I think it appropriate here briefly to explain my method of 

 investigation, whose main point is the so-called "castration" invented 

 by C. Raunkiaer (1903): 



The seeds of the species to be examined are always sown in 

 small pots with baked soil to prevent contamination by other 

 Hieracia; when they have germinated, they are transplanted to 

 somewhat larger pots with good nutritive soil (the baked soil is 

 not good for the growth of the seedlings), and when they have 

 reached a sufficient size, generally 5^} individuals are planted out 

 in beds. If the sowing has taken place in spring, some species come 

 to flowering in autumn, and the experiments may then be made, 

 but a good and plentiful flowering and favourable conditions for ex- 

 periments are generally not obtained before the next summer. When 

 the flowering has begun, a number of heads of one individual, and 

 preferably in the same shoot of the plant, are chosen for castration; 

 they must not yet have opened, but must be so near flowering, that 

 it can be expected in 2 or 3 days. The castration is made by cutting 

 off, with a sharp razor, the upper half of the head. By this operation 

 are removed: the upper part of the involucral bracts, the corolla to 

 the tube, the anthers and the upper part of the styles with the 

 stigmas, besides the uppermost part of the pappus rays. There remain: 

 the lower part of the bracts, the ovary and the lower part of the 

 pappus. It is thus not a castration in the proper sense of the word, 

 but an operation which removes both the male element and the con- 

 ducting parts of the female element. 



The castrated head becomes covered on the wounded surface 

 with a quickly coagulating latex, which, however, later, as a rule, drops 



1) In experiments to show the heterogeneity or homogeneity of crosses and the 

 like, of course, all the material is used. 



