2^0 (Jstenielü. 



Of course these countings are based on gatherings of castrated 

 and non- castrated heads of the same individual (generally even from 

 the same shoot) and taken at the same time. How necessary it is 

 to make the comparison only between gatherings from the same year 

 is shown by the following case: in 1908 the offspring of a form of 

 H. boreale Fr. (from Lille) turned out in favour of apogamy 

 (difference: 7 p. ct.), while in 1909 the same individual gave the 

 reverse result (8 p. ct. in favour of fertilization). 



The objection that such a sorting from the outer appearance into 

 full and empty fruits is uncertain and ought to be replaced by germi- 

 nation experiments, will apply in the same degree both to the castrated 

 and to the non-castrated fruits and may therefore be dismissed as 

 unimportant for a purely comparative consideration '). 



If we go through Table I, we still find some deviations so 

 considerable that we cannot explain them as casual. The above 

 mentioned form of H. alpimwi seems to have had advantage from 

 the castration (86: 66), but I think we ought not to regard this case 

 as convincing, as only the fiuits in one non-castrated head have 

 been counted, and in these species with few and large heads the 

 largeness of the heads varies considerably. 



The other more noteworthy deviations are, i": //. sabaiidnm L. 

 where the castration no doubt has diminished the fruiting power 

 (65:92); 2", the case is similar with H. laevigaium Willd., of which 

 there are countings of two different sets and of which one set has 

 a great deviation (58 : 85), the other only a small one (62 : 71), both 

 in favour of fertilization, 3", and finally, a H. vulgatuvi Fr. and a 

 H. rigidiDii Fr., both from Svendborg in Denmark, seem to show 

 the same (55 : 83 and 73 : 94). As to these deviations, the differences 

 in percentage between castrated and non-castrated fruits seem to 

 indicate, that these species have not wholly lost the power of having 

 fertilized fruits-). If we were to try to make hybrids within the 

 Archieracia, we have here hints, which species we are to use. 



'j In order to get an idea how great the germinating power is in the fruits 

 considered as full, I have made a simple germination experiment, laying "fuU" fruits 

 on wet filter-paper under glass in an ordinary room. The result was in one case 

 83 p. ct. germinated seeds, in another 70 p. ct. 



^) The supposition is only of a restrained value, as the investigations are few, 

 and most of the deviations are from 1909 with its rainy summer. Thus it is possible 

 that the frequent rains may have had more influence on the heads made open through 

 castration than upon the intact heads, which are protected by their bracts. 



