eng 
bilities of error which are worth considering, are 1°, that the pollen- 
grains might have germinated in the anthers before we castrated 
the head, and 2°, that the stimulus of castration might have been 
the exciting cause of the development of the embryo. Although 
both those suppositions seem improbable, they cannot be refuted 
until other experiments with an exclusively female species has 
reduced the possibility to zero. 
The castrated flower-heads develop like the un-wounded ones, 
but are easily recognizable by their 
shortness (see the text-figure) ; when 
the fruits are fully ripened, they 
force the bracts backwards as usual, 
but the aspect of such an opened 
castrated head with fullgrown fruits 
is rather curious because of the 
short pappus-rays. 
The plants used in the experi- 
ments have all grown in the open 
ground except some few which 
were transplanted into pots and 
isolated in a cold-house, but no 
other precautions were taken, and 
the castrated heads developed, un- 
covered, from the time of castra- 
tion to the harvest of the fruits. 
The fruits were sown as soon 
as possible, partly in the autumn 
of the same year, and partly in the 
next spring and always in baked soil to prevent foreign seeds. 
When they had germinated, most of them were thrown away, but 
in some case the plants have been kept for other purposes. 
With regard to the determination of the species it has been a 
very difficult business, and indeed it is practically impossible 
for one who is not a specialist to identify the species in this 
genus. I have done my best, but I fear that some names may be 
wrong. For some of the species Mr. Dahlstedt of Stockholm 
has kindly given me the names, and I wish here to express my 
best tanks to him for his valuable kindness. In our case a deter- 
mination going to the elementary species is also not necessary; 
when the collective species is correct, it will be sufficient. 
Fig. 1. Three heads of Hieracium 
sp., of which two grow out after 
castration. 
