452 Tricotylous Races. 
(p. 332). It is natural to inquire, therefore, whether this 
vigor of the seeds can be increased by improved nutri- 
tion of the plant. Unfortunately we must, in such cases, 
be content with averages, and these obviously give much 
smaller differences than the single extreme cases. Some- 
times I found that the external conditions exerted a con- 
siderable influence, but, as a rule, I could not detect any 
difference. 
In the first place I shall refer to the effects of manu- 
ring. The prevailing opinion is that manuring with ni- 
trates favors the development of leaves, but that phos- 
phates favor flowers and fruits. As a nitrate manure 
I employed horn-meal combined with dried guano and 
as phosphate manure the ordinary superphosphate ; but 
I find that the former usually produces more abundant 
flowering and a larger harvest of seed than the latter. 
In 1899 I instituted a comparative experiment with Ocno- 
thcra hirtella by planting out the seedlings of a single 
self-fertilized plant. During the period of flowering each 
inflorescence was guarded from the visits of insects by 
a parchment bag. The nitrate plants grew more rapidly 
and luxuriantly at first than those on the phosphate bed. 
Moreover they began to flower somewhat earlier, and 
consequently set more seed ; for fertilization must be 
stopped before all the flowerbuds have opened, since in 
our climate flowers which open in September have not 
time enough left to ripen their seed. 
The harvest for each individual plant was examined 
separately, and from this was calculated, in the way we 
have so frequently described, the percentage value for 
each plant. In the following summary these are associ- 
ated in groups of from 15-25, 26-35, 36-45; the means 
of the groups being 20, 30, 40 etc. 
