Influence of External Conditions on Tricotyly. 451 
becomes a double one. To these is added a third factor, 
namely the union of the two germ cells, which may itself 
be influenced by outward conditions in different ways. 
I have gone into the questions only so far as was 
necessary for the choice of the conditions of my experi- 
ments. The course of these has, as a rule, been very 
regular. The results of selection seldom have been dis- 
turbed, to any great extent, by other influences. Some- 
times, however, this did obviously occur ; and it is exactly 
experiments of this kind which justify us in ascribing 
an important part to external conditions in the determi- 
nation of the proportion of tricotylous seedlings. But 
when, from these facts, we proceed to analyze this in- 
fluence, we do not succeed in making experiments in 
which the same influences have the same results. 
Two instances will suffice. Under very peculiar con- 
ditions which affected the whole growth of the plant, 
the hereditary capacity of my race of Amarantus spe- 
ciosus suddenly rose, in 1891, from 4.5 to 13%, with- 
out, however, maintaining that proportion even under 
selection (p. 407). On the other hand, in the summer 
of 1896 the whole culture of Silcnc inflata underwent 
considerable deterioration as the result of unfavorable 
conditions, the mean of all values falling from 53 to 
32%. One year may obviously affect the plants quite 
differently from another, even when the treatment to 
which they are subjected is as uniform as posoible. The 
effect is perhaps produced in the early stages of the plant, 
perhaps also in the development of the sexual cells and 
at the time of fertilization. Abundant starting-points 
for further investigations may here be found. 
As a rule the likelihood of obtaining anomalies in- 
creases with the vigor of the seeds which produce them 
