Tricotylous Races Do Not Arise by Selection. 405 
spending number of offspring which manifested these 
numbers amongst their seeds). 
P 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 21 25 
A 16 27 30 18 18 11 6 4 2 2 2 3 2 1 2 1 2 1 l 
This series only relates to the tricotyls selected for 
my experiment. If I had planted out the whole progeny 
of the parent in question without selection, the number 
of low values would most certainly have been somewhat 
larger; and the apex of the curve would perhaps have 
stood over the instead of over 3%. But the chief point 
in this series is that from 3% onwards the figures regu- 
larly and continually diminish in such a way that the 
two extreme variants with 21% and 25% round off the 
series continuously; i. e., in the ordinary manner pecu- 
liar to physiological curves. It is obvious that they do 
not stand in discontinuous relations to it. 
In the estimation of the higher individual values a 
latitude of 5% must be allowed, i. e., the figure 15 de- 
notes a real capacity of from 10 to 20%, the figure 25 
one of from 20 to 30%. In the case of extreme devia- 
tions it is always well to assume that these errors may 
have operated in the same direction. We may only state, 
therefore, that by means of a cultivation through eight 
generations, a ratio of 20% has been attained. But 
whether or not this is the limit, we do not know. Also, 
it is uncertain whether the parent with 25% was really 
better than those with 15 to 21%. But it is certain that 
the best seed-plant, as judged by its hereditarv value 
must be one of this group. 
Therefore, at this stage in the development of our 
race, the process of selection has become much more 
certain than before and less dependent on chance. It is 
only necessary not to limit our choice to those which 
