450 Tricotylous Races. 
ennial form ; and therefore both of them afford material 
for comparing the harvest of the same plant as produced 
in successive years. Here again I failed to find differences 
of any significance. I have given above the values de- 
rived in 1892 and 1893 from a single plant of Silene, 
which flowered in isolation; they were 3 and 4% for 
these two years. In Scrophularia I made a series of ob- 
servations at the beginning of my cultures when the hered- 
itary values were still small, and repeated them in the 
period of 1896-1899, when they had become higher (15% 
and more). In these years, 1897, 1898 and 1899 six 
plants gave the following values, the bracketed number 
referring to the second year: A 22 (25), B 25 (17), 
C 22 (17), D 23 (25), E 27 (25), F 23 (22). Ob- 
viously these figures do not justify a conclusion as to any 
diminution or increase in the ratio in which tricotyls are 
produced. 
The result of all of these experiments is such as to 
justify my practice of limiting the individual harvest to 
the quantity of seed necessary for sowing. 
8. THE INFLUENCE OF EXTERNAL CONDITIONS ON 
TRICOTYLY. 
Which seeds in a fruit produce aberrant seedlings? 
This question is at once one of the most simple and one 
of the most difficult presented by experimental breeding. 
If some day we could succeed in solving it and thereby 
make a control of this process possible, much light would 
be thrown on a whole series of phenomena connected 
with the origin of races. 
In dealing with this question we are thrown back 
on egg cells and pollen grains and the problem at once 
