414 Tricotylous Races. 
-23 25 and 27%. I planted out 165 tricotylous seed- 
lings from amongst their offspring in pots, as soon as 
they proved themselves also to be ternary in their first 
whorls of leaves. In the following whorls about half of 
them reverted to the decussate arrangement. These were 
thrown away, and only 72 plants which remained ternary 
were ultimately planted out. They were fairly uniformly 
distributed over the crops of the five parents. 
From these 72 plants, 72 values, in five groups, were 
calculated in the following spring (1900). The lowest 
values, arranged in a series corresponding with the in- 
creasing values of the five parents, now grandparents, 
were 9, 8, 13, 8 and 11%, the means 16, 17, 18, 17, and 
19%, and the maxima 19, 22, 26, 22 and 26%. As we 
see the five groups did not exhibit any essential differ- 
ence. The mean value of the previous generation, 12%, 
had now been exceeded, but the maximum remained the 
same. 
Let us now summarize the whole experiment in the 
following table : 
VALUES FOUND IN THE FOLLOWING 
TION FLOWERING 
GENERA- YEAR OF SPRING 
LOWEST MEAN HIGHEST 
I. 1890 
II. 1891 
III. 1892 1 
IV. 1893 0-0.3 1-2 5.4 
V. 1894 0.5 2 5.5 
VI. 1895 2 4.5 8 
VII. 1896 2-3 6-7 14-15 
VIII. 1897 2-3 7-8 15 
IX. 1898 2 12 25-27 
X. 1899 8 16-19 26 
Progress is, as we can see, a continual one, and the 
selection has been, although perhaps only indirectly, (p. 
