A Genetic Study of Plant Height in Phaseolus Vulgaris. 35 
vigorous of all the plants grown in 1912. Tho both the parents 
are among the earliest races of beans, the Fi plants of the cross 
were later in flowering and in ripening seeds than any of the other 
crosses considered here. The same peculiarity was observed 
in Fi plants of this cross as grown near Boston, Massachusetts, 
in 1911. The large mean number of internodes, over 31, shown by 
the Fi plants of this cross is probably due directly to the vigor 
of growth and lateness induced by heterozygosis. This is indi- 
cated by the fact that in Fo very few plants had as many inter- 
nodes as the mean of Fi. The Fo plants of the cross between 
Snowflake and Red Marrow had practically the same mean 
number of internodes as did the Fi plants. 
Unfortunately no records are available of the number of 
internodes in F^ of the cross between the two pole beans. Snow- 
flake and July, or of the crosses between July and the bush races 
Triumph and Red Marrow. The variation shown in Fi of the 
cross between Snowflake and July was slightly less than that 
exhibited by the parent races, but the differences are too small 
to be significant. In Fi of the Snowflake-Triumph, Snowflake- 
Red Marrow, and July-Red Marrow crosses, the variation was 
less than in the pole-bean parents but somewhat greater than in 
the bush-bean parents (Table 1). The small number of plants 
grown in Fi families may have been in part responsible for this 
comparatively small variation. In F2 the variation in the crosses 
between Snowflake and these bush races, as measured by the 
coefficient of variation, was greater than in even the pole-bean 
parent and much greater than in the Fi plants. In F2 of the 
Snowflake-Red Marrow cross, the coefficient of variation was 
only 3.81 ±2.08 greater than that of Snowflake but 9.51 ±1.44 
greater than that of Red Marrow. The coefficients of variation 
for Red Marrow, Snowflake, the Fi generation of the cross be- 
tween them, and the F9 pole-bean segregates of the cross were 
respectively 10.76±0.57, 16.46±1.61, 10.73 ±1.29, and 20.27± 
1.32 per cent. It will be recalled that the F2 bush segregates of 
this same cross also showed considerable variation in number of 
internodes (Table 1), the coefficient of variation being 19.65 ± 
2.18 per cent. The standard deviations are widely different for 
these lots, but standard deviation is not a good measure of 
variation for comparisons between plants so unlike as pole and 
bush beans. 
The full significance of the greater variation in the F2 genera- 
tion than in Fi or in the parents, as presented in Tables 1 and 2, 
and discussed above, is appreciated only when it is realized that 
the coefficients were determined for the pole and bush segregates 
separately. If the whole F2 generation of either cross between 
