A Genetic Study of Plant Height in Phaseolus Vulgaris. 
33 
individual plants used in making the crosses. Owing to the small 
number of plants in any one progeny, progenies of like breeding 
were lumped together in making up the arrays. 
In mean number of internodes Red Marrow and Triumph 
differed by only 0.20 ±0.06 of an internode, or about three times 
the probable error of the difference. The means for the Fi and 
Fs generations were almost exactly half way between the means of 
the parents. It is possible that the two parent races are not 
inherently different but that the slight difference observed in 
mean number of internodes was a mere matter of chance. The 
same would then apply to the intermediate number of internodes 
of Fi and F2. The number of Fi individuals was so small that 
the determination of the mean for plants of that generation has 
little value. It should also be noted that the difference between 
the mean number of internodes for Fo and that for either parent 
lot was less than twice the probable errors of these differences. 
The variation in internode number as measured by the coeffi- 
cient of variation was somewhat greater in Fo than in either parent. 
The differences, however, are only from two to three times their 
probable errors. The most that can be said, therefore, is that, 
if this difference between the parents and the F2 generation has 
any significance, it is that there is a slight inherent difference 
between the parents and that the somewhat increased variability 
in F2 indicates segregation of the factors that differentiate the 
parents.^ 
The bush bean segregates in F2 of the crosses between Triumph 
and July and between Triumph and Snowflake exhibited about 
the same intensity of variation as the F2 plants of the cross be- 
tween Triumph and Red Marrow. The variation was somewhat 
greater in case of the F2 bush segregates of the cross between 
Red Marrow and July, the coefficient of variation being 13.98 =1= 
1.69 per cent. While this is greater than that of Red Marrow, 
which was 10.76 ±0.57 per cent, it is less than that of July (see 
Table 2), which was 15.61 ±1.42 per cent. Considered in con- 
nection with their probable errors, the differences — 3. 22 ±1.78 
and 1.63 ±2.21 respectively — cannot be considered significant. 
It cannot therefore be told, without further breeding tests, 
whether the variation exhibited by these F2 bush plants was 
due to segregation of factors influencing internode number. It 
seems possible, however, that a race of indeterminate growth 
^ Here and consistently elsewhere in this paper the facts are interpreted 
in terms of the multiple-factor hypothesis solely because that hypothesis seems 
to the writer to afford the simplest adequate interpretation. He desires, 
however, to disavow any dogmatic notions as to the truth of this or any other 
hypothesis. A discussion of the matter appears later in this paper. 
