ON THE CORRELATION BETWEEN SOMATIC CHAR- 
ACTERS AND FERTILITY. II. 
ILLUSTRATIONS FROM PHASEOLUS VULGARIS 
J. Arthur Harris 
I. Introductory Remarks 
The purpose of this paper, like that of the one which preceded it,^ 
is the presentation of quantitative data toward the solution of the 
relationship between the degree of development of the somatic organs 
of the plant individual and its fertility. 
The materials are drawn exclusively from pedigreed dwarf varieties 
of garden beans, Phaseolus vulgaris. These varieties are being de- 
scribed in biometric terms elsewhere, as shown in the bibliography; 
hence it is only necessary to designate the series by the key letters 
which open the detailed information of other papers. 
The characters here treated are number of pods per plant, number 
of ovules formed per pod and number of seeds matured per pod. 
II. Analysis of Data 
Table I shows the correlations, r-poi rps, and the partial correlation 
oTps, when p = pods per plant, o = ovules per pod, 5 = seeds per pod. 
The actual number of plants and the number of pods counted are also 
entered. Because of the great bulkiness of the data it is not feasible 
to publish even in condensed form the 64 correlation tables from which 
these constants were deduced.^ 
The correlations of this table are also shown in diagram i . 
I turn first to the question of linearity of regression. This is es- 
sential statistically, since upon the distribution of the means of the 
arrays of ovules or seeds in a sensibly straight line depends the va- 
1 Biometrika 8: 52-65. 1911. 1 
2 The number of pods per plant was weighted with the number of pods counted 
in calculating mean and standard deviation of number of pods. Sheppard's modi- 
fication was nowhere applied, variation being by discrete units. The number of 
pods in which the number of ovules and seeds was determined was used as n in the 
calculation of the probable errors. 
398 
