HARRIS: FACLORS INFLUENCING THE WEIGHT OF BEAN SEED 491 
given number of pods is small, this weighting will, I believe, tend 
to result in abnormal irregularity of the means of arrays. 
biological reason for this condition I hope to discuss later. 
The 
Notwithstanding the conspicuous irregularity of the means, 
the number of regression diagrams in the two figures indicate that, 
taken by and large, a straight line represents the change in mean 
seed weight associated with variation in the number of pods on the 
plant as well as any other single equation would do. 
It is quite idle to apply mathematical tests for goodness of fit 
in these series, for these involve the calculation of a measure of 
interdependence which is greatly influenced by the special dif- 
ficulties. which are met in graphical tests for linearity. 
Thus, notwithstanding difficulties inherent in the materials 
the correlation coefficient seems the best numerical expression of 
the degree of interdependence between the variables. 
Numerically they are of a low 
All of the 27 values are positive. 
Fic. 3.. Comparison of correlations between number of pods per plant and 
seed weight with that between number of pods per plant and number of ovules per 
pod in 27 cultures of Phaseolus. 
