08-4 



THE AMERICAN NATURALIST [Vol. XLVII 



which developed to maturity. The method is perfectly 

 legitimate, providing the samples planted be drawn in a 

 purely random manner (as they were in these experi- 

 ments), but the probable error of random sampling is a 

 two-fold one, and this increases the difficulty of deter- 

 mining the statistical significance of a given constant. 



(b) The result seemed, a priori, improbable. Other 

 studies 2 had demonstrated a moderately low but consist- 

 ently positive correlation between the weight of the seed 

 planted and the number of pods on the plant produced. 

 It seemed reasonable to assume that since the larger seed 

 produce the heaviest plants they are in general more 

 vigorous, and hence should be more viable. 3 If the seeds 

 increase in vigor and viability from the smallest to the 

 largest, one would anticipate an increase in the mean of 

 the survivors and a decrease in their variability result- 

 ing from the mortality in the lower part of the range of 

 variation instead of a reduction in variability without a 

 change in type (mean). These were the biological hy- 

 potheses which led me to question the generality of the 

 statistical findings. Further work was therefore under- 

 taken along various lines. Additional field cultures in 

 which it will be possible to compare the constants of the 

 seeds developing with those of the seeds failing, were 

 made. 4 Such cultures can only substantiate, refute or 

 modify the conclusions drawn from the experiments 

 already carried out, but will not advance our knowledge 

 of the proximate causes of the differential mortality. To 

 this end, physiological (including chemical and physical) 

 studies must be instituted. 



The purpose of this paper is to discuss the results of 



2 Harris, J. Arthur, "On the Relationship between the Weight of the 

 Seed Planted and the Characters of the Plant Produced," Biometrika, 9, 

 pt. 1, 1913; also "The Size of the Seed Planted and the Fertility of the 

 Plant Produced," Amer. Breed. Mag., 3: 293-295, 1912. ^ ^ 



and size of plant produced is not merely the result of extra reserve food m 

 the larger seeds. 



4 The results of these and of other data from experiments made long 



