274 



J. ARTHUR HARRIS 



lower to the higher grades the time required for its germination 

 also becomes longer. 



The evidences for this conclusion are (a) the significantly larger 

 number of experiments which give positive correlations, (b), the 

 numerically larger mean values of the positive correlations, and 

 (c), the greater trustworthiness (as indicated by probable errors) 

 of the positive coefficients. 



The fact that both positive and negative values of the coeffi- 

 cients occur may tend to shake the confidences of the reader in 

 the generality of the conclusions, even for the varieties dealt 

 with. Of course, it is too early to say that certain conditions 

 of growth or treatment of seed may not regularly result in a 

 negative correlation between weight and time, but it seems much 

 more likely that the negative correlations are simply the extremes 

 of the kind of fluctuations invariably seen in any series of experi- 

 ments involving the measurement of delicate relationships and 

 in which external conditions can be but grossly controlled. Their 

 occurrence is indeed no more remarkable than the fact that no 

 two records in a series of transpiration readings agree exactly. 



Finally, I must caution the reader against the extension of 

 the conclusions here drawn beyond the materials — taxonomically 

 or physiologically limited — of this paper. There are strong sug- 

 gestions that some varieties show a much higher correlation 

 between the two variables under consideration than do others. 

 It is also quite possible that the conditions under which the 

 seeds were grown or the age of the seed have measurable influence 

 upon the intensity of the relationship. Thus while there can 

 be little doubt that the condition here demonstrated is the gen- 

 eral rule in Phaseolus vulgaris, it is quite possible that some 

 series might be found giving results at variance with our own. 



The data here given are so keyed that in conjunction with 

 other published papers on Phaseolus, it is possible to consider 

 them with regard to all these points, but in view of the fact 

 that materials for other experiments are being gathered more 

 detailed analysis seems superfluous. 



In the meantime, the contents of this paper stand as recon- 

 naissance findings, merely. Their presentation is justified by 

 the clearness and definiteness of the results. 



