igi3] /. Arthur Harris and Ross Alken Gortner 529 



two are of the order r = — 0.300. Thus, of the four series of data 

 which can be scientifically analyzed, three give correlations which are 

 negative in sign while one is positive. One negative and one positive 

 coefficient are very low; two of the negative coefficients are more 

 substantial, agreeing fairly well with the values previously pub- 

 lished by us. 



Surely such facts as these form a very slender basis for the con- 

 clusion (widely circulated by uncritical reviewers) that in beets of 

 the same strain there is no negative correlation between weight and 

 sugar content ! Nevertheless one must recognize the possibility of 

 the correctness of the conclusion. Should it prove to be valid, the 

 Suggestion follows that the negative correlation demonstrated in 

 commercial cultures has a genetic origin, i. e., that strains character- 

 ized by large root size are also characterized by low sugar content, 

 and that when these strains are intermingled and intercrossed in 

 field cultures there results a negative correlation between the weight 

 of the individual beet and the sugar content of its juice. Such a 

 result would be of the greatest interest to breeders. 



