CORRELATION OF BRAIN AND SPINAL CORD. 511 



number of variables dealt with. The test for linearity is made 

 through a determination of the correlation ratio and a comparison 

 of this with the coefficient of correlation r. The methods used 

 for the determination of the coefficient of correlation, the corre- 

 lation ratio and the partial correlation coefficient are those 

 described by Pearl (4). 



The raw data and the correlation tables which serve as the 

 basis of this study are not given because of space limitations. 

 They are on file at The Wistar Institute. 



In Table I. are given the zero order coefficients of correlation 

 (ro), the correlation ratios (77), the values for (tf -- r 2 ), the 

 probable errors for r and i", and the quotients r/E r and f / f of 

 the several comparisons made in this study. In addition to these 

 values there is to be found in the literature the following figures: 

 Donaldson (5), (6) in a series of rats of scattered ages and with 

 the sexes combined found the correlation between body weight 

 and brain weight to be 0.764: the correlation between body 

 weight and cord weight to be 0.856: between body length and 

 brain weight to be 0.86: between body length and cord weight 

 to be 0.99. Hatai (7) records a value of 0.516 in the male and of 

 0.692 in the female for the coefficient of correlation between 

 body weight and cranial capacity in the mature albino rat. 

 Jackson's (8) values are based on too few observations to be 

 any more than suggestive. For adult man, Boas (9) records a 

 correlation between body weight and head length of 0.43 in the 

 male and 0.41 in the female; and between body weight and head 

 width of 0.32 in the male and 0.33 in the female. Blakeman 

 (10) found the coefficient of correlation between body length and 

 brain weight to be 0.289 in the male and 0.367 in the female; 

 while Pearl (n) got values ranging from 0.166 to 0.183 in the 

 male and from 0.183 to 0.349 in the female for this association, 

 and 0.167 in the male and 0.226 in the female for the degree of 

 association between body weight and brain weight. 



The values for man are of little help in the problem because the 

 raw data from which they were derived are complicated by a 

 multiplicity of interfering variables which vitiate any comparison 

 with the figures obtained from the more uniform population of 

 albino rats. 



