Pearl, Correlation in Brain- Weight. 477 



More and better material for determining the relation be- 

 tween brain-weight and skull characters is much to be desired. 



9. The Weight of the Cerebrum. The Boyd-Marshall Eng- 

 glish material placed at my disposal by Professor Donaldson 

 made it possible to determine the variation in this character, 

 and its correlation with age and stature. The cerebral hemis- 

 pheres are markedly more variable in weight than is the entire 

 encephalon. This character is slightly less closely correlated 

 with stature and age than is the weight of the entire encephalon. 



ID. Characteristic Equations. Exact comparison of the 

 brain-weight in different races has hitherto been impossible on 

 account of the lack of precise methods of allowing for the influ- 

 ences of age, stature, sex and other bodily differences. A 

 knowledge of the degree of correlation existing between brain - 

 weight and these other characters at cnce places in our hands 

 the data through which such exact comparisons can be made. 

 With a knowledge of the coefficients of correlation and the 

 standard deviations we can form so-called "characteristic equa- 

 tions" which will give the probable brain-weight for arrays of 

 individuals of given types of the various other characters. 

 Such equations, besides enabling us to predict probable brain- 

 weight from a knowledge of other characters, also make clear 

 the relative influences of these characters on the weight of 

 the brain. 



I have put together here all the characteristic equations 

 calculated in the course of the work. In these equations the 

 letters have the following significance : 



W = probable brain-weiglit in grams. 



A = known age in years for the period 15 (or 20) to So.* 



« = " " " " " " " 50-' 



S = stature in centimetres, for individuals between 15 (or 20) and 80 years. ^ 



j. __ <i i< 14 .1 44 44 41 44 44 2ci\d. 50 years. "^ 



W^^weight of body in kilograms 



L = length of skull in millimetres 



B = breadth " " " 



.2 ^ standard deviation (in grams) of the array having W as the mean. 



'The equations containing the letters a and s are based on the "young" ser' 

 ies, and those containing A and S on the "total" series. 



