3!) 2 



Oh the Laws of Inhcritanve in Man 



(x) Gross Fratermd Reseinblance. 



I turu to tho rolatiiiiisliiiis betweoii dirtereut Organs in pairs üf siblings. These 

 are tabulatud below. 



TABLE XII. 



Correlation Coefficients für Cross Fraternal Heredity. 



The saine general remark inust again be inade here, i.e these cross-correlations 

 are romarkably high, — as high as a few years ago \ve anticipated tliat the direct 

 fraternal eorrelations would bi\ 



Tlie series being rathor shoit — thrce to t'our huudnd brolliers* — the results 

 are more irregulär tlian \ve niight have hopcd tor. In particnlar the cross- 

 correlation bet\s'ecn brother's stature and brother's forearm is distinctly less than \ve 

 might have expccted. A resiilt of slightly over ••l would clearly be niore consonant 

 witli the other results, but I have not beon able to discover any slip in the 

 arithmetic. In the brother-sister eorrelations we tind tliat within the limits of 

 the probable errors of ranilom sami)ling the cross-correlation coeftieients are pair 

 and pair e((ual, e.g. the relation of brother's stature to sister's forearm is sensibly 

 that of brother's forearm to sister's stature. 



To obtain au empirieal fonuula, I assumed that as in (vi) p. .SXö we should have 



7-,v = 6'(i-n'-i-/-+-'-2!'Vj) (ix). 



I deterniini'd (' l'roni the twelve series and found for its mean value ".^ö.SS. This 

 is within two per cent. of the value of C found for tiie cross-parental relationships, 

 aud I think the agroemeut is as close as we could hope fort. 



* See Appendix of correlation tabics. 



t To obtain tlie desirablo end of using onc formulii iuatead of two. I am not conviuced tbat eqnality 

 is a priori to be expccted. 



