388 



Oti tlie Laws of Inhrritanre in Man 



find good agrecmcnt in thi' 'ö valiie. In tho Table below each serics involves 

 1000 to 2000 cases. 



TABLE X. 

 Fraternal Reseinblance in Boys at School. 



Mean of Eifjht Characters •539. 



Tliis cüinpares well wiih •öl 7 ut ihc ubuve Family Kocords for Brothers. At 

 the same time several of the abovo results are under revision, uamely the absolute 

 niejisurcments of the head. In this ca.se ever}' boy wjus reduced to a Standard age 

 12 by adding to or subtracting froni his age the average growth found to takc place 

 between his age at measurement and the age of 12. A more subtlc method of 

 detennining the allowance to be raade for growth has been recently given by me 

 in a paper published this ycar, only it involves a vory great ainount of nunierical 

 work, i.e. fivi> instead of two correlatiun tables, and thus we have only been able 

 so far to iniHlify the correlation in the matter of head length*. I consider it 

 better, howevcr, to place above the correlatiuns a.s found by one uniform method 

 until \ve are in positiou to publish all the results revised. I auticipate that both 

 head breadth and auricular height will ultimately be found to be really nearer to 

 ".5 than apj)ears above. Meanwhile it seeins (|uitc safe to siim up onr results for 

 fraternal correlation in man as follows : 



(i) The degree of resemblance of brethren is clo.sely the same for all 

 characters. 



(ii) The twi) sc.xes appear to be e(]ually iiitlnenced by hercdity. 



(iii) The inteusity of fraternal ((irrrlat inii in man is close to h, possibly 

 slightly great er. But for praetiral purposes \ve may convenieutly work with ^ as 

 a rounil niimber. 



I now turn to what personally I consider one of the raost obscure points in the 

 quantitative determination of inheritance, namely : the manner in which fraternal 

 resemblance varies fi'om sj)ecies to species, while paternal inheritance reniains fairly 

 constant. If we look at Table V. we .see that witliin moderate liniits j)arcntal 

 induence appro.\imates to the same value for vory different species and very 

 diverse characters. This cannot be asserted with the same accuracy of fraternal 

 correlation. I have found values of it ranging Crom ^4 to '7 for largo and 



* R. S. Proc. Vol. 71, pp. 290—4. 



