ON VAEIATIONS. 347 



The proof given by Professor Pearson ^ of the exist- 

 ence of a selective death rate in the case of man seems 

 to me scarcely to entitle him to speak of it as a case of 

 ' Natural Selection." Thus he shows that there is a 

 fairly close correlation (r = .26) between the ages at 

 death of brother and brother, and a less close one be- 

 tween those of fathers and sons (r = .12 to .14). For 

 instance, the mean age at death of men not dying as 

 minors is 61 years. If, however, one brother of a pair 

 dies at 25, then the other will, on an average, die at 

 51.6, or 9.4 years earlier than the mean; if one dies at 

 85, then the other will, on an average, die at 67.2, or 6.2 

 years later. There is something in the constitution of 

 a man, therefore, which to a large extent determines 

 when he shall die, or undergo elimination. His death 

 is not at all a matter of chance. Further analyses of 

 data by Miss Beeton and Professor Pearson f indicate 

 the same thing, though they also lead to other and some- 

 what unexpected conclusions. Thus, from the pedigree 

 records of members of the society of Friends, the au- 

 thors found that elder (adult) brothers and sisters on an 

 average live distinctly longer than younger (adult) 

 brothers and sisters, and that the greater the difference 

 in age, the greater the difference in expectation of life. 

 For instance, a man born 6 years after his elder brother 

 will probably live 4 years less than he will; one born 

 10 years after, 7 years less, and one born 17 years after, 

 as much as 12 years less. Put in other words, the 

 eldest children of a family have the best chance of life, 



*" Grammar of Science," p. 497; also Beeton and Pearson, Proc. 

 Roy. Soc., Ixv. p. 290, 1899. 

 f Biometrika, i. p. 50, 1901. 



