300 



The Journal of Heredity 



sisters) compared with the number of 

 them in each family. 



This table (I) deals with 2,964 indivi- 

 duals. If a random sample of a hun- 

 dred be taken, and the career of each 

 one followed, their history will be repre- 

 sented by the graph in Fig. 6. The 

 very large and \x^ry small families are at 

 the sides, the middle-sized families in 

 the middle of the figure, and it is evident 

 that the larger and smaller families 

 had a relatively large death rate in 

 childhood and early youth and a smaller 

 proportion of members living to old age. 



"only child" carries handicap 



In the forty-one cases set down as 

 "only child" a majority (58.5%) died 

 young. If the families containing only 

 two children be added, it is found that 

 no less than 47.6% of the total (126) 

 died before reaching maturity, and onl\- 

 4.8% lived to be over eighty. 



There were 168 persons belonging to 

 very large families containing thirteen 

 or more children. No less than 46.4% 

 of these per.sons died in childhood (under 

 twenty) and 5.9^/'( lived to be over 

 eighty. 



It is evident that in this case both too 

 few and too many brothers and sisters 

 were unfavorable to long survival. 



The proportion who died young was 

 least, and the pro]jortion who lived to 

 be old was greatest, where the ]jersons 

 belonged to families containing nine and 



ten children. There were 683 such per- 

 sons; 32.8% of them died young, and 

 this is the smallest jiroportion dying 

 under twenty of any of the groups shown 

 in Table I. ' On the other hand, 9.7% 

 of these persons lived to be over eighty; 

 and this is the largest proportion living 

 to old age in any of the grovips shown in 

 Table I. 



In short, the ]jro]jortion of children 

 who lived to old age increased with the 

 size of family up to ten children, and 

 beyond that it fell. There thus seemed 

 to be a limit to the size of family con- 

 sistent with the production of long-lived 

 ofTs])ring, but this limit is very much 

 higher than ijojjular ideas would lead one 

 to believe. Certainly ten children con- 

 stitute a pretty sizable family; even 

 the most enthusiastic eugenist is not 

 likely to ask suj^erior mothers to have 

 as many offsj^ring as that. Neverthe- 

 less the child with nine brothers and 

 sisters has (statistically speaking) just 

 about twice as good a chance of living 

 to old age as has the child with only a 

 single brother or sister, in a normal, 

 healthy population. 



But if the analysis of the figures is 

 carried a ste]j further, an even more 

 striking result ai)pears. It is evident 

 from Fig. 6 that the greatest difference 

 between the families of various sizes 

 is in the amount of child mortality. 

 It is im])ortant to know how the various 

 sizes of family will rank, if differences in 



T.MJl.li I. 



-Duration of Life of Individual Compared with the Number of His Sihs; Percentages 

 Who Died at Ages Named 



