270 Mr Punnett, 



From this table it will be seen that Set 1 (with birth dates) 

 differs from Set 2 (without birth dates) in shewing a greater 

 proportion of ,-fs and a much smaller average number per family. 

 There would seem to be some relation between the large families 

 with their greater proportion of $ s in Set 2, and the unwilling- 

 ness shewn here to return the dates of birth of the children. 

 From what has gone before we should expect an increase in the 

 size of the family to be correlated with an increased relative 

 proportion of $s. This is actually the case, though to a greater 

 extent than we should have been led to look for. If we possessed 

 the birth dates in Set 2 it is not improbable that they might 

 reduce somewhat the relative preponderance of Js among the 

 first born, though it does not seem possible that such an effect 

 would be very marked. For although $s tend to preponderate 

 more in large families than in small ones, yet the proportion of 

 jfs is just as great in the first children of large families as 

 of small. In 197 families taken from Set 1 and containing 6 or 

 more children the first born was in 117 cases a £ and in 80 cases 

 only a $, a proportion of ^s (146 per 100 $s), which tallies 

 closely with the general result given in Table VI. 



As this feature of a more marked preponderance of </s for 

 the first birth in a family is of considerable importance it may 

 be worth while digressing for a moment in order to consider 

 the evidence in favour of it. The phenomenon was familiar to 

 Diising 1 , nearly twenty years ago when he stated that " die 

 Erstgeburten liberhaupt zeigen einen relativ grossen Knaben- 

 iiberschuss," and gave statistics in support of his statement. More 

 recently Orchansky 2 has brought forward additional evidence. He 

 distinguishes two types of family — Type I where the first born 

 is a </, and Type II where it is a %. From the data given by 

 him I have been able to calculate that the proportion of ^s to 

 $s among first born (on a total of 2442 births) is 104*6 : 100, 

 whilst the proportion of ^s to % s for total births is 101"1 : 100. 

 The excess of g first births though quite distinct is not so well 

 marked here owing probably to the heterogeneity 3 of the author's 

 material. Such additional evidence as I have been able to obtain 

 upon the subject all points in the same direction. In a total of 

 6938 first births from the records of Queen Charlotte's Lying-in 



1 C. Diising, Jen. Zeit. 1884, p. 608. 



2 J. Orchansky, Die Vererbung im gesunden und krankhaften Zustande u. s. w. 

 (Stuttgart, 1903), p. 103. 



3 Orchansky's material contained amongst others 542 Jewish families in which 

 the proportion of J to ? first births was only 81 "9 : 100. If we consider the 

 remainder of his material apart — 1900 non-Jewish families— the proportion of <? to 



? first births is 11T8 : 100, the proportion of total <? births to total ? births re- 

 maining nearly the same. This behaviour of Jewish infants is somewhat remark- 

 able and is not supported by Diising's evidence {Jen. Zeit. 1884, p. 625). 



