150 



CARNEGIE INSTITUTION OF WASHINGTON. 



a 1 : 1 ratio basis in families of different sizes. This formula has been 

 used in the calculations herewith appended (tables 15 and 16). 



Table 15. — Hemophilia. 



The difference between the expected and observed numbers of bleed- 

 ers is an excess of "observed, " which is 8.8 times its probable error. 



The data from pedigrees of color-blindness are much more meager, 

 but are interesting in support of the figures previously quoted in the 

 case of hemophilia. 



Table 16. — Color-blindness. 



The difference in this case is 3.1 times its probable error and is prob- 

 ably significant. 



Another indication of the presence of sex-linked lethal factors is to 

 be expected when the sex-ratio of families in which all the males are 

 either hemophilic or color-blind are contrasted with families in which 



'The ± value given represents the fluctuation in the actual numljer of hemophilic males, due 

 solely to chance. On the ba^s of this value the actual number observed may be compared 

 with the theoretical expectation as indicated. 



