Ixx TcM*for8tati$tici<in*an<l Biomrtrician* [XXXV XLVI 



with origin at the mode = 16 - 1486 iu actual percentages, ~ 3 - 2297 in working 

 units. To calculate // we take the origin at and have 



logy = 26-134,8705 + 2-0917 log (*+ -6774)+ 17'6838 log (3G 2017 -<r), 

 \vln'iv ./ ni:iv In- put 0, 1, 2, 3, 4, 5... working units c<>rrrs|><>ii<ling to 0, 5, 10, I " 

 20,25... actual percentages. The curve is shewn in the acrmn|>anying diagram, 

 and considering the nature of the data is a reasonable graduation. 



10 15 20 25 30 35 40 45 5O 55 60 65 70 75 80 

 Per cent, of Black in Skin Colour. 



TABLE XLVIII (pp. 8997). 



Percentage frequency of Occurrences in a Second Sample of m after p Occur- 

 rences in a First Sample n. (M. Greenwood, Biometrika, Vol. IX. pp. 69 90.) 



If we assume the tnith of Bayes' Theorem then an event having occurred 

 p times and failed q times in n trials, the chance that it will occur s times and fail 

 m s times in a second series of m trials is : 



/I 



I x(\-xyidx 



Jo 



These results can be evaluated as all the indices are integers and the series 

 C,+ C, + C t + ... + C t + ... expressed iu the usual hypergeometrical form : 



m (m - 1) (m - 2) 



\q \n + m + 



n+l [i + P + 1 + ^_(^il) (P 

 n+l\ ll^y + TO (2 (q + 



5T+ UX). 



