clx Tables for Statisticians and Biometricians [XXXII XXXIV 



On the accompanying diagram the curve plotted from these ordinates is given 

 and compared with the ordinates of the curve 



, 2-0917 



1 - 



.17-6838 



33-03207 

 which is the curve fitted to the same data, i.e. 



Mean = 22-8361, Standard Deviation = 13'5078, 



1 = '6783, fa = 3-7342, 



the equation to the curve, however, being expressed in five actual units as working 

 unit. 



ORDINATES OF (p,H ) CURVE . COHPARED WITH 

 PEARSON CURVE FITTED FROM FIRST FOUR rTOPTENT? . 



40 



-IO 



O iO 2O 3O 4-O JO 00 



Percentage of Black i Skin-Colour 



iOO 



The process is considerably shorter than computing an adequate series of ordin- 

 ates from the above curve. It illustrates the principle that if two distributions have 

 the same fti, /3% the curves deduced from these will closely accord, if we superpose 

 their means, and equalise their standard deviations. 



The result in the previous section shows a very good correspondence between 

 the Pearson curve and an r-curve of the same first four moments (Mean, S.D., /3i and 

 /3 2 the same), even for n as low as 13. This is certainly true as the value of n in- 

 creases. For n = 25, p = 0'6 the accompanying diagram (p. clxi), in which the true 

 ordinates are represented by dots and the continuous curve is the Pearson curve : 



/ . \ 57536 / \178-5135 



9-64157/ 



