Ethel M. Eldbrton and Karl Pearson 
557 
later periods. Our actual correlations have all been found without grouping by 
the direct product-moment formula, but we give on pp. -550 and 558 two grouped 
tables to illustrate the difficulties which arise in analysis. Our first table is for 
the total numbers isolated and the total deaths registered. It will be seen at once 
that the marginal distributions are intensely skew, crowding up into the corner of 
few deaths and few cases isolated, so that they appear to asymptote to the zero 
values of the coordinates. Further, Diagram I shows that the regression curve 
Total isolated. 
0 250 500 750 1000 1250 1500 1750 2000 2250 2500 
Diagram I. 
of deaths on total number isolated is, if just sensibly, still not markedly skew. 
Turning to the actual numbers given by this table we have the following series 
of constants : 
Numbers Isolated (/) 
Mean « =475-33 
Standard Deviation ... ... o-^ =571'25 
Coefficient of Variation ... ... = 1-20 
( = S.D./Mean) 
Correlation Coefficient and Ratio rj-^= -8348*+ -0163 
Registered Deaths {d) 
d =103-42 
a-rf= 118-61 
Va= 1-15 
-8564 1 
* Agrees reasonably well with the non-grouped values for the two separate periods, 
t Found by taking means of all 13 column-arrays. 
71-2 
