K. Pearson 
183 
Lastly if we consider that the question of health determines the skewness, we 
have in the dotted ciu've the weight-distribution of G99 hearts stated to be 
"healthy." We see that there is still the same essential skewness; the pathologist 
has merely cut off a small portion of the tail on tlie left and tar too much of the tail 
on the right, i.e. unusually big hearts were discarded as necessarily " unhealthy." 
The form of the curve undoubtedly indicates that many of these large hearts are 
abnormal, but any continuous curve fitted to the remainder, the " healthy hearts," 
would not only be significantly skew, but would project a long way into the 
portion of the tail discarded as " unhealthy." The list of asymmetrical distributions 
might be indefinitely extended, but these must sutfice to indicate that asymmetry 
cannot be lightly put on one side in the manner adopted by Ranke and Greiner. 
380 
30O 
340 
\ 
\ 
ato 
too 
leo 
160 
110 
130 
\ — 
j 
\ 
100 
1/ 
■A 
80 
£0 
-if— 
' ) 
// 
i 
«0 
30 
-■LI a 
— - 
— - 

, 
»fl ^ ^ ^ O 5 io »0 5 ^ ~, 1^ 5 to O «5 io U3 5 m iQ « « lO tO iD 
'°'"'°°'23a2S22r:22ogagsg»5aD»05j.»sgg 
Weifiht of Hi'in t in Kiiglisli oz^. 
Fk;. 1. Frequency Polyfjons of Weight of Heart in Males (Greenwood*). 
I. * -* 25 to -5.5 years. Witliout specific disease of heart. Number of cases 1382. 
II. O O 25 to 35 years. Without specific disease of heart. Number of cases 358. 
III. • • 25 to 55 years. Definitely sound hearts. Number of cases 66!l. 
The scale is four times as great for II. and twice as great for III. as for I. 
If we pass to discrete variates, we find as large a number, if not a larger 
number of distributions in which skewness is well marked, for example, fertility 
in the Aphis Hyaloj)terus Trirhodusf, fertility in man|, fecundity in race-horses§, 
and fertility and fecundity in mammals generally. I illustrate this with an 
example of fertility in English mothers in Fig. 2. It will be seen at once that 
no normal curve could be used to describe this distribution. It is equally 
* Biometrika, Vol. in. p. 45 et seq. 
t Warren : Biometrika, Vol. i. p. 127. 
X Pearson : Phil. Tram. Vol. 192 A, p. 257, and The Chances of Death, Vol. i. p. 63. 
§ Pearson: Biometrika, Vol. i. p. 292. 
