Karl Pearson and David Heron 
313 
that according to that law there might be in rare instances such a combination of 
circumstances that a normal child might be born from two parents that function as 
feeble-minded. For practical purposes it is, of course, pretty clear that it is safe 
to assume that two feeble-minded parents will never have anything but feeble- 
minded children*." 
The italics are ours. They are very typical of the manner in which an elastic 
record and a plastic theory are made to fit. No account is anywhere provided 
of this extension of Mendelian theory which "in rare instances" provides a normal 
offspring to the two parents with an abnormal dominant character. Possibly, as 
in Dr Pearl's case, it is due to the occurrence of " physiologically extremely 
favourable " matings. Anyhow the old definite simplicity of Mendel's Mendelism 
has gone ; with an elastic record and a plastic theory any data may be Mendelian 
— or not — according to the views of the investigator who moulds his theory and 
stretches his facts. 
How welcome to such an one must be the Yulean theory of association ! 
" Whatever the nature of the classification, however, natural or artificial, definite or 
uncertain, the final judgment must be decisive; any one object or individual must 
be held either to possess the given attribute or not" (Yule, Theory of Statistics, 
p. 9). 
In the face of such a direction, how could Dr Pearl have been so foolish as to 
balance a number of his hens on the dichotomic fence of 30-eggs, and allow a 
moiety of each such hen to possess one Mendelian attribute and the other moiety 
its alternative ? 
It is not difficult to understand, however, why Dr Pearl does not like Pearson's 
criterion of the goodness of fit of theory and observation. On p. 255 of his 
memoir he gives a table " showing the observed and expected distributions of 
winter egg production for all matings taken together." He remarks on this 
table that " the lumped figures do not give an altogether fair estimate of the 
matter, but some sort of a summary is necessary." We agree very cordially 
because a number of the impossible green balls of the subsidiary tables do not 
appear as such when the tables are lumped, but taking the data for what they 
are worth we have the following four series : 
Winter Egg Production 
Over 30 
"Under SO 
Zero 
I. 
Observation ... 
365 5 
259-5 
31 
Theory 
381-45 
257-25 
17-30 
II. 
Observation ... 
2 
23 
15 
Theory 
0 
25 
15 
III. 
Observation ... 
36 
79 
8 
Theory 
26-5 
86-75 
9-75 
IV. 
Observation ... 
57-5 
98-5 
23 
Theory 
68-60 
95-0 
15-40 
* The Kallikak Family, p. 114. 
Biometrika ix 40 
