K. Pearson 
155 
coefficients of fraternal resemblance in man. I do not believe it to be at all 
comparable with the irregularities that arise from random sampling and occasional 
carelessness in measurement or in appreciation of character. 
Now turn to Table IV. giving the degree of resemblance in the mental and 
moral characters. What do we find in it ? Perhaps slightly more irregularity in the 
TABLE IV. 
Inheritance of the Mental Characteristics. 
School Observations on Children. 
Character. 
Correlation. 
Brothers. 
Sisters. 
Brother and Sister. 
Vivacity 
•47 
•43 
•49 
Assertiveness ... 
•53 
■44 
■52 
Introspection ... 
•59 
•47 
•63 
Popularity 
•50 
•57 
•49 
Conscientiousness 
•59 
•G4 
•63 
Temper 
•51 
•49 
•51 
Ability 
•4G 
•47 
•44 
Handwriting ... 
•53 
•56 
•48 
Mean 
•52 
•51 
•52 
values than in the case of the physical characters. The judgment required is much 
finer; and the classification is much rougher. Let me frankly admit the difficulties 
of the task, both for observers and computers. I will lay no weight whatever, if 
you like, on the second place of decimals. But what is the obvious conclusion ? 
Why, that the values of the coefficient again cluster round '5. If anythiug the 
average degree of resemblance for the psychical is rather less than for the physical, 
it certainly is not greater. Personally I would lay not a grain's weight on the 
difference. 
I have illustrated the whole result in Diagram XIII. The two lines 
representing physical and psychical qualities go bobbing up and down, and cutting 
and re-cutting one another. No wise man, however, would venture to assert that 
one or other is sensibly uppermost, or that any of those rises or falls have real 
significance. We are forced absolutely to the conclusion that the degree of 
20—2 
