316 



MR. G. UDNY YULE ON THE ASSOCIATION 



TABLE X. Showing the number of Cases in which the Association between Defects is 

 greater for Males and greater for Females (based on preceding tables). 



* Total associations only. 



associations are greater for the young than the old. I have not, however, been able 

 to say with certainty whether the decrease in association was due to growth (change 

 in the individual) aided by selection, or growth opposed by selection. The only 

 available evidence suggested that selection acted in the same direction as growth. 



The total associations between defects were always high, except in some cases for 

 the very old, but the partial associations for defective universes were in most cases 

 low, and in some cases even certainly negative. The partial associations for un- 

 defective universes were, on the other hand, higher than the total coefficients, a fact 

 which I held to imply that no one of the defects recorded was necessary as a connecting 

 link between any other pair. 



70. During the course of the present investigation I have naturally been led 

 to study pretty thoroughly the " Report " of the Committee on Childhood, and the 

 papers by Dr. WARNER, bearing on the same subject, in the ' Journal of the Royal 

 Statistical Society ' papers which first drew my attention to the need of a 

 theoretical study of the whole subject of association. I may be pardoned, then, for 

 offering some suggestions as to the work of the Committee as regards both the mode 



