298 MK. G. UDNY YULE ON THE ASSOCIATION 



older than fourteen years or so. For purposes of comparison, and in order to be able 

 to follow the association of defects from childhood to old age, I have made some 

 use of the material available in the Census.* This consists of the numbers of those 

 who are blind, deaf and dumb, or mentally deranged, and the numbers of those 

 suffering from combinations of these defects ; the sexes are separated and the 

 numbers in different age groups are given. The figures are very unreliable in the 

 early age groups (0 5, 5 10, 10 15), especially for "mental derangement,"! but 

 so far as I know they are the only figures of the kind published. I have not used 

 the age group 5 at all, and the others appear to give much the sort of associa- 

 tions one would expect, quite comparable with those given by the results of 

 Dr. WARNER'S investigations. 



53. The case to be treated is by no means a simple one. Suppose a certain 

 group of young children to be observed at some time and the frequencies of all com- 

 binations of certain defects noted ; let the survivors be again observed after a lapse 

 of years, and the defects be again noted. Changes in the relative frequencies and 

 the associations observed between defects will have taken place for three reasons 



( 1 ) Because some of those originally observed have died ; 



(2) Because some have outgrown or lost certain defects ; 



(3) Because others have acquired defects. 



We have taken the case as referring to children, but the first and third causes of 

 change are equally, or more, effective in the case of adults. 



Now if the observations were made as supposed, and a record kept of the same 

 individuals, the effect of each change could be distinguished. Those who had either 

 lost or acquired defects^ during the intervening period could be struck out of both 

 series ; the resulting changes would be due to selection only, and so on. But un- 

 fortunately this is not the case in any published statistics of which I am aware, even 

 in the " Childhood Committee's " work, where such a procedure might well have 

 been adopted. All that is given is that a certain group, closely centered round a 



* Census of 1891, vol. 3, Tables 15, 16, p. Ivii. ; 1881, vol. 3, Tables 14, 15, p. xlv. In the Census of 

 1871 the numbers with combinations of defects are not given, so the material is not available for the 

 present purpose. In the Census of 1881 those who are " Idiot or Imbecile " are distinguished from the 

 " Insane,"|both in the first and second order groups, and I have made some use of this (rf. below, p. 312). 

 In 1891 those " Mentally Deranged from Childhood " and " Blind from Childhood " are distinguished from 

 others, but the same distinction is not continued where these defects are combined with others. The 

 notation of the Census is not clear : by the courtesy of the Registrar-General I am informed that 

 "Blind" includes "Blind and Dumb," and so on; but "Blind and Dumb" does not include "Blind, 

 Dumb, and Mentally Deranged," i.e., the frequencies given are (A) (B) (C) (ABy) (A/3C) (aBC) (ABC). 

 This should be made clear in a future Census. 



t Cf. remarks in the Census Reports for 1881 : General Repdrt, p. 68. 



J I speak of " defects " as they form our subject matter, but the reasoning applies to attributes of uny 

 kind. I would define a defect, tentatively, as an attribute the possessors of which have a death-rate above 

 normal, but I do not know that this applies to all, e.g., of the defects noted by Dr. WARNER. 



