A SECOND STUDY OF THE WEIGHT, VARIABILITY 
AND CORRELATION OF THE HUMAN VISCERA. 
By M. GREENWOOD, Junp,, and J. W. BROWN. 
(From the Statistical Laboratory of the Lister Institute.) 
In 1904, one of us published an account of certain applications of biometric 
methods to data obtained from the pathological department of the London 
Hospital*. The object of that essay was to throw light upon the mean values, 
variabilities and correlations of the human visceral organs, in the hope of (a) Placing 
on record results comparable with the numerous studies of anthropologists which 
deal with other organs or dimensions : (b) Providing materials for the study of the 
changes produced by various diseases in the absolute measures and inter-relations 
of the organs observed and, therefore, p - o tanto, illustrating the work of selection 
upon the human species. 
In the paper cited, attention was directed to the specialised characters of a 
" General Hospital Population " and it was pointed out that results obtained from 
such a population could not be generalised without inquiry as applicable to the 
race at large. It was also shown that groups selected upon various principles 
within the " General Hospital Population " exhibited marked deviations from the 
prevalent type, a selection of the more healthy organs being attended by a rise in 
correlation and a fall in variation, while a group characterised by a preponderance 
of specifically diseased organs exhibited changes in the opposite sense. 
The difficulty which had to be faced in that investigation was the task of 
ascertaining the characteristics of a healthy " General Hospital Population." By 
this expression we define the class of pei'sons of the physical, racial and economic 
type to which hospital patients conform but who are not suffering from any disease 
likely to cause death or seriously affect health. The nature of the material avail- 
able in 1904 did not permit of such a selection being made and all that could be 
done was to choose from the data those organs which, so far as the records showed, 
were not obviously diseased. Such cases formed the normal or " healthy " material 
Biometrika, Vol. in. p. 65. 
