M. Greenwood 
73 
If we consider in broad lines the general results of otir investigation, we should 
say that they may be summed up in the statement that both special diseases and 
the general want of health to be found in a hospital population tend in the same 
directions, namely, to increase the variability of the organs dealt with and to 
reduce their correlation. As we pass from the general hospital population to a 
healthy population we find that variability sinks and correlation rises. To what 
extent this is an antecedent or concomitant of the diseased state it might not be 
alwaj'S possible to assert. 
In taking any population, low variabilities and high correlations are the two 
factors which measure closeness to type. As a general rule, under a given environ- 
ment closeness to type is a condition of stability, we may almost say, of low selective 
death-rate. Hence we may look upon disease as less stringent approach to type, 
and high variability and low correlation as a sign of instability *. 
Of course the capacity to vary absolutely and to alter the relationship of organs 
must exist, or a race will not be able to effect a change in type with a changing 
environment. Still, if we trust the theory of correlation by natural selection at 
all, death before senility as far as it is selective is the destruction of the less fit, 
i.e. of those not approaching within certain limits the type suitable to the environ- 
ment. Thus it comes about that we shall expect on the Darwinian theory to find 
the individuals who die of disease in adult life to be more variable and less highly 
correlated in their organs than the " healthy." This is precisely what we do find, 
and the post-mortem room provides direct evidence in favour of the action of 
natural selection in the case of man. Indeed, in a not very conscious way the 
medical world has been expressing these very truths of evolution in other words. 
The figures we have considered showing lowered correlation in the diseased state 
are a biometric illustration of the truth of Dr Sutton's aphorism, " Disease is 
absence of rhythm -f-." In the normal or healthy group we see a population 
possessing the characteristic marks of stability, small variability, and high corre- 
lation. In the two special and the general diseased groups we have conditions 
tending in the opposite sense. In the healthy class we get a closer quantitative 
relationship between the weights of the viscera, in the diseased there is greater 
variety of proportions. Indeed, to adopt a well-worn definition, " Among the 
diseased each organ has a life and growth of its own, irrespective of the needs of 
the organism as a whole." Our biometric investigation shows us this independent 
life and growth leading to increased variability and to lessened correlation, shortly, 
to those deviations from type which beyond certain limits are incompatible with 
survival under a given environment j. 
* In the evolution of subspecies it seems probable that the hardiest and most prolific groups 
have the least coefficients of variation. Thus in an investigation recently made by Mr A. Bacot 
and the author on the variability of Sinlosoma Urticae, it was found that in several series of broods 
the groups with the largest coefficients of variation had the least net fertility. 
t Medical Pathology, 1886, p. 95 et seq. 
i I desire to take this opportunity of expressing my gratitude to Prof. Karl Pearson, to whose staff, 
among other acts of kindness, I owe the correction of many arithmetical slips in the above results. 
Anything of interest in this essay is due, either directly or indirectly, to him. 
Biometrika lu 10 
