MISCELLANEA. 
Local Death Rates. 
Would it not be worth while for ciii evokitionist statistician to give some attention to the 
mass of material accumulated in the Decennial Supplements to the reports of the Registrar- 
General for England and Wales ? These contain for each intercensal decade (1851 — 60, 1861 — 70, 
etc.) a series of tables giving the deaths from different causes and at successive age-groups for 
each of the 632 Registration Districts in England and Wales. The mean population during the 
decade at each age-group is also tabulated, so that the rates can be easily worked out. The 
correlations of death rates at different ages would, for instance, form a very interesting study. 
Thus the question suggests itself, e.g. are the childhood and adult death rates for different 
districts always positively correlated — i.e. should we in general expect to find a high adult rate 
where there is a high mortality in infancy and childhood l Very little inspection will show 
that the general death rates are thus positively correlated, but it is at least open to question 
whether death rates from specific causes are so ; I would instance the death rates from diseases 
of the nervous system. Should death rates from some causes show a much lower correlation 
than others, the question would arise whether the reductions might be due to the selectivity 
of the death rate ; were the death rate highly selective, a high infantile or childhood mortality 
might lead to a reduced adult mortality and so to an actuall}' negative correlation. It is doubtful 
however whether this would really occur : high death rates are in general due to bad local 
conditions of one sort or another, and it must be remembered that any selectivity of the death 
rate acting on the young may be counterbalanced by a corresponding weakening of the survivors 
due to these very conditions. It must also be borne in mind that death rates have changed with 
great rapidity in many parts of England, and that the adults now existing are the survivors of a 
much severer childhood mortality than the present. Unless, moreover, a careful selection be 
made and the rapidly growing urban districts taken by themselves, the influence of migi'ation 
may make itself felt. From the point of view of selection many difficulties might be avoided if a 
group of districts with little migration could be formed, and the change in childhood death rate 
between the two decades, say 1851 — 60, 1861 — 70, for each district compared with the change (in 
the corresponding age-groups) of adult death rate between 1871 — 80 and 1881 — 90. Were the 
childhood death rate markedly selective one would expect the districts exhibiting the greatest 
decreases of childhood death rate in the earlier period, to exhibit the smallest decreases of adult 
death rate in the later — i.e. the changes would be negatively correlated. Any investigation would 
certainly present great difficulties as to interpretation of results, but it would seem worth under- 
taking. 
G. U. YULE. 
