Ethel M. Elderton and Karl Pearson 551 
(iii) Are the districts which have adopted most isolation really urban 
districts where isolation was easiest to adopt and where possibly economic or 
social conditions favoured the spread of the disease or, in the case of the death- 
rate, the disease encountered a less resistant population ? 
(iv) What evidence is there to show that the districts which have rapidly 
increased their isolation-rates have subsequently lower attack- or death-rates ? 
If no one of these problems can be fully answered, — even in the case of a single 
disease — with the data at present available, at least light can be thrown on the lines 
which their solution in the future must take; and further something can be done 
to prevent hasty generalisation and excessive dogmatism as to the advantages or 
disadvantages of the isolation system. It can never be too strongly insisted upon, 
because it is so often forgotten, that preventive medicine is essentially an 
experimental science, and that in nine cases out of ten the efficiency of any line 
of action can only be adequately tested by statistics and by statistics collected after 
the expenditure of many thousand pounds, possibly spread over a long period of 
years, in carrying out this line of action *. 
(2) Material. In endeavouring to throw some light on the above problems we 
have fortunately received data of very considerable value from Dr E. H. Snell, the 
Medical Officer of Health for the City of Coventry. He obtained for a period of 
nine years, 1904-1912 inclusive, for about eighty towns or districts of large popula- 
tion but of very varying local conditions, (i) the annual number of diphtheria cases, 
(ii) the number removed to hospital, (iii) the number of deaths. We have added 
to this material the estimated population of the town or district, and further 
certain data as to the economic and social conditions. Unfortuna.tely there is no 
existing adequate measure of the general sanitary condition of individual towns, 
although the construction of a general sanitary " index number " would be of 
remarkable value in many forms of inquiry. We took as our measures of social 
condition : 
(a) Death-rate of infants under a year. 
(b) Amount of overcrowding, that is to say the percentage of the population 
in private families living more than two in a room. 
(c) Density of population, i.e. the number of persons to the acre. 
* Assert that it is most desirable to test the effect of sanatoria and of tuberculin in cases of 
tuberculosis, but do not dogmatically proclaim them as " cures " for phthisis, until statistics have been 
collected in sufficient amount and have been adequately and dispassionately examined to prove or 
disprove your statements. Insist on compulsory inoculation for enteric in the case of all recruits, but 
do not make it optional and then publish letters in the newspapers giving perfectly idle statistics, or 
go round to the camps giving popular lantern lectures to the recruits showing the gravestones of 
uninoculated persons, the portraits of persons dying of enteric, or much enlarged pictures of bacilli ! 
If you think it experimentally worth doing, inoculate ; but don't bring inoculation about by emphasising 
the dread of pain or the fear of death, both of which it is the first essential for a soldier wholly to 
disregard. 
