In the preceding pages, an analysis has been made of the correlation coefficients found between the number of fatal cases from bronchitis, pneumonia and respiratory diseases, or the two summed together, in children under five years of age in London, Glasgow, Edinburgh, Aberdeen and Dundee, and the meteorological factors, mean and mean minimum temperature, mean relative humidity and mean rainfall, for a period of from 40 to 50 years; and between the death-rates from bronchitis, pneumonia and respiratory diseases amongst children at the same period of life in the registration counties of England and Wales in the decennia, 1901–10 and 1891–1900, and their mean annual temperature and mean annual rainfall and it seems warrantable to draw the following conclusions.

1. The meteorological factor, amongst those investigated, that seems to exercise the greatest influence in predisposing to an increased mortality from, and presumably an increased incidence of, bronchitis in children in the cities, is the prevailing temperature. In accordance with the average duration of the disease, the mortality is most intimately related to the mean temperature of the preceding week and is, on the average, as closely related to the temperature of two weeks before as it is to that of the corresponding week. The relationship is inverse, the lower the mean temperature the greater the fatality from bronchitis. As suggested by Dr Leonard Hill, the cold weather probably compels children to remain indoors in stuffy, overheated rooms where infection is intensified and health weakened by heat stagnation, a lowered metabolism and want of sunlight.

2. While there is a suggestion from some of the correlation coefficients found, that pneumonia in children may have some inverse relationship to mean temperature this association is not shown in the monthly or weekly data generally.

3. There is a definite inverse relationship between the mortality from the respiratory diseases, or bronchitis and pneumonia summed together, and the mean temperature; and, as occurs with bronchitis, the relationship to the mean temperature of the preceding week is closer than to that of the corresponding week.

4. The mortalities from bronchitis, pneumonia and respiratory diseases (bronchitis and pneumonia summed together), in children under five years in the registration counties of England and Wales, are inversely associated with the mean annual temperature in the corresponding counties. This association still persists after allowance is made for the proportion of each county which is urban, a proportion which has been taken as a rough index of industrial conditions therein.

5. The mortalities from bronchitis and pneumonia in children in the cities examined and in the registration counties are not influenced in any consistent manner or degree by the amount of rainfall.

6. While the coefficients of correlation between the corrected monthly deaths from bronchitis and respiratory diseases, respectively, and the corresponding mean monthly humidity in Glasgow seem to suggest the existence of a significant direct association between these variables, viz. the moister the atmosphere, the higher the death-rate from bronchitis; such a relationship is not definitely indicated in the coefficients found for the other cities. The magnitude of the coefficients for Glasgow is apparently determined, in some degree, by the periodicity in the mortality figures. It is probable, however, that the periodicity is not wholly responsible for the correlation found as the correlation coefficients between the mean temperature and the deviations in the same monthly data from the five-yearly moving average—the method adopted to eliminate the effect of the periodicity—are not very different from those found by using the actual figures. There would appear to be some evidence for the view that a high relative humidity, when associated with a low temperature, has some influence in predisposing to an increased mortality from the respiratory diseases.