Although the national statistics of the notifications of infectious diseases are defective with regard to age and sex, it has been established from the returns of London and other large towns that the mean age of attack from diphtheria has been increasing during the present century. The age shift has been from the pre-school to the school age. Several investigators have commented on the change in age incidence, Murphy (1907) reviewed the trend of mortality in London, and Chalmers (1913) the mortality of Scotland. Cheeseman, Martin & Russell (1939), as a result of an analysis of data kindly provided by the Medical Department of the London County Council, concluded that the change might be due to the decrease in size of family, especially in the more crowded districts. This would Jead to a decrease of morbidity at early ages and, as a consequence, a lower level of herd immunity at school ages, so that the epidemiological experience of the poorer districts would approximate to those of the middle-class districts. In this paper (pp. 188–9) reference was made to Turner's observation (1923) that both in London and Manchester the mean age of attack in epidemic years of scarlet fever was above the average, and it was shown that if the maximum weekly case rate of diphtheria in London (1897–1930) were correlated with the ratio of the number of cases at ages 0–4 to the number at 5–9, a significant negative value was reached. This finding was consistent with Turner's result, but it seemed of interest to apply a slightly more direct test by comparing years of heavy prevalence with adjacent years in which diphtheria was at a lower level.