a1 School of Biological Sciences, Derby Building, University of Liverpool, P.O. Box 147, Liverpool L69 3BX, UK
a2 Control Systems Centre, Department of Electrical Engineering and Electronics, UMIST, P.O. Box 88, Manchester, M60 1QD, UK
There was a marked rise in scarlet fever mortality in England and Wales in the mid-nineteenth century and spectral analysis of the registration details, 1847–80, shows that the interepidemic interval was 5–6 years, but after 1880 the endemic level fell and the fatal epidemics disappeared. The dynamics of the scarlet fever epidemics can be represented by a linearized mathematical model and because the system is lightly damped, it could be driven by an oscillation in susceptibility. Epidemics were significantly correlated with dry conditions in spring/summer (P < 0·001), suggesting that these produced a low amplitude oscillation in susceptibility which drove the system. Epidemics also correlated (P < 0·001) with an oscillation in wheat prices but at a lag of 3 years, suggesting that malnutrition during pregnancy caused increased susceptibility in the subsequent children which interacted synergistically with seasonal dry conditions. Scarlet fever mortality was sharply reduced after 1880 in parallel with falling wheat prices suggesting that the remarkable period of high scarlet fever mortality (1840–80) was dependent on poor nutritive levels during that time.
(Accepted June 25 1996)