Modelling the incidence of virus yellows in sugar beet in the UK in relation to numbers of migrating Myzus persicae

1, The incidence of virus yellows diseases in sugar beet crops in eastern E ngland during 1965-96 was analysed in relation to numbers of migrating Myzu s persicae, the principal vector. 2, A non-linear model was fitted to the data incorporating dual routes of i nfection: primary infection. arising from winged immigrant aphids carrying the virus, and secondary infection, arising from local dispersion of their wingless offspring transferring the virus from infected to uninfected plant s. 3. The model was fitted to the data assuming that (i) the epidemic commence d when the first migrating aphid was recorded, (ii) the rate of primary inf ection was dependent on the total number of aphids migrating up to the end of June, and (iii) the rate of secondary infection was dependent on the pro portion of diseased plants in the crop. 4, A good description of virus yellows incidence in August depended on allo wing the rate parameters for primary and secondary infection to decrease si gmoidally over the years during 1965-96. The shape of this function depicte d a major change in the epidemiology of virus yellows since 1974-76. Amongs t several contributory factors, this change coincided with the introduction and subsequent wide use of systemic granular pesticides. 5, Good fits also depended on allowing the rates of disease progress to dec ay with time to accommodate effects of increasing host resistance with plan t age to feeding aphids. This meant that early epidemics rose faster than l ate epidemics, which is consistent with general observations. 6, Given that the changes in the epidemiology of virus yellows over the yea rs are associated with improvements in pest management practices, the model presents a useful extension to disease forecasting by providing prediction s of disease risk in the absence of pesticides. Such predictions indicate t hat during the last decade up to five severe epidemics would have been like ly.