On superinfection in malaria.

The task of proving or disproving that there is superinfection of human beings in malaria has been unresolved since the concept of superinfection was introduced by Macdonald (1950). We present a method of analysis the results of which indicate that there is superinfection in children up to the age of about five years, but not in older human hosts. We use maximum-likelihood estimation in an extension of ideas introduced by Bekessy, Molineaux, & Storey (1976). The estimation is based on a model for the infection of human hosts with variable degree of superinfection. This model contains a parameter S that equals the maximum number of infections in the host. In the absence of superinfection, S = 1, while S = infinity in the unlimited superinfection assumed by Macdonald. The estimation of S in a number of age bands allows us to identify the age bands where superinfection occurs.