A simple numerical procedure is presented for the problem of estimating the parameters of models for the distribution of eggs oviposited in a host. The modelling is extended to incorporate both host density and time dependence to produce a remarkably parsimonious structure with only seven parameters to describe a data set of over 3,000 observations. This is further refined using a mixed model to accommodate several large outliers. Both models show that the level of superparasitism declines with increasing host density, (...) and the rate declines over time. It is proposed that the differing behaviours represented by the mixed model may reflect a balance between behavioural strategies of different selective benefit. (shrink)