The distribution of eggs per host in a herbivorous insect--intersection of oviposition, dispersal and population dynamics.

1. The dynamics of parasitic organisms depend critically upon the frequency distribution of parasite individuals per host. However, the processes giving rise to this frequency distribution have rarely been modelled and tested for organisms with complex host selection behaviour. 2. In this study Microrhopala vittata, a chrysomelid beetle, was used to investigate how oviposition behaviour, movement and density of host plants interact in shaping the frequency distribution of egg clusters per host in the field. 3. Enclosures were stocked with two different host species and different beetle densities and various stochastic process models were fitted to egg cluster count data obtained from these enclosures. The different models were derived considering different scenarios, in particular whether or not plant density limits oviposition rate, whether or not ovipositing females actively seek out the most attractive plant within their perception radius and whether a female's oviposition rate is determined by plant intrinsic factors, the plant's egg cluster load or the surrounding beetle density. 4. The model parameters fitted to cage data were used to describe the frequency distribution of egg cluster counts obtained in a release experiment in the field. A total of 220 beetle pairs were released at five locations in a field where this beetle was not observed previously. Each release point was at a border between the two host species. 5. One model predicted for the preferred host species the egg cluster count frequencies in the field from parameters estimated in the cages. This model assumed that egg clusters present on a plant increased subsequent oviposition on this plant. All other models could not describe the distribution of egg cluster counts for either of the two host species. 6. The results suggest that females seek out attractive hosts actively and the attractiveness of a plant increases with its egg cluster load. This behaviour creates a frequency distribution of egg clusters per host that depends only on beetle density but not on plant density. This conclusion has important implications for modelling insect-plant interactions.