Estimation of population densities of carabid beetles in cereal crops.

A model is presented which aims to predict population densities of carabid beetles from pitfall-trap data. This is not normally possible because of interspecific differences in activity patterns, resulting in differences in the rate at which individuals of different species encounter traps. Experiments carried out both in the laboratory and in the field on 10 species of Carabidae reveal that another factor, the rate of avoidance of traps following encounters, also differs interspecifically and should be taken into account before any predictions of abundance using pitfall-trap data are attempted. Data on the activity patterns of 5 of these species is used to simulate movement of individual beetles on a hypothetical grid of pitfall traps in the memory of a computer. Activity is simulated for a number of different durations and the proportion of individuals of each species encountering traps is determined for each duration. This parameter is regarded, for each species at each duration, as an estimate of the probability of an individual encountering a trap in the field. This parameter is combined with the avoidance rate to modify pitfall- trap data collected over an equivalent period. This gives a modified estimate (corrected for activity and avoidance) of the relative abundance of species in pitfall traps. This relative abundance is compared interspecifically with absolute abundance estimates from the field to determine the accuracy of the model. A discussion of the assumptions accompanying the model is followed by suggestions for further refinements so that it might be used in the future to predict the absolute abundance of carabids which are natural enemies of agricultural pests such as aphids