Analysis of the spatial component of pesticide side effects on non-target invertebrate populations and its relevance to hazard analysis.

The relative levels of recovery by populations of Carabidae, Staphylinidae and Linyphiidae were measured, following exposure to dimethoate, in two winter wheat field experiments. Regression analysis of the duration of recovery (the time taken for treatment plot population densities to approach those in control plots) with respect to distance from the plot boundary, revealed differences in the reinvasion potential of the three families. Staphylinidae recovered the most rapidly and Carabidae the most slowly with the Linyphiidae taking an intermediate position. The source of reinvasion differed between the three families with the carabid populations being confined within the field in contrast to the other two families which recolonized treated areas from the field boundary and beyond. The rate at which staphylinid and linyphiid populations dispersed into the plots raised questions concerning the most appropriate scale from which to obtain estimates of the duration of side-effects. More serious interpretative problems arose for the carabids since a large proportion of their apparent recovery may have been an experimental artefact. Regression analysis may provide a simple means of predicting the duration of side-effects on different taxa from field experiments, as long as these interpretative constraints are recognized. This approach incorporates an important feature of life history into predictions of pesticide hazard and may thus be more realistic than predictions based on pesticide toxicity alone. Key-words: Carabidae, dimethoate, dispersal rate, Linyphiidae, modelling, pesticide hazard analysis, Staphylinidae, temporal and spatial metapopulation dynamics