|
Résumé :
|
Many studies on carabid ecology rely on the comparison of carabid assemblages among various habitat types, assuming that pitfall trap catches are not biased by the environmental conditions. However, some authors have already suspected that the structure of the vegetation surrounding pitfall traps may have an impact on the trapping efficiency by modifying carabid movements, either through a barrier effect of ground-vegetation or though micro-climatic conditions influencing beetle activity. Therefore, we wonder whether there might be a strong bias in capturability of carabid species between habitat types: does capture probability of carabid species strongly vary with canopy-closure degree in forest stands ? Without real capture-marked-recapture experiments of individuals, a probability of capture for a set of species can yet be estimated by using spatially-repeated measures of carabid assemblages, thanks to probabilistic methods coming from capture-mark-recapture theoretical models applied to community ecology. Thus, to evaluate the difference in capture probability between various habitat types, we analysed a carabid data set of 28 plots of different forest stages within a French temperate oak forest (Montargis state forest, 4000 ha). In each plot, 4 traps, half-filled with a solution of 50% ethylene-glycol, were arranged in the four corners of a 14 m x 14 m square and were activated during 3 one-week periods in late spring and late summer. The canopy-closure degree of the silvicultural stage modified the form of the species rarefaction curves from a 1-plot-by-stage strategy to a n-plot-by-stage strategy (n varying from 4 to 8 plots) : the seeding cuts seemed to be less exhaustively sampled than the other stages. However, as capture probabilities, estimated and compared using COMDYN program, did not significantly differ between silvicultural stages (Chi-square tests), the comparison of carabid species richness at the stage scale does not seem to be biased by the canopy-closure degree of the stage.
|
