Fluctuations of density and survival of carabid populations.

Standardized and all-year pitfall-sampling during a number of years gives reliable estimates of the yearly changes in size (year-catches) of the local breeding populations (interaction groups) of carabid species. From sampling runs at 89 sites, 2,893 estimates of coefficients of net reproduction (R) of 64 carabid species were obtained. The R-values of all species appeared to be lognormally distributed (LND). For each species realistic fluctuations of density could thus be simulated after estimation of mean and standard deviation of the fitted LND (Appendix). Simulation results were compared with field data in two ways: 1. The range between highest and lowest density (logarithmic range: LR), only for those sampling runs of 4-12 years that are not interrupted by zero year-catches. These LR-values were compared with those from simulated runs at similar catch-levels (n), i.e. runs with the highest catch between 1/2·3 and 1/2·3 , if n=1, 2,..., 7. As the number of uninterrupted (by zero year-catches) time series of catches decreases at lower catch-levels, by which the distribution of LR-values becomes skewed, each sampling run from the field could only be compared with simulations at the same catch-level. It thus appeared, that in the field density usually fluctuated within wider ranges than could be expected from the concerning simulations (where density fluctuated according to R-values drawn at random from the fitted LND). This means, that in the field the succession of R-values generally is less favourable than at random. This was especially evident for species with high powers of dispersal (T-species; glossary). In species with low powers of dispersal (L-species) the density range (LR) of field populations only at higher catch-levels agreed well with the expected values from the simulations. 2. The number of zero year-catches in sampling runs of 4-12 years taking into account the catch-level. In the field very low densities as a cause of zero year-catches cannot be distinguished from extinction and/or (re) founding (turnover). In the simulations the density level (start density) can be chosen such that only very low densities can still be a cause of zero year-catches (pseudo-turnover), the chance of which is higher the lower the catch-level. Correcting the field data for pseudo-turnover for each species gives estimates of the actual turnover. Local populations of carabid species thus appear to survive for only restricted periods, in T-species around 10 years, in L-species around 40 years. This fits in well with previous findings on differences in the Distribution of Population Sizes (DPS; glossary) between species (Den Boer 1977). The survival times of local populations are not so much determined by the instability of the habitat, but more by the instability of the populations themselves. The low survival times of local populations of T-species seem directly to result from the high level of dispersal.