2Abstr act. In nature conservation it is important to understand under which circumstances populations can survive by compensating local extinctions with colonisations. Many saproxylic (= wood-dwelling) insect species have declining populations and are regarded as threatened due to low habitat availability in managed forests. Several methods have been used to better understand the dispersal biology and colonisation ability of saproxylic insects with declining populations. In this paper, I summarize and compare the results of such studies.When the same species have been studied using several methods, the results are consistent, but different aspects of dispersal biology are revealed with different methods. Capturerecapture and telemetry are direct methods that can be used to quantify dispersal rate and range in the field. Studies of genetic structure and occupancy patterns are complementary, as they reveal the consequences of dispersals that have taken place over a larger spatial and temporal scale than is possible to study with direct methods. Because colonisation, rather than dispersal, is most important for population persistence, colonisation experiments provide useful information. To obtain information relevant for conservation work dispersal studies should be conducted on model species that are representative of threatened species.Colonisation ability probability differs between common and rare species, and therefore it is important to also study dispersal of rare species, even if it is more difficult.Key terms: capture-recapture, colonisation, dead wood, Osmoderma eremita, telemetry, tethered flight 3
Intr oductionIn order to understand population dynamics and assess extinction risks, knowledge about species' dispersal rate and range are important. Dispersal is necessary for gene flow (e.g. Slatkin 1987), colonisation of empty habitat (e.g. Hanski et al. 1994), and may also affect the population dynamics (Pulliam 1988) and extinction risks of local populations (Brown and Kodrick-Brown 1977). Therefore, dispersal affects how the populations are able to cope with changes in the environment, such us habitat loss and fragmentation (Thomas 2000). Models that do not explicitly involve costs for dispersal (such as Levins ' (1969, 1970) metapopulation model) suggest that dispersal is always beneficial. If the dispersal rate is high, however, an increased local extinction risk may make species with high dispersal rates more vulnerable in fragmented landscapes (Hanski and Zhang 1993). Empirical data support the view that there is no simple positive relationship between dispersal and population persistence in fragmented landscapes; Thomas (2000) related the recent decline of British butterfly species with their proneness for dispersal and found that species in the intermediate category had experienced the most serious declines, while both more mobile and more sedentary species had persisted better. This pattern is dependent on the degree of fragmentation; at a later stage in the fragmentation process, we should expec...