For many species, not all required resources are contained in breeding habitat. Such species depend on landscape complementation, i.e., linking together different landscape elements through movement, to complete their life cycles. We suggest that the dichotomous habitat classification of many metapopulation analyses (habitat vs. nonhabitat) masks our ability to detect metapopulation effects for such species. We tested this using a species for which landscape complementation is obligate and metapopulation structure is likely: Rana pipiens, the northern leopard frog. We used breeding chorus survey data to index relative abundance of leopard frogs in 34 ''core'' ponds and conducted Poisson regression analysis to determine the effects on frog density of local pond habitat, availability of summer habitat (landscape complementation), and number of occupied ponds in the surrounding landscapes (metapopulation structure). All of these factors had statistically significant effects on frog density. However, when summer habitat was not included in the statistical model, the metapopulation structure was no longer significant; i.e., its effect was masked. Our results suggest that one must be cautious in applying the results of metapopulation analyses to species for which the habitat vs. nonhabitat categorization of the landscape is not appropriate. The potential for rescue and recolonization to maintain a regional population must be assessed within the constraints of the entire landscape.

No abstract

No abstract

No abstract

To examine the effects of forest fragmentation on within-population genetic structure of Acer saccha rum, the spatial distributions of allozyme variation in the first-year seedling cohorts of four forest patch populations (patches) were compared with those of four populations within continuous forest (controls). Forest patch populations exhibited less spatial mixing of genotypes than controls at the smallest scale examined (10-14.1 m), possibly as a result of reduced overlap of seed shadows in patches, which generally had lower densities of reproductive trees. Patch populations exhibited greater mixing of genotypes than controls at the largest scale examined (113.1-141.4 m), possibly as a result of the incorporation of immigrant pollen pools into mating events at forest patch edges. This may extend the spatial range of patch population breeding associations which might otherwise be truncated owing to limited forest patch area. Overall, results suggest that mating events are probably the primary determinant of spatial genetic structure within these cohorts in both forest patch and control populations of A. saccharum and that forest fragmentation has affected genetic structure by changing patterns of gene flow within, and possibly among, forest patch populations.