To assess the effects of landscape configuration on local plant species richness, we tested whether local species richness of forest understory plants is affected by the total forest area and forest edge length in the adjacent landscape. We also tested whether the landscape effect on species richness is different for forest and edge species. We estimated species richness from 113 forest sites in four regions in Northern Europe. At each site, we studied two plots, one at the edge and one in the core of the forest. Total forested area and forest edge length in circles with a 1-km radius, together with plot-specific variables of environmental conditions and temporal continuity of forests, were recorded at each plot. The amount of forest and the length of the forest edge in the adjacent landscape had a significant positive effect on local species richness of all understory plant species. As expected, edge species were positively affected by increasing length of the forest edge in the landscape, but surprisingly there was no effect of forest area on species richness of forest species. Temporal forest continuity had a negative effect on species richness of edge species but no effect on species richness of forest species. Our results suggest that forest edge length had a stronger landscape effect on understory plant species richness than forest area. Implications of these findings for the management of forest landscapes depend on priorities given to different species groups in biodiversity conservation, i.e. if emphasis is in total species richness or species richness of forest or edge species.

Aim There is an intense debate about whether habitat fragmentation has a negative or positive effect on biodiversity. We examined whether species richness and incidence of forest plants were negatively or positively associated with fragmented forest configuration. We also analysed whether the results support the fragmentation threshold hypothesis with fragmentation effects only in landscapes with small habitat amount. Location Södermanland province, south‐eastern Sweden (8,388 km2). Methods Data consisted of plant distribution maps and landscape data on forest amount and configuration in 2.5 km × 2.5 km quadrats. We carried out models including forest area together with clumpiness index (CL models) or edge density (ED models) as the measure of habitat configuration. We focused on plant taxa with positive association between incidence and forest area (163 taxa in CL models; 119 taxa in ED models). Results Responses to fragmented configuration were negative more often than by random (33 and 22 taxa in CL and ED models, respectively; includes only models without significant forest area × configuration interaction), whereas positive responses were rare (four taxa in both models). When forest area × configuration interaction was significant, the most common response had a negative effect of fragmented configuration when forest area was low and no effect of configuration when forest area was high, which agrees with the fragmentation threshold hypothesis. Species richness also had this type of response. In another common interactive response, the effect of fragmented configuration was negative at low forest area and positive at high forest area. Main conclusions Responses to fragmented forest configuration, when significant, were usually negative. When responses to fragmented configuration were modulated by forest area, they were negative when forest area was low. The findings of complex interaction between forest area and configuration have implications for selection of appropriate patch sizes in sustainable forest management.