Area- but no edge-effect on woody seedling abundance and species richness in old Afromontane forest fragments

Question: What is the relative importance of area- and edge-effects on woody seedling diversity in old Afromontane forest fragments? Location: Mistbelt Afromontane forests, KwaZulu-Natal midlands, South Africa. Methods: Woody seedling abundance and species richness in 590 1-m2 plots were sampled at the forest edge (< 10 m from the edge) and interior in 31 old (> 60 a) Afromontane forest fragments (0.05 – 328.5 ha) with closed edges in an ancient grassland matrix. Results: Unlike young (< 20 a) Amazonian fragments, there was no edge- or area-effect on sample plot seedling density and species richness, although these increased significantly with increasing herb cover (less disturbance). Seedling density, but not species richness, declined significantly with herbivory of seedlings, regardless of forest size or plot location. Seedling community composition and richness did not differ significantly between the edge and interior of forests across the range of forest sizes (i.e. no edge-effect). Community composition was nested with small forests retaining a subset of the seedling flora of larger forests. Overall, cumulative seedling species richness increased with forest area (i.e. area-effect). Conclusions: Holocene climatic extinction filtering events and area-dependent species relaxation have potentially selected for tree species with convergent life histories adapted to local fragmentation-effects. Stable environmental conditions at old edges in these naturally fragmented forests cause similar regeneration conditions and seedling species composition between edge and interior. Consequently, seedling density and species richness are controlled more by response to gradients of local disturbance (habitat area, herb cover, herbivory) than by proximity to the edge. Large patches (> 50 ha) with intact edges had the highest tree seedling diversity and are a conservation priority. Although small patches contain no unique species they preserve landscape processes, have conservation value, and require protection. Conservation principles derived from recently created Amazonian fragments and that emphasize edge-effects, require critical evaluation for application to old Afromontane patches. Nomenclature: Arnold & de Wet (1993).