Main content area

Disentangling fragmentation effects on herbivory in understory plants of longleaf pine savanna

Levey, Douglas J., Caughlin, T. Trevor, Brudvig, Lars A., Haddad, Nick M., Damschen, Ellen I., Tewksbury, Joshua J., Evans, Daniel M.
Ecology 2016 v.97 no.9 pp. 2248-2258
annuals, biotic factors, demography, edge effects, forbs, grasses, habitat fragmentation, habitats, herbivores, landscape ecology, landscapes, leaves, observational studies, perennials, probability, rare species, savannas, shrubs, understory
Habitat fragmentation affects species and their interactions through intertwined mechanisms that include changes to fragment area, shape, connectivity and distance to edge. Disentangling these pathways is a fundamental challenge of landscape ecology and will help identify ecological processes important for management of rare species or restoration of fragmented habitats. In a landscape experiment that manipulated connectivity, fragment shape, and distance to edge while holding fragment area constant, we examined how fragmentation impacts herbivory and growth of nine plant species in longleaf pine savanna. Probability of herbivory in open habitat was strongly dependent on proximity to forest edge for every species, increasing with distance to edge in six species (primarily grasses and annual forbs) and decreasing in three species (perennial forbs and a shrub). In the two species of perennial forbs, these edge effects were dependent on fragment shape; herbivory strongly decreased with distance to edge in fragments of two shapes, but not in a third shape. For most species, however, probability of herbivory was unrelated to connectivity or fragment shape. Growth was generally determined more strongly by leaf herbivory than by distance to edge, fragment shape, or connectivity. Taken together, these results demonstrate consistently strong edge effects on herbivory, one of the most important biotic factors determining plant growth and demography. Our results contrast with the generally inconsistent results of observational studies, likely because our experimental approach enabled us to tease apart landscape processes that are typically confounded.