Edge contrast modulates ant community responses to edge distance in agricultural landscapes

Although edge influence has been studied extensively, there is no consensus on how insect communities respond to edge environments. Studies highlight the importance of an experimental design that considers samples in adjacent matrix and the distance from the edge as a continuous factor. We sampled ants along edges composed by Cerrado vegetation and three types of crop matrix: sugarcane, eucalyptus and pasture. We assessed the response patterns of ant communities to edge contrast and distance using linear models. Ant richness increases linearly towards Cerrado core in edges composed by sugarcane, while it did not change along edges composed by eucalyptus. In both cases, ant composition on both sides of these edges tended to be highly different. In pasture, ant richness is higher than Cerrado, while the composition changes gradually along this edge, indicating that low contrast edges imply a smoother effect of edge influence on communities. Our results suggest that edge influence is dependent on the structural contrast between the habitats that compose the edges. We highlight the importance of the two‐sided approach on edge studies and the use of more than a single biodiversity dimension to provide an accurate picture of edge influence on insect communities. Responses of species richness (A‐C) and species composition (D‐F) of ant communities to edge distance in three types of anthropogenic edges in agricultural landscapes in São Paulo State, southeast Brazil. Colored lines represent plausible models: red lines represent no changes in ant diversity along the transect (null model), pairs of orange lines indicate that changes in ant diversity occur only between environments that compose the edge (habitat‐matrix model), and purple lines represent a linear change of ant diversity along the edge (linear model). Positive values of edge distance represent edge distance within Cerrado remnants, whereas negative values represent edge distance in adjacent agricultural matrix.