Which is the appropriate scale to assess the impact of landscape spatial configuration on the diet and behavior of spider monkeys?

Understanding the response of species to changes in landscape configuration is required to design adequate management and conservation strategies. Yet, the most appropriate spatial scale (i.e., landscape size) to assess the response of species to changes in landscape configuration (so-called "scale of effect") is largely unknown. In this paper, we assess the impact of landscape forest cover, forest fragmentation, edge density, and inter-patch isolation distance on the diet and behavior of six communities of spider monkeys (Ateles geoffroyi) in the fragmented Lacandona rainforest, Mexico. We evaluated the strength of the relationship between each landscape predictor and each response variable within ten different-sized landscapes (range = 50-665 ha) to identify the landscape size that best predicted changes in diet and behavior. The strength of most associations varied across spatial scales, with the 126-ha landscape showing the strongest relationships between landscape predictors and response variables in many cases. Yet forest cover represented the main driver of the diet and behavior of spider monkeys, being positively associated with time traveling and time feeding on wood, but negatively related to time resting and time feeding on leaves. Although weaker, the impact of edge density was opposite to forest cover for most response variables. Forest fragmentation and isolation distance showed the weakest associations with the diet and behavior of this species. Our findings thus indicate that different landscape attributes operate on different response variables at different spatial scales. Therefore, the scale of effects cannot be generalized to all response variables and to all predictors, and a multi-scale analysis will be required to accurately assess the impact of landscape configuration on species' responses.