Habitat fragmentation and land use changes threaten neotropical habitats and alter patterns of diversity at forest edges. Like other arthropod assemblages, neotropical fruit-feeding butterfly communities show strong vertical stratification within forests, with some recent work showing its potential role in speciation. At forest edges, species considered to be forest canopy specialists have been observed descending to the forest understory, with the similarity in light conditions between the canopy and understory strata at edges hypothesized to be responsible for this phenomenon. We conducted a study using standardized sampling to document and quantify this edge effect, characterize edge and forest strata, and estimate the relative contributions of temperature and light conditions to changes in nymphalid butterfly stratification at forest edges. We found strong evidence of an edge effect in these butterflies and confirmed strong differences in light and temperature, showing that the edge understory differs little from forest canopy conditions. Of 41 species common to both forests and edges, 28 shifted to have a lower canopy probability at the edge, and our model detected a decrease in canopy probability of 0.165. Furthermore, our analysis indicated the relative abundance of canopy taxa increased at the edge, and the tribes Haeterini and Morphini were especially sensitive to edge effects. However, the analyses here did not clearly implicate temperature or light magnitude in causing changes in neotropical nymphalid vertical stratification at forest edges. Instead, our results point to other mediator variables as being important for changes at tropical forest edges. From our data, edge-responsive species can be separated into two different categories, which likely relates to their resilience to anthropogenic disturbance. We also note that structural causal models have a potential place in future work on tropical conservation, given they can provide causal estimates with observational data.

Paired, baited understory and canopy butterfly traps were placed at edge, ridge, and valley locations (a total of 32 trap pairs) at La Selva Biological Station May 2015 - August 2016. From 26 May 2015 to 12 August 2015, traps were checked four consecutive days a week; from 12 September 2015 to 10 August 2016, traps were checked five consecutive days in the second week of each month. Bait was added to the traps on the first day of each sampling week and supplemented as needed to maintain potency. Butterflies were photographed and marked, and HOBO loggers recording temperature and light data were attached to traps.