Sampling Hymenoptera along a precipitation gradient in tropical forests: The effectiveness of different coloured pan traps

Measuring species richness of tropical insects is an important but considerable challenge. Several techniques have been developed to quantitatively sample the non-formicid Hymenoptera (bees and wasps). One of the most common is the use of colored pan traps. Although it is known that Hymenoptera are attracted differently by different colors, it is not yet known if these preferences shift in different habitats and hence affect comparisons of Hymenoptera diversity. We studied the effectiveness of differently-colored pan traps along a latitudinal, climatic, and forest structure gradient from tropical to subtropical forests. Overall, we found a strong increase in individual numbers from north to south. Yellow traps sampled significantly more individuals than blue ones, mainly due to the responses of the families Ichneumonidae, Nyssonidae, Pompilidae, and Crabronidae, but trap catch was also related to canopy cover. Notably, traps located at forest edges had yellow/blue ratios similar to those of forests with comparable canopy cover. This suggests that, in contrast to the overall number of individuals caught, the relative effectiveness of yellow vs. blue traps was driven by canopy cover and hence light conditions or visibility of the traps. Thus comparisons of pan trap results between forests having different structures should only be made with great care.