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 themost 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 fromnorth 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 bemade with great care.