OBJECTIVE: To explore possible changes in the community attributes of haematophagous insects as a function of forest disturbance. We compare the patterns of diversity and abundance, plus the behavioural responses of three epidemiologically distinct vector assemblages across sites depicting various levels of forest cover. METHODS: Over a 3-year period, we sampled mosquitoes, sandflies and biting-midges in forested habitats of central Panama. We placed CDC light traps in the forest canopy and in the understorey to gather blood-seeking females. RESULTS: We collected 168 405 adult haematophagous dipterans in total, including 26 genera and 86 species. Pristine forest settings were always more taxonomically diverse than the disturbed forest sites, confirming that disturbance has a negative impact on species richness. Species of Phlebotominae and Culicoides were mainly classified as climax (i.e. forest specialist) or disturbance-generalist, which tend to decrease in abundance along with rising levels of disturbance. In contrast, a significant portion of mosquito species, including primary and secondary disease vectors, was classified as colonists (i.e. disturbed-areas specialists), which tend to increase in numbers towards more disturbed forest habitats. At pristine forest, the most prevalent species of Phlebotominae and Culicoides partitioned the vertical niche by being active at the forest canopy or in the understorey; yet this pattern was less clear in disturbed habitats. Most mosquito species were not vertically stratified in their habitat preference. CONCLUSION: We posit that entomological risk and related pathogen exposure to humans is higher in pristine forest scenarios for Culicoides and Phlebotominae transmitted diseases, whereas forest disturbance poses a higher entomological risk for mosquito-borne infections. This suggests that the Dilution Effect Hypothesis (DEH) does not apply in tropical rainforests where highly abundant, yet unrecognised insect vectors and neglected zoonotic diseases occur. Comprehensive, community level entomological surveillance is, therefore, the key for predicting potential disease spill over in scenarios of pristine forest intermixed with anthropogenic habitats. We suggest that changes in forest quality should also be considered when assessing arthropod-borne disease transmission risk.
