The effects of pesticides on tomato pollinators depend on application practices and floral visitor assemblage

Pesticide use affects biodiversity and ecosystem service provision. Yet, such effects may vary between pesticides, dosage, application timing and pollinator assemblage in the study area. Understanding such diversity of effects is essential to enable integrated pest management practices that minimize negative impacts on pollinators. Here, we use a controlled experiment to compare different strategies of pesticide application in tomato crops: no application, low-intensity applications (every 7 days) and high-intensity application (every 3-4 days) with the last being the usual practice used by farmers in the study region. We focus on imidacloprid that is a neonicotinoid insecticide commonly used in Brazil. We show the negative effects of imidacloprid on the pollinator visitation rate of tomato flowers varied between flower visitor species. While the bee Paratrigona lineata was negatively affected by the imidacloprid application, no effects were detected when analysing all other bees as a group, indicating a null net effect. Although some studies have shown this insecticide is extremely harmful to bees' health, others showed that some bees do not avoid it, which may explain our results. However, the visitation rate recovery for P. lineata was only detected in the less than the more intensive treatment after imidacloprid applications ceased. The fact that no differences in fruit production were detected between treatments (including control blocks) could result from a null net effect of the negative effects on pollinators and positive effects on pest control. Further studies would be needed to disentangle the two effects.