The reliability of honey bee density estimates from trapped drones

Australia has an abundance of feral honey bee colonies. Understanding their densities is important to assess their current economic and ecological impact and the need for mitigation should the mite Varroa destructor become established. Inferring colony density based on the genotypes of honey bee drones (males) caught in a Williams trap has been identified as a promising approach. This method assumes that (a) drones are attracted to the trap from an area bounded by the drone flight range, (b) sufficient colonies present within that radius are represented in a sample and (c) colonies that do not produce drones are small and of little ecological consequence. Here, we investigate whether known feral colonies were represented in drone samples and whether drone contribution per colony correlated with the relative colony size or the distance between the colony and the trap. We found that one-third of known colonies were not represented in the drone sample, and this proportion did not correlate with colony size or distance. For colonies that contributed at least one drone, there was a correlation between the number of drones caught per colony and the distance of the colony from the DCA, and at distances beyond 0.9 km, there was substantial non-detection. Further work is needed to determine an appropriate correction factor that converts the estimated number of colonies represented in a drone sample to colony density.