Relative reliability of four field assays to test defensive behaviour of honey bees (Apis mellifera)
Ernesto Guzmán-Novoa, Daniel Prieto-Merlos, José L Uribe-Rubio and Greg J Hunt
Many assays for measuring defensive behaviour of honey bee (Apis mellifera) colonies are reported in the literature, but their reliability in most cases has not been assessed. This study was conducted to determine the reliability and convenience of four field assays: (1) a test to score behavioural traits that differ between defensive and gentle bees (ratings assay), (2) two stinging tests in which a leather patch is presented to the bees above the brood chamber (flag assay), or (3) in a confined space (box assay), and (4) a test to measure the numbers of bees recruited at the hive entrance in response to a chemical stimulus (pheromone assay). Six colonies each of three different bee types known to vary in their defensive responses (European, Africanized, and hybrid) were established and tested with the assays. To determine the reliability of the assays, coefficients of variation (CV) among trials for single colonies were calculated. Additionally, colonies were ranked according to their degree of defensiveness and their rankings compared for each and between the different assays. The mean CV for the ratings assay was 17.0%, vs. 71.8, 90.4, and 67.3% for the flag, box, and pheromone assays, respectively. The pursuing behaviour of the experimental bees was the component of the field assays that best separated gentle and defensive colonies. The mean number of pursuing Africanized bees was> 81 times higher than that of pursuing European bees. No significant interactions between day of test and colony were found. Rankings for all colonies were comparable and similar across all assays, except the box assay. All assays were able to separate gentle from defensive bee types, but the ratings assay was the most reliable, economical and practical.