SummaryPrevious results have demonstrated that results of laboratory tests of hoarding, length of life and pupal weight of workers were positively correlated with colony honey production. Here, it is shown that hoarding was not significantly correlated with either life-span (r= 0·396) or pupal weight (r= 0·449), and that life-span was not significantly correlated with pupal weight (r= 0·452). Therefore under these conditions each test measured a different component of the complex behaviour of colony honey production, and each component appears to be controlled by a different group of genes. Selection for increased honey production on the basis of performance on all three tests should thus be feasible.