Division of labour during honeybee colony defence: poetic and scientific views

Abstract

Poets, philosophers and politicians have used bees, and often projected an idealized human society into their view of how beehives are organized, from the ancient Greeks to present times. We first review how division of labour in honeybees was perceived by human observers, before presenting our current understanding. We focus specifically on defensive behaviour and show that this model provides an interesting case study for our conceptual understanding of division of labour as a whole. We distinguish three phases of the defensive response: detection of an intruder, recruitment of individuals into collective defence and attack. Individual bees may selectively contribute to one or more of these steps. Guard bees monitor entering conspecifics or attacking mammals, and release an alarm pheromone to recruit stinging soldiers. However, we are still far from understanding why only subsets of bees become guards or soldiers (or even if soldiering can be considered a task per se). We discuss the stimuli associated with each of these steps, how they define the number of bees needed and how they might combine with individual and developmental characteristics such that individuals take on a particular task. We also highlight pending questions and interesting avenues for future research.This article is part of the theme issue 'Division of labour as key driver of social evolution'.

Keywords: defence; guard bees; honeybee; soldier bees.

Figure 1.

Stinging responsiveness to a rotating dummy alone (solvent control with mineral oil (MO), in blue) or to the same rotating dummy in the presence of alarm pheromone (10% vol/vol IAA in mineral oil, in red) for bees taken from 30 different colonies in 4 different countries (AU: Australia; FR: France; GE: Germany; NZ: New Zealand). The bees were tested once in a small arena, in pairs, as described in [8]. A stinging trial corresponds to at least one of the bees stinging the dummy. The numbers in parentheses indicate the number of bee pairs tested with MO and IAA, respectively. The colonies are ranked according to their response to the MO control. IAA generally increases stinging responses, but not always.

Figure 2.

(A) Bee actively releasing the sting alarm pheromone (SAP). Credit: Jessica Helgen, UMN Bee Squad. (B) Stinging bees and embedded stingers on a black suede flag. Credit: David Vogel. (C) Prior to the disturbance, the number of bees on the landing board of the colonies tested were similar for the two subspecies carnica and mellifera. This number was obtained by averaging the number of bees visible at 3 time points: 2 min, 1 min and 1 s before the test. Mann–Whitney U-test, U = 1734, z = −0.403, p = 0.687. (D) More individuals from A. m. mellifera colonies were observed actively releasing the SAP during the disturbance, which consisted of a black suede flag moving up and down in front of the colony entrance. This test lasted 2 min. The flag was initially placed 4 cm away from the entrance and was moved to 12 cm after 1 min. Mann–Whitney U-test, U = 1276, z = −4.272, p < 0.001. (E) A. m. mellifera colonies developed a much stronger defensive response, as evidenced by the higher number of stingers embedded in the flag after 2 min. Mann–Whitney U-test, U = 1410, z = −3.189, p < 0.01.

Figure 3.

Stinging responsiveness of bees from the same 4 colonies at two different times of the year in Brisbane, Australia. The bees were faced with a rotating dummy in a small arena, in pairs, with or without the alarm pheromone (IAA) as already described for figure 1. Winter in Brisbane (June–August) is mild thus the colonies never stop foraging completely, but available resources are scarce. In winter condition, IAA did not significantly increase stinging propensity. Fisher exact tests, *p < 0.05, ns: not significant. Sample sizes are indicated inside each bar.