The superseded female's dilemma: ultimate and proximate factors that influence guarding behaviour of the carpenter bee Xylocopa pubescens

Both solitary and primitively social nests of the facultatively social carpenter bee Xylocopa pubescens can be found throughout most of the breeding season. In social nests there is reproductive division of labour between a dominant forager and a guarding female. Two types of guarding females can be discerned: the young pre-reproductive guards, and older, formerly reproductive guards. The latter type of guard is found when, after a take-over of reproductive dominance either by a nestmate (mostly a daughter) or an intruder, the defeated female stays in the nest instead of leaving to try and found or usurp another nest. She is then manipulated into the role of a guard. The dominant female profits from the presence of the guard since she protects the nest against pollen robbery by conspecifics (Hogendoorn and Velthuis 1993). We have studied why superseded females might prefer to remain as a guard, rather than try their luck somewhere else. The hypotheses investigated pertain to (1) the difficulty for the defeated female of finding a new nest and of restarting reproductive activities due to (a) ecological constraints (nest and pollen shortage) and (b) the effect of age and wear on the defeated female; (2) the effects of guarding in terms of inclusive fitness. We found that superseded females remained as guards significantly more often when a nestmate (not necessarily close kin) took over reproductive dominance than when an intruder did so. Other factors associated with the decision of the defeated female to stay or leave were her age and the number of her own young still present after the supersedure. The probability of finding or constructing a new nest was lower for old than for young females. After finding a nest, old females produced less brood than young foundresses. As a result of these two factors old superseded females gained, in terms of inclusive fitness, by staying as guards, whereas young females profited from leaving the nest. We interpret these results as an indication that guarding behaviour has evolved due to kin selection. However, kin discrimination apparently did not occur. Therefore we conclude that in this species kin selection is not, in the proximate frame of reference, based on kin recognition and preference for helping kin. Correspondence to: K. Hogendoorn