Naturwissenschaften (2000) 87 : 395–397 Q Springer-Verlag 2000 R.F.A. Moritz (Y) 7 U.E. Simon Institut für Zoologie, Martin-Luther-Universität Halle-Wittenberg, Kröllwitzer Str. 44, 06099 Halle/Saale, Germany e-mail: r.moritz6zoologie.uni-halle.de Tel.: c49-345-5526223 Fax: c49-345-5527264 R.M. Crewe Department of Zoology and Entomology, University of Pretoria, Pretoria 0002, South Africa SHORT COMMUNICATION R.F.A. Moritz 7 U.E. Simon 7 R.M. Crewe Pheromonal contest between honeybee workers (Apis mellifera capensis) Received: 2 July 2000 / Accepted in revised form: 28 July 2000 Abstract Queenless workers of the Cape honeybee (Apis mellifera capensis) can develop into reproduc- tives termed pseudoqueens. Although they morpholog- ically remain workers they become physiologically queenlike, produce offspring, and secrete mandibular gland pheromones similar to those of true queens. However, after queen loss only very few workers gain pseudoqueen status. A strong intracolonial selection governs which workers start oviposition and which re- main sterile. The “queen substance”, 9-keto-2(E)-dece- noic acid (9-ODA), the dominant compound of the queen’s mandibular gland pheromones, suppresses the secretion of queenlike mandibular gland pheromones in workers. It may act as an important signal in pseudo- queen selection. By analysing the mandibular gland pheromones of workers kept in pairs, we found that A. m. capensis workers compete to produce the strongest queen-like signal. Establishing dominance hierarchies in animal societies often involves fighting and agonistic interactions (Wil- son 1980; Engels 1990). But the honeybee colony ap- pears as a harmonious unit, free of conflict, and an ex- ception to this rule (Ratnieks 1988). This is surprising, because excessive multiple mating of the queen causes a low intracolonial relatedness. As a consequence, the colony is composed of many subfamilies (Moritz et al. 1995; Oldroyd et al. 1997) causing an inherent potential for conflict over the distribution of the colony’s re- sources. One mechanism of conflict avoidance is queen dominance and the suppression of worker reproduction by the queen’s mandibular gland secretion (QMS) (Slessor et al. 1988; Winston et al. 1989; Winston and Slessor 1992; Naumann et al. 1991). In concert with brood pheromones and other queen signals it sup- presses ovary activation in workers. QMS is dominated by 9-keto-2(E)-decenoic acid (9-ODA), the “queen substance”, which is also essential for many other regu- latory processes in honeybee biology (Moritz and Southwick 1992). In spite of this pheromonal control, some workers nevertheless activate their ovaries and lay eggs, most of which are eliminated by worker polic- ing (Ratnieks and Visscher 1989). Workers’ offspring are extremely rare in queenright honeybee colonies (Visscher 1996). In queenless colonies both control mechanisms break down. Workers develop into “pseudoqueens” that lay eggs and themselves produce QMS (Crewe and Velthuis 1980; Page and Erickson 1988). This process is particularly dramatic in the Cape Honeybee, Apis mel- lifera capensis. In a process of extreme intracolonial se- lection, only very few workers manage to establish themselves as pseudoqueens (Moritz et al. 1996) where- as all others remain sterile. Because pseudoqueens have an extreme reproductive advantage, one might ex- pect fierce competition for reproduction similar to the lethal fights of virgin honeybee queens. However, al- though behavioural studies report increased worker ag- gression in queenless colonies, those workers develop- ing into pseudoqueens do not participate in such physi- cal fighting (Velthuis et al. 1990). If dominant workers are neither being killed nor kill rivals, what are the mechanisms that govern this extreme selection, and why do we not see many more pseudoqueens in the col- ony? We used A. m. capensis workers as experimental subject to address this question. These workers rapidly activate their ovaries under queenless conditions (Ruttner and Hesse 1981) and QMS has been detected in A. m. capensis workers within 24 h after emergence (Simon 1998). We suspected that the mandibular gland signals were an important cue in pseudoqueen selec-