False queens: A consequence of mandibular gland signals in worker honeybees

Abstract

Observations on queenless honeybees have indicated that it is possible for workers to become egg layers [1-7]. What is even more curious is that some of these laying workers are capable of becoming what Sakagami has called "false queens" [3, 5]. The criteria for this designation being dependent on their ability to elicit retinue behaviour and in being able to inhibit emergency queen rearing in other workers should female larvae be available. Laying workers and false queens are both able to inhibit to some extent ovarial development in other workers. These behavioural and physiological effects can be understood in terms of the production of chemical signals a false queen emits which allow her to mimic queen signals [6]. The queen mandibular gland signals, within the hive, are involved in eliciting retinue behaviour and in the inhibition of both emergency queen rearing and ovarial activation in the workers [7]. The chemical constituents of this gland have been extensively investigated. The major component of the queen mandibular gland has been identified as (E)-9-oxo-2decenoic acid (9-ODA) [8, 9], the so-called queen substance. The worker mandibular gland produces (E)-10-hydroxy-2-decenoic acid (10-HDA), which is also a component of royal and worker jellies [t0] and 2-heptanone which is thought to be an alarm substance [tl, 12]. Recently a number of other fatty acid constituents have been identified in royal jelly and the mandibular glands of workers and queens [13]. Thus worker mandibular gland contents are said [14, 15] to differ markedly from those of the queen and hence the signals workers may be capable of releasing should be very different from those of the queen. The distinction between queen and worker signals has been blurred by the discovery that Apis mellifera capensis workers can synthesize the major component of the queen mandibular gland secretion, 9ODA, in their mandibular glands [16, 17]. Workers of A. capensis are capable of becoming laying workers very readily [1] and these workers, when placed in queenless groups of other races [16], very rapidly establish themselves as false queens. That these laying workers are capable of mimicking queen mandibular gland secretions and thus could produce queen-like chemical signals, destroyed the distinction [14, 18] that had previously been drawn between queen and worker mandibular gland products. As with ovarial development, the effects of caste determination on the biosynthetic capabilities of the mandibular gland are not irrevocable. The fact that A. capensis workers (amongst other pecularities) can become laying workers and false queens so readily, suggested that they occupied an intermediate position between A. melliJera workers and A. mellifera queens in terms of their ability to produce queen-like signals. However, the occasional appearance of false queens among A. mellifera laying workers [4, 7] was the spur which made us reinvestigate these mandibular gland signals. The mandibular glands of the series of honeybees (Table 1) were investigated by removing the pair of mandibular glands from each individual and analysing their contents gas-chromatographically (Fig. 1) [19, 20]. The bees used in these experiments were reared and maintained according to methods described in [5]. The A. mellifera workers in group A (Table 1), as has been reported previously [14, 15], have secretions dominated by 10HDA. This fatty acid is said to be "characteristic" of workers [18] yet is present in all the individuals examined. The two egglaying individuals produced a greater range of acids (10-HDA, 10-HDAA, 9HDA) than workers with underdeveloped ovaries with one exception. The exceptional individual with underdeveloped ovaries also produced a range of acids similar to that of the laying workers. The enhanced production of acids in this individual suggests that the production of appropriate mandibular gland signals may precede ovarial development. The secretions of the A. mellifera laying