Antra. Behav., 1971, 19, 299-304 MATING BEHAVIOUR OF DRONE HONEY BEES WITH QUEEN MODELS (APIS MELLIFERA L.) BY NORMAN E. GARY & JERRY MARSTON Department of Entomology, University of California, Davis, California 95616 Abstract. Wooden models of queen bees, when treated with queen pheromones and elevated (5 to 15 m), stimulated mating behaviour of flying drones. When sting chamber depths of 1.6, 4.8 mm or 'infinite' were tested at diameters of 1.6, 2.4, 3.2 and 4.0 ram, respectively, the 'mating' frequency for the respective diameters (all depths pooled) was 4.6, 20.0, 46.7 and 44.8 per cent (N = 240), respectively, for 762 drones that mounted the models. Sting chamber dimensions affected the degree of drone genital eversion. The median time between mounting and eversion was 2.4 s. Prolonged mount duration (median = 17.8 s) was observed when models had sting chambers too small to stimu- late eversion. The data document the brevity of mating and the open sting chamber requirement. Honey bee queens, Apis mellifera L., mate while flying at unknown distances from the hive. Observations of mating behaviour in nature have been extremely rare. Many observations of mating behaviour have been made during experi- ments in which tethered queen bees or queen pheromones were suspended approximately 5 to 20 m above the ground in restricted areas where drones could be attracted (Gary 1962, 1963). Tethered queens do not behave normally, but it is possible to observe the mating be- haviour of flying drones under relatively con- trolled conditions. Queens produce one or more pheromones, primarily 9-oxo-2-decenoic acid, that attract flying drones (Gary 1962 Pain & Ruttner 1963 Butler & Fairey 1964). In ad- dition, drones are strongly attracted visually to the queen and to other flying drones near the pheromone source. When drones mate with a queen, there is an instantaneous eversion of the drone's genitals into the sting chamber of the queen. Concomitantly, drones are instantly paralysed and fall to the ground where they soon die (Gary 1963). The mating act is repeated during the same flight, with approximately seven to ten drones (Taber & Wendel 1958). After mating, the queen returns to the hive. Gary (1963) reported that drones frequently mounted tethered queens, but rarely mated. After remaining in the mount position for a few seconds, they dismounted and joined the group of drones that hovered about the queen, A series of fortuitious observations indicated that the queen's sting chamber must be open before drone genital eversion was stimulated. Thus, even a dead tethered queen stimulated eversion or 'mating' if her sting chamber were open. Furthermore, queens that had a pseudo- sting chamber, produced by removing the term- inal abdominal segments and exposing the body cavity, induced drone genital eversion. Butler (1967) investigated the mounting be- haviour of drones by displaying dead worker and queen bees, with open and closed sting cham- bers, that were treated with synthetic 9-oxo- 2-deeenoie acid, queen extracts, or parts of queens. He reported that significantly more drones mounted the specimens with open chambers. 'Matings' of drones with the speci- mens were not reported, and presumably did not occur. The research reported herein was undertaken to elucidate the physical requirements of the sting chamber that stimulate eversion of drone genitals after drones have mounted the queen's abdomen in the typical dorsal position. Specific- ally, experimental evidence was sought to define the effects of varying the depth and diameter of sting chambers. Methods Model queen bees (Fig. 1) were constructed in order to control the diameter and depth of the sting chambers that were tested. The overall dimensions of a normal queen were duplicated approximately in the models, made of black walnut wood. No attempt was made to simulate the colour and body markings of normal queens. A hole of uniform diameter was bored com- pletely through the centre of the models, along the longitudinal axis, leaving openings of the same diameter at both ends. Into this hole was inserted a tightly fitting wooden dowel that could be adjusted anteriorly or posteriorly to create sting chambers of various depths. In addition, the dowel could be adjusted posteriorly until 299

300 A N I M A L B E H A V I O U R , 19, 2 Fig. 1. Suspension apparatus for wooden, modei queen. Sting chamber dimensions are variable. Flying drones 'mate' with the artificial queen, after being attracted to queen pheromones on the queen model, and from the living, captive virgin queen (diagrammatic). it was flush with the abdomen tip, in order to simulate a queen that has a closed sting chamber. Each model queen was glued permanently at the ventral, anterior end to the 'head' end of an insect pin. Thus mounted, the model could be affixed conveniently (by sticking the pin in the cork) to the end of a small cylindrincal, eight- mesh wire cloth cage that contained a captive, living virgin queen 1 to 2 weeks old. The captive queen cage was supported by a wire that was inserted into the cork at the upper end. The support wire was bent at the approximate angle to suspend the queen model and captive queen cage at an angle of 15 to 25 degrees from horiz- ontal, relative to their longitudinal axes. This angle presumably approximates the pitch of the queen's body during a mating flight. The wind caused constant movements of the models, and aecordingly caused a variable pitch angle. Rotation of the queen model apparatus was facilitated by its suspension on a swivel. Thus mounted, the queen model and captive cage oriented windwardly, thereby facilitating the approach and mounting of the models by drones flying windward towards the pheromone source. Movements of the models seemed to enhance the attraction of drones. In order to make the wooden queen models more attractive to drones on mating flights, each model was treated topically, just prior to ele- vation and testing, with mating attractant pheromones from the head of a virgin queen I to 2 weeks old. Application was made by merely crushing the head, including the pair of man- dibular glands that contain the pheromones, and spreading the contents as evenly as possible over the entire model. Additional pheromones emanated from the captive queen a living queen was used to insure the sustained release of 'total queen odour', in the event that deleterious chemical changes took place in the pheromones originating from the crushed head contents. Pheromones from the living queen presumably passed directly over and around the model, mounted in a downwind direction to achieve this objective. Four queen models were treated with phero- mones, then attached at stations spaced at 18-era intervals along a horizontal monofilament nylon line (0.635 mm diameter) that was elevated immediately to 5 to 15 m. The precise height was selected according to the optimum response of drones. Daily weather variations, especially wind, alter the flight altitude. Queen models were coded and arrayed on the support line by an assistant to prevent possible observer bias. During each test, all four models were ad- justed to identical sting chamber depths, leaving