Aversive conditioning in honeybees (Apis mellifera).

Abstract

Learning in honeybees, previously studied by appetitive techniques, was studied here by aversive techniques. In one series of experiments, a shuttle box was used, with the odor of formic acid as the aversive stimulus. A punishment contingency was found to suppress shuttling more in master animals than in yoked controls, whereas escape and unsignaled avoidance contingencies facilitated shuttling in master animals compared with yoked controls. In a second series of experiments, the subjects were unrestrained foragers flying back and forth between the hive and the sill of an open laboratory window to take sucrose solution from targets so constructed that shock could be delivered while the proboscis was in contact with the solution. A group of animals trained to discriminate between two differently colored targets, one providing sucrose and the other sucrose plus immediate shock, performed as well as a group trained with sucrose and tap water, and better than a group trained with sucrose and sucrose plus delayed shock. Animals for which a signal was paired with shock while they were feeding from a single target quickly learned to avoid the shock by flying off the target. The effectiveness of the pairing was demonstrated both by an explicitly unpaired procedure (which retarded acquistion when the signal and shock subsequently were paired) and by differential conditioning. This work substantially extends the range of vertebrate learning phenomena found in honeybees. Recent experiments on appetitive conditioning have revealed some striking similarities between the learning of honeybees and vertebrates, whose nervous systems have evolved independently for the most part and differ substantially both in organization and complexity. The technique used in the majority of these experiments has been to train free-flying foragers to shuttle back and forth from the hive to the laboratory where they take sucrose solution from targets distinguished by color, odor, or position (Couvillon & Bitterman, 1980, 1982, 1984; Couvillon, Klosterhalfen, & Bitterman, 1983; Klosterhalfen, Fischer, & Bitterman, 1978). This simple technique has made possible a variety of sophisticated learning experiments in which an array of vertebrate conditioning phenomena have been demonstrated. Among them are overshadowing, potentiation, and within-compound association in compound conditioning; dimensional shift in choice problems; the overlearning-extinction effect and its dependence on magnitude of reinforcement ; and successive negative incentive contrast. With foragers temporarily confined on each visit to permit better control of certain critical variables, Sigurdson (198 la, 1981b) did a variety of fully automated, massed-trials experiments demonstrating such phenomena as behavioral contrast, probability matching, progressive improvement in reversal learning, and the partial reinforcement effect. Work has been done also on the classical This article is adapted from a doctoral dissertation submitted to Boston University. The work was done at the University of Hawaii with the support of Grant BNS 83-17051 from the National'Science Foundation. The author is indebted to M. E. Bitterman for providing the facilities and for guidance throughout the course of the work.