Foraging honeybees (Apis mellifera), trained with 2 successively presented targets labeled with different odors, one target baited with a small drop of 50% sucrose solution and the other baited with a small drop of 20% sucrose solution, soon come to respond more promptly on 50% than on 20% trials (prospective effect) and more promptly after 20% than after 50% trials (retrospective effect), with a pronounced interaction between the two effects. In training with unlabeled targets, the retrospective effect is absent, which argues against postingestive inhibition as an explanation, but the effect appears precipitously, along with the prospective effect and the interaction, when odor labels are introduced (Experiment 1). Three subsequent experiments provided no evidence for an associative explanation of the retrospective effect in terms of discriminatio n supported by adaptation-based differential reinforcement. In recent work on the role of sucrose concentration in the learning of honeybees (Loo & Bitterman, 1992, Experiment 3), individual foragers were trained with two successively presented targets labeled with different odors, one of which always contained a small drop of 50% sucrose solution and the other, a small drop of 20% sucrose solution. Latency of response to the targets, which was substantial at first, declined sharply with continued practice. In addition to the practice effect, there were two concentration-related effects on latency—one prospective (lower latency on 50% trials than on 20% trials), the other retrospective (lower latency after 20% reinforcement than after 50% reinforcement)— and an interaction between them (a larger retrospective effect on 20% trials than on 50% trials). The prospective effect, reminiscent of results obtained in runway experiments with rats (Goodrich, 1960; Kraeling, 1961), is most simply explained on the assumption that associative strength increases with sucrose concentration, although a representational or incentive-motivational explanation is equally tenable. Our concern in this study is with the source of the retrospective effect. Responding after 20% sucrose may have been speeded by something akin to frustration (Amsel & Roussel, 1952) if the animals had some expectation of 50% sucrose, but not much weight can be given to that possibility because the retrospective effect persisted long after the odor discrimination was well established. The frustration assumed to be generated in such experiments by encounters with the