Investigações sobre discriminações simples e discriminações condicionais em abelhas

Abstract

Research on stimulus equivalence constitutes an experimental approach to assess variables underlying the development of symbolic behavior, which depends on the establishment of conditional relations between stimuli. Matching-to-sample procedure is typically used for teaching conditional discriminations, and it has been also used in early studies with honeybees. Bees are capable of learning conditional relations between stimuli, such as specific identity relations, arbitrary relations between visual stimuli, arbitrary relations between colors and odors and generalized identity matching. Therefore, basic processes involved in symbolic behavior could be studied using honeybees as an experimental model. The main objective of the present study was to establish conditional relations in honeybees. Another goal was to test for select and reject-control relations in simple discrimination baselines, since these controlling relations are involved in responding by exclusion. In Experiment 1, eleven bees Melipona quadrifasciata and Melipona rufiventris were individually trained to perform an identity matching-to-sample task. Identity relations were demonstrated by most of the bees. In Experiment 2, four M. rufiventris were individually trained on arbitrary relations. One of these bees demonstrated an increasing trend in accuracy in an arbitrary MTS task over the last sessions. In Experiment 3, twenty Apis mellifera and four M. rufiventris were trained to perform arbitrary relations. Additionally, Apis mellifera bees were presented with a symmetry test. Apis mellifera bees had acquired arbitrary MTS, but Melipona rufiventris performed at chance level. In the symmetry test, Apis mellifera performed at chance level. In Experiment 4, twenty M. rufiventris were individually trained to perform a simple discrimination task. Subsequently, each honeybee was presented to a select and reject control test. Select as much reject control had developed in simple discrimination training with M. rufiventris. On one hand, these data indicate that in most cases the arbitrary matching performance was not accurate. Moreover, symmetry relations did not emerge from training. On the other hand, most subjects responded highly accurately in simple discrimination and identity matching tasks. Most subjects had also demonstrated select and reject-control relations. These advancements in research on basic learning processes in bees may be a promising starting point for investigating the prerequisite repertoires involved in symbolic behavior.