Observational conditioning in flower choice copying by bumblebees (Bombus terrestris): influence of observer distance and demonstrator movement.

BACKGROUND: Bumblebees use information provided inadvertently by conspecifics when deciding between different flower foraging options. Such social learning might be explained by relatively simple associative learning mechanism: the bee may learn to associate conspecifics with nectar or pollen reward through previous experience of foraging jointly. However, in some studies, observers were guided by choices of 'demonstrators' viewed through a screen, so no reward was given to the observers at the time of seeing other bees' flowers choice and no demonstrator bee was present at the moment of decision. This behaviour, referred to observational conditioning, implies an additional associative step as the positive value of conspecific is transferred to the associated flower. Here we explore the role of demonstrator movement, and the distance between observers and demonstrators that is required for observation conditioning to take place.

METHODOLOGY/PRINCIPAL FINDINGS: We identify the conditions under which observational conditioning occurs in the widespread European species Bombus terrestris. The presence of artificial demonstrator bees leads to a significant change in individual colour preference toward the indicated colour if demonstrators were moving and observation distance was limited (15 cm), suggesting that observational conditioning could only influence relatively short-range foraging decisions. In addition, the movement of demonstrators is a crucial factor for observational conditioning, either due to the more life-like appearance of moving artificial bees or an enhanced detectability of moving demonstrators, and an increased efficiency at directing attention to the indicated flower colour.

CONCLUSION: Bumblebees possess the capacity to learn the quality of a flower by distal observation of other foragers' choices. This confirms that social learning in bees involves more advanced processes than simple associative learning, and indicates that observational conditioning might be widespread in pollinating insects, raising intriguing questions for the underlying mechanisms as well as the spread of social information in pollinator-plant interactions.