Predators can reduce bee pollination and decrease plant fitness through fear: prey avoiding predators. However, the effects of predator-induced fear on bee communication are poorly understood. For bees that are mass foragers, such as honey bees, the effects of fear alter individual foraging choices and colony foraging allocation through recruitment communication. Honey bees (Apis mellifera) can recruit nestmates to visit food sources, thereby amplifying the effects of their individual foraging. It was not known if honey bees would alter their foraging communication, factoring in the risk of predation, when confronted with live predators on floral-type resources. We show that the potential risk of predation from live predators, not just the evidence of recent predation, alters the dance language in an adaptive way. In this study, honey bees were given a choice of safe and dangerous feeders to determine if they would avoid praying mantises (Tenodera sinensis), a predator that bees were not previously known to avoid. Bees exhibited fear (avoidance) and significantly avoided the live mantis (at all instar sizes >̲ 4cm) and avoided mantis odor (cuticular hydrocarbon extract) and visual cues (plastic mantis). Both olfactory and visual cues were equally effective deterrents. Honey bees decreased the number of waggle dance circuits (a measure of food source quality) for a rich food patch with a live mantis, but not for a mantis-free control, which is the first demonstration that wariness of a live predator alters bee communication inside the nest and suggests that information flow within a colony plays a role in the ecology of fear

Predators can reduce bee pollination and plant fitness through successful predation and non-consumptive effects. In honey bees, evidence of predation or a direct attack can decrease recruitment dancing and thereby magnify the effects of individual predation attempts at a colony level. However, actual predation attempts and successes are relatively rare. It was not known if a far more common event, just detection of a predator, could inhibit recruitment. We began by testing honey bees' avoidance of the praying mantis (Tenodera sinensis). Larger predators (later mantis instars, ≥4.5 cm in body length) elicited significantly more avoidance (1.3 fold) than smaller mantis instars. Larger instars also attempted to capture honey bees significantly more often than did smaller instars. Foragers could detect and avoid mantises based upon mantis odor (74% of bees avoided an odor extract) or visual appearance (67% avoided a mantis model). Finally, foragers decreased recruitment dancing by 1.8 fold for a food source with a live adult mantis, even when they were not attacked. This reduction in recruitment dancing, elicited by predator presence alone, expands our understanding of predator non-consumptive effects and of cascading ecosystem effects for plants served by an important generalist pollinator.