ORIGINAL PAPER Judith Reinhard Mandyam V. Srinivasan Shaowu Zhang Complex memories in honeybees: can there be more than two? Received: 12 July 2005 / Revised: 22 November 2005 / Accepted: 25 November 2005 / Published online: 20 December 2005 Ó Springer-Verlag 2005 Abstract Foraging honeybees are likely to learn visual and chemical cues associated with many different food sources. Here, we explore how many such sources can be memorized and recalled. Marked bees were trained to visit two (or three) sugar feeders, each placed at a dif- ferent outdoor location and carrying a different scent. We then tested the ability of the bees to recall these locations and fly to them, when the training scents were blown into the hive, and the scents and food at the feeders were removed. When trained on two feeder locations, each associated with a different scent, the bees could correctly recall the location associated with each scent. However, this ability broke down when the number of scents and feeder locations was increased to three. Performance was partially restored when each of the three training feeders was endowed with an addi- tional cue, namely, a distinct colour. Our results suggest that bees can recall a maximum of two locations when each is associated with a different scent. However, this number can be increased if the scent cues are augmented by visual cues. These findings have implications for the ways in which associations are established and laid down in honeybee memory. Keywords Associative learning Memory Vision Olfaction Honeybee Introduction A honeybee colony’s search for food is a well-coordi- nated process requiring precise orientation over dis- tances of several kilometres (Gould 1993; von Frisch 1993). During foraging flights, honeybees learn naviga- tional information about the location of a food source, in terms of its distance from the hive and its direction relative to the sun (Wehner and Rossel 1985; von Frisch 1993; Esch and Burns 1995; Srinivasan et al. 2000). Bees are also known to learn additional visual cues to aid the process of navigation, such as the landmarks expected en route and at the destination, or a flower’s colour and shape, as well as olfactory cues associated with a food source, such as floral odours, or the taste and fragrance of the nectar (Wehner 1981; Collett 1992; Gould 1993; von Frisch 1993). The honeybee’s amazing capacity to learn environ- mental cues has been investigated extensively (e.g. Wehner 1981; Collett 1992; von Frisch 1993; Srinivasan 1994; Menzel and Mueller 1996; Srinivasan et al. 1998; Zhang et al. 1999, 2004; Menzel et al. 2000; Giurfa et al. 2001). Recent studies exploring learning and memory under natural conditions in the field demonstrated that honeybees form and recall associative memories of food sources across sensory modalities (Reinhard et al. 2004a, b). Bees that were previously trained to forage at dif- ferently scented feeders, each positioned 50 m away from the hive, could be induced in subsequent tests to visit a specific feeder location merely by blowing into the hive the scent that was associated with that location, even though the feeders were empty and unscented during the tests. Perceiving the scents in the hive trig- gered recall of visual and navigational memories of the feeders associated with the specific scents, guiding them to the correct locations. Relying on scent-triggered recall of previously learnt information, honeybees navigated quickly and success- fully to the appropriate feeders, irrespective of the scents used and regardless of whether the bees had to recall a specific location, or a specific colour (Reinhard et al. 2004b). But this was only true if the number of feeders was restricted to two. In a preliminary experiment using three feeders, injecting a scent into the hive did not cause the trained bees to return preferentially to the specific location that was associated with the scent during training. This raises the question of whether the J. Reinhard (&) M. V. Srinivasan S. Zhang Visual Sciences Group, Research School of Biological Sciences, The Australian National University, P.O. Box 475, 2601 Canberra, ACT, Australia E-mail: judith.reinhard@anu.edu.au Tel.: +61-2-61259701 Fax: +61-2-61253808 J Comp Physiol A (2006) 192: 409–416 DOI 10.1007/s00359-005-0079-0