- Email: [email protected]
Foraging Bumble Bees Weigh the Reliability of Personal and Social Information
Report
Foraging Bumble Bees Weigh the Reliability of Personal and Social Information Highlights
Authors
d
Foraging bumble bees use the reliability of both social and personal information
Aimee S. Dunlap, Matthew E. Nielsen, Anna Dornhaus, Daniel R. Papaj
d
If social information is at all reliable, it is preferred over personal
Correspondence
d
Both personal and social information are successfully learned when reliable
[email protected] (A.S.D.), [email protected] (M.E.N.)
In Brief Dunlap et al. find that bumble bees use the reliability of information gained from both personal experience and other bees to make foraging decisions. They are especially sensitive to reliability of social information and trust others above themselves when information from other bees is reliable, even if personal information is more reliable.
Dunlap et al., 2016, Current Biology 26, 1195–1199 May 9, 2016 ª 2016 Elsevier Ltd. http://dx.doi.org/10.1016/j.cub.2016.03.009
Current Biology
Report Foraging Bumble Bees Weigh the Reliability of Personal and Social Information Aimee S. Dunlap,1,2,* Matthew E. Nielsen,1,* Anna Dornhaus,1 and Daniel R. Papaj1 1Department of Ecology and Evolutionary Biology, Center for Insect Science, University of Arizona, 1041 E Lowell Street, Tucson, AZ 85721, USA 2Department of Biology, University of Missouri-St. Louis, One University Boulevard, St. Louis, MO 63121, USA *Correspondence: [email protected] (A.S.D.), [email protected] (M.E.N.) http://dx.doi.org/10.1016/j.cub.2016.03.009
SUMMARY
Many animals, including insects, make decisions using both personally gathered information and social information derived from the behavior of other, usually conspecific, individuals [1]. Moreover, animals adjust use of social versus personal information appropriately under a variety of experimental conditions [2–5]. An important factor in how information is used is the information’s reliability, that is, how consistently the information is correlated with something of relevance in the environment [6]. The reliability of information determines which signals should be attended to during communication [6–9], which types of stimuli animals should learn about, and even whether learning should evolve [10, 11]. Here, we show that bumble bees (Bombus impatiens) account for the reliability of personally acquired information (which flower color was previously associated with reward) and social information (which flowers are chosen by other bees) in making foraging decisions; however, the two types of information are not treated equally. Bees prefer to use social information if it predicts a reward at all, but if social information becomes entirely unreliable, flower color will be used instead. This greater sensitivity to the reliability of social information, and avoidance of conspecifics in some cases, may reflect the specific ecological circumstances of bee foraging. Overall, the bees’ ability to make decisions based on both personally acquired and socially derived information, and the relative reliability of both, demonstrates a new level of sophistication and flexibility in animal, particularly insect, decision-making. RESULTS The natural environment encountered by pollinator species such as bees offers a multitude of cues that may be used to identify resources, all of which will vary in the reliability with which they indicate reward. Diverse environmental cues such as the color, shape, and other attributes of flowers can be used to distinguish species offering accessible, high-quality reward from less-
rewarding or non-rewarding species [12, 13]. However, several factors reduce reliability of these cues and thus personally gathered information: the reward of individual plants of the same species may vary due to microhabitat differences; some flowers may have been emptied by competing foragers, and some floral cues may be more difficult to detect and discriminate among than others, resulting in a change in the reliability of these cues from the point of view of the forager [12, 13]. Social information for pollinators can also vary in reliability. When a floral resource is abundant, conspecifics can provide a good indicator of reward presence; however, conspecifics may themselves make lessthan-perfect choices in resources. Even if a conspecific is visiting the best resource, that resource has now likely decreased in value because it has been at least partially exploited by the same conspecific providing the cue [14, 15]. An individual pollinator’s foraging success may thus depend on its ability to assess the relative reliability of the two forms of information and to respond appropriately. Nectar-foraging bumble bees provide an excellent system for studying the use of personal and social information. Bumble bees forage on flowers, an ephemeral resource, often in the company of other bees; both personal and social information can be used by bees to make floral choices (reviewed in [16]). In terms of personal information, bumble bees demonstrate individual trial-and-error learning, using a variety of floral cues associated with nectar reward, including color, odor, shape, texture, and spatial location [17]. In addition to personal learning of floral cues, bumble bees also use a variety of social cues to inform their flower choices. Social cues range from direct cues such as the sight of conspecifics on particular flowers or the odors left in a conspecific’s ‘‘footprints’’ to indirect cues such as floral scents carried into the nest [18–24]. Social information can even facilitate individual learning of floral cues by bumble bees [25– 27]. Finally, bumble bees show natural responses to artificially manipulated representations of floral and social cues (e.g., [25–29]), enabling us to control the reliability of floral (= personal) and social cues independently and precisely. To assess how the reliability of personal versus social information affects their use by bumble bees during foraging, we trained bees on arrays of artificial flowers in which we independently manipulated flower color (a floral cue providing personal information), the presence or absence of a pinned conspecific (a social cue providing social information), and the presence of a concealed sucrose reward (Figure 1). Each cue, floral or social, was made a fully reliable, moderately reliable, or unreliable indicator of the sucrose reward, in a fully factorial design (Table 1).
Current Biology 26, 1195–1199, May 9, 2016 ª 2016 Elsevier Ltd. 1195
Table 1. Predictions and Outcomes of Bee Preferences for Flowers Associated with Floral and Social Cues as a Function of Their Reliability Social Reliability 50% Floral Reliability
83%
100%
50% predict: neither predict: social predict: social result: neither U result: social U result: social U 83% predict: floral result: floral U 100% predict: floral result: floral U
predict: either predict: social result: social result: social U but not floral U predict: floral result: social X
predict: either result: social but not floral U
Checkmark (U) indicates that result met prediction; X indicates that result did not meet prediction.
Figure 1. Diagram of the Colored-Disc Artificial Flowers and Dried Pinned Bumble Bees Used as Social Cues The final testing array consisted of three of each of these types, for a total of 12 flowers. See also Figure S1.
Under these conditions, bees making perfect economic choices should use the most reliable information source, whether floral or social. They may also use the other source if it can provide any increase in total reliability, but they should never pay attention to unreliable sources of information. We tested these hypotheses by measuring the preferences of bumble bees on a non-rewarding array after training, finding that bees do not use unreliable information sources, personal or social, but favor social information over personal information so long as it has at least some reliability. Effect of Reliability on Information Use The preference during the test of bees for the floral versus social cue generally depended on their relative reliability during training as an indicator of sucrose reward (Table 1; Figure 1). In five of six treatments where one type of cue was more reliable than the other, the more reliable cue was preferred (Table 1). Nevertheless, bees did not respond equivalently to the reliability of the two types of cue. The response to the reliability of the social cue was strong and consistent. When the social cue was unreliable (reliability = 50%, i.e., a random chance of indicating food), bees never chose flowers bearing that cue more often than random (Table 1; Figure 2). In an extreme case, when social cues were unreliable (50%) and floral cue reliability moderate (83%), bees actually avoided the social cue (Figure 2A; t9 = 2.32, p = 0.02), which is not predicted in terms of reliability alone. On the other hand, when the social cue provided any reliable information (reliability R 83%), the response to the social cue was always significantly greater than 0.5 (Table 1; Figures 1B and 1C), meaning bees sought out flowers with conspecifics. Overall, these patterns gave rise to a strongly significant effect of social cue reliability on bee preference (repeated-measures ANOVA, F2,81 = 5.362, p < 0.0065; Table S1), as well as an interaction between social cue reliability and the focal cue type being measured (repeated-measures ANOVA, F2,81 = 17.026, p < 0.00001; Table S1). 1196 Current Biology 26, 1195–1199, May 9, 2016
In contrast, the bees’ response to the reliability of the floral cue was complex. When the floral cue was unreliable (50%), bees never chose flowers bearing that cue more often than random (Table 1; Figure 2), consistent with expectations. When the reliability of the floral cue was R83%, though, the response to the floral cue depended on the reliability of the social cue. If the social cue was unreliable (50%), bees responded significantly to a reliable floral cue (Figure 2A). However, when the social cue was reliable (83% or 100%), the floral cue was ignored, even when it was more reliable than the social cue (Figures 2B and 2C). Overall, the effect of floral cue reliability alone was not statistically significant (F2,81 = 0.937, p = 0.396) but interacted significantly with the effect of social cue reliability and the effect of focal cue type (social reliability 3 floral reliability 3 cue type interaction, F4,81 = 3.223, p = 0.0166; Table S1). Performance during Training We also examined how bees performed during the training phase by tabulating the number of rewarded visits across each of ten blocks of ten trials each. Bees showed improvement in terms of choosing rewarded flowers over blocks of trials (repeated-measures ANOVA, F9,729 = 5.84, p < 0.000001; Figure S2). Both social cue reliability and floral cue reliability had significant effects on how successful bees were at obtaining reward while foraging during training (F2,81 = 8.067, p = 0.0006 and F2,81 = 4.36, p = 0.016 for social and floral reliability, respectively; Figure 3). These results suggest that bees learned to use each of the cues. None of the interactions among block, floral cue reliability, and social cue reliability effects were statistically significant (Figure S2; Table S2). DISCUSSION We have shown here that bumble bees assess and react to the reliability of both personally acquired information and information acquired from conspecifics. To our knowledge, this is the first time that the reliability of personal and social information has been manipulated simultaneously and the relative impact on choice behavior assessed. Our study indicates that bees do not treat social and non-social information equivalently with respect to reliability: in our experiments, bees preferentially used social cues when these were reliable indicators of reward
Figure 2. Responses to Bee Presence and Trained Flower Color for All Reliability Treatment Combinations Left: all levels of floral cue reliability when social cue reliability = 50%. Middle: all levels of floral cue reliability when social cue reliability = 83%. Right: all levels of floral cue reliability when social cue reliability = 100%. Each point shows the mean ± SD for the ten bees tested in each treatment. Single asterisks indicate significant (p < 0.05) deviations from what is expected by chance (50%); double asterisks indicate points remaining statistically significant following Bonferroni corrections. See also Table S1.
yet avoided them when they were not. Social information was preferred to such an extent that when both the social and the floral cue were equally reliable, or even when the floral cue was more reliable but the social cues provided some information, bees used the social cue exclusively. The dominance of social information was not due to bees only being able to learn the social cue. During training, bees increased their performance for either
Figure 3. Visitation Rates during Training to Social and Floral Cues Based on the Reliability of Those Cues The main effects of social cue reliability and of floral clue reliability on overall performance during the training phase in terms of the proportion of rewarded landings (mean ± SE). For social reliability, each point is statistically significantly different from every other point (Fisher’s least significant difference [LSD], all p < 0.0019). For floral reliability, the mean number of rewarded landings per block when flowers are 100% reliable significantly differs from other levels (Fisher’s LSD, both p < 0.00008). See also Figure S2 and Table S2.
cue, particularly when that cue was reliable, independently of the reliability of the other cue. Thus, bumble bees were able to, and did, learn to use floral cues as well as social ones during training, and their differences in response to each type of cue in our later choice tests were not due to differences in information acquisition. Overall, social information should be more valuable than personal information when an individual is naive, uncertain, or less effective at gathering information than other individuals because it is less costly in time and energy to gather (e.g., [30–33]). Thus, the strong attraction to social cues in our study might reflect the fact that floral resources are ephemeral and must be found and exploited quickly. A conspecific bee at a novel flower type may commonly be a first and best indicator of an as-yet-untapped floral resource, and, indeed, closer tracking of changing environments is invoked as a benefit to using social information [34–36]. Consistent with this explanation are studies showing that bumble bees are more responsive to social cues when the floral resource at which conspecifics are found is a novel one [21, 26]. The strong preference for informative social cues might also support the common idea that social information is less costly to acquire and permits a forager to track a quickly changing environment without sampling resources directly [34, 36]. Although personal information may often be more reliable than social information in nature [37, 38], when social information is also reliable, this low cost would lead it to be preferred. Here, we have shown that bumble bees can evaluate the reliability of the cues they use when making foraging decisions and use this reliability to determine responses to cues, in this case deciding between social and personal information. A wide variety of other insects, fish, birds, and mammals have also been shown to discriminate between personal and social information, most often in a foraging context but also in other circumstances such as predator detection [2–5, 39]. The reliability of the personal and social information should affect these decisions as well. Nevertheless, we found that bees do treat social and personal information differently and are more sensitive to reliability of the former. Future studies will be needed to determine whether this difference in sensitivity is realized broadly across the spectrum of floral attributes (color, shape, odor, petal microtexture, etc.) and social contexts and whether a similar pattern is Current Biology 26, 1195–1199, May 9, 2016 1197
found in other animals engaging in other activities. Meanwhile, our results show that even relatively small-brained animals are able to make sophisticated decisions about which information to use based on its reliability. EXPERIMENTAL PROCEDURES In order to assess the effects of reliability of information on bee foraging, we constructed an array of artificial ‘‘flowers’’ in a small enclosed arena connected to a bumble bee nest. We then manipulated the reliability with which a floral cue (one of two colors of artificial flower) or a social cue (presence or absence of a pinned, dried conspecific bee) predicted a sucrose solution reward at a given flower. We defined the reliability of a cue as the conditional probability of sucrose solution being present at a flower, given the presence of the cue. We evaluated three levels of reliability for each cue: completely reliable (100%), moderately reliable (83%), and unreliable (50%) and all combinations thereof. We maintained a constant overall frequency of 50% rewarding flowers in each treatment, so the alternate cue was rewarded with the inverse frequency of the focal cue. We first allowed individual bees to interact with floral and social cues in their randomly assigned treatment combination and to collect sucrose solution for 100 flower visits. The solution was manually replenished in any emptied flowers immediately after the visit in order to maintain the statistical properties of reliability and reward, despite any potential revisits. After this experience, we gave bees a simple choice test, to evaluate their response to the floral and social cues. In the test, bees were allowed to search in an array in which flowers of each possible cue combination (each color without pinned bees and each color with pinned bees; Figure 1) were present with equal frequency. Bees were thus not forced to choose one cue at the expense of the other cue; they could potentially respond strongly to both. None of the test flowers offered sucrose solution reward. We estimated the bee’s response to the social cue as the proportion of landings on which a pinned bee was present. Similarly, we estimated the bee’s response to the floral cue as the proportion of landings that were on the trained color. See Supplemental Experimental Procedures for more details on the experimental design.
2. Kendal, R.L., Coolen, I., and Laland, K.N. (2004). The role of conformity in foraging when personal and social information conflict. Behav. Ecol. 15, 269–277. 3. Giraldeau, L.A., and Beauchamp, G. (1999). Food exploitation: searching for the optimal joining policy. Trends Ecol. Evol. 14, 102–106. 4. Webster, M.M., and Laland, K.N. (2011). Reproductive state affects reliance on public information in sticklebacks. Proc. Biol. Sci. 278, 619–627. 5. Rieucau, G., and Giraldeau, L.A. (2011). Exploring the costs and benefits of social information use: an appraisal of current experimental evidence. Philos. Trans. R. Soc. Lond. B Biol. Sci. 366, 949–957. 6. Searcy, W.A., and Nowicki, S. (2005). The Evolution of Animal Communication: Reliability and Deception in Signaling Systems (Princeton: Princeton University Press). 7. Bradbury, J.W., and Vehrencamp, S.L. (1998). Principles of Animal Communication (Sunderland: Sinauer Associates). 8. Blumstein, D.T., Verneyre, L., and Daniel, J.C. (2004). Reliability and the adaptive utility of discrimination among alarm callers. Proc. Biol. Sci. 271, 1851–1857. 9. Seyfarth, R.M., Cheney, D.L., Bergman, T., Fischer, J., Zuberbu¨hler, K., and Hammerschmidt, K. (2010). The central importance of information in studies of animal communication. Anim. Behav. 80, 3–8. 10. Dunlap, A.S., and Stephens, D.W. (2009). Components of change in the evolution of learning and unlearned preference. Proc. Biol. Sci. 276, 3201–3208. 11. Dunlap, A.S., and Stephens, D.W. (2014). Experimental evolution of prepared learning. Proc. Natl. Acad. Sci. USA 111, 11750–11755. 12. Raguso, R.A. (2004). Flowers as sensory billboards: progress towards an integrated understanding of floral advertisement. Curr. Opin. Plant Biol. 7, 434–440. 13. Leonard, A.S., Dornhaus, A., and Papaj, D.R. (2011). Forget-me-not: complex floral displays, inter-signal interactions, and pollinator cognition. Curr. Zool. 57, 215–244.
SUPPLEMENTAL INFORMATION
14. Giraldeau, L.A., Valone, T.J., and Templeton, J.J. (2002). Potential disadvantages of using socially acquired information. Philos. Trans. R. Soc. Lond. B Biol. Sci. 357, 1559–1566.
Supplemental Information includes Supplemental Experimental Procedures, two figures, and two tables and can be found with this article online at http://dx.doi.org/10.1016/j.cub.2016.03.009.
15. Hall, C.L., and Kramer, D.L. (2008). The economics of tracking a changing environment: competition and social information. Anim. Behav. 76, 1609– 1619.
AUTHOR CONTRIBUTIONS
16. Dukas, R. (2008). Evolutionary biology of insect learning. Annu. Rev. Entomol. 53, 145–160.
The experiment was designed by A.S.D., A.D., and D.R.P. Data were collected by A.S.D. Data were analyzed by A.S.D., D.R.P., and M.E.N. The paper was written by A.S.D., M.E.N., A.D., and D.R.P. ACKNOWLEDGMENTS We thank Brian Poulakidas, Zoe HarnEnz, Kevin Yee, and Eugene Choi for assistance in running the experiment. We also thank our four anonymous reviewers and the editor for their helpful comments and questions. This work was supported by an NIH-IRACDA-funded PERT fellowship to A.S.D. through the Center for Insect Science at the University of Arizona and by an NSF grant to D.R.P. and A.D. (IOS 0921280). Received: May 29, 2015 Revised: February 17, 2016 Accepted: March 2, 2016 Published: April 28, 2016 REFERENCES 1. Dall, S.R.X., Giraldeau, L.-A., Olsson, O., McNamara, J.M., and Stephens, D.W. (2005). Information and its use by animals in evolutionary ecology. Trends Ecol. Evol. 20, 187–193.
1198 Current Biology 26, 1195–1199, May 9, 2016
17. Laverty, T.M. (1994). Bumble bee learning and flower morphology. Anim. Behav. 47, 531–545. 18. Dornhaus, A., and Chittka, L. (2005). Bumble bees (Bombus terrestris) store both food and information in honeypots. Behav. Ecol. 16, 661–666. 19. Saleh, N., Scott, A.G., Bryning, G.P., and Chittka, L. (2007). Distinguishing signals and cues: bumblebees use general footprints to generate adaptive behaviour and flowers and nest. Arthropod-Plant Interact. 1, 119–127. 20. Baude, M., Dajoz, I., and Danchin, E. (2008). Inadvertent social information in foraging bumblebees: effects of flower distribution and implications for pollination. Anim. Behav. 76, 1863–1873. 21. Kawaguchi, L.G., Ohashi, K., and Toquenaga, Y. (2007). Contrasting responses of bumble bees to feeding conspecifics on their familiar and unfamiliar flowers. Proc. R. Soc. B 274, 2661–2667. 22. Kawaguchi, L.G., Ohashi, K., and Toquenaga, Y. (2006). Do bumble bees save time when choosing novel flowers by following conspecifics? Funct. Ecol. 20, 239–244. 23. Molet, M., Chittka, L., and Raine, N.E. (2009). How floral odours are learned inside the bumblebee (Bombus terrestris) nest. Naturwissenschaften 96, 213–219. 24. Jones, P.L., Ryan, M.J., and Chittkas, L. (2015). The influence of past experience with flower reward quality on social learning in bumblebees. Anim. Behav. 101, 11–18.
25. Worden, B.D., and Papaj, D.R. (2005). Flower choice copying in bumblebees. Biol. Lett. 1, 504–507.
33. Giraldeau, L.-A., and Dubois, F. (2008). Social foraging and the study of exploitative behavior. Adv. Stud. Behav. 38, 59–104.
26. Leadbeater, E., and Chittka, L. (2005). A new mode of information transfer in foraging bumblebees? Curr. Biol. 15, R447–R448.
34. Krebs, J.R., and Inman, A.J. (1992). Learning and foraging: individuals, groups, and populations. Am. Nat. 140 (Suppl 1 ), S63–S84.
27. Leadbeater, E., and Chittka, L. (2007). The dynamics of social learning in an insect model, the bumblebee (Bombus terrestris). Behav. Ecol. Sociobiol. 61, 1789–1796.
35. Giraldeau, L.A., and Caraco, T. (2000). Social Foraging Theory (Princeton: Princeton University Press).
28. Leadbeater, E., and Chittka, L. (2009). Bumble-bees learn the value of social cues through experience. Biol. Lett. 5, 310–312.
36. Galef, B.G., Jr., and Giraldeau, L.A. (2001). Social influences on foraging in vertebrates: causal mechanisms and adaptive functions. Anim. Behav. 61, 3–15.
29. Plowright, C.M.S., Ferguson, K.A., Jellen, S.L., Xu, V., Service, E.W., and Dookie, A.L. (2013). Bumblebees occupy: when foragers do and do not use the presence of others to first find food. Insectes Soc. 60, 517–524. 30. Laland, K.N. (2004). Social learning strategies. Learn. Behav. 32, 4–14.
37. Boyd, R., and Richerson, P.J. (1985). Culture and the Evolutionary Process (Chicago: University of Chicago Press).
31. Danchin, E., Giraldeau, L.A., Valone, T.J., and Wagner, R.H. (2004). Public information: from nosy neighbors to cultural evolution. Science 305, 487–491.
38. Boyd, R., and Richerson, P.J. (1988). An evolutionary model of social learning: the effects of spatial and temporal variation. In Social Learning: Psychological and Biological Perspectives, T.R. Zentall, and B.G.J. Galef, eds. (Hillsdale: Lawrence Erlbaum Associates), pp. 29–48.
32. Giraldeau, L.-A., Caraco, T., and Valone, T.J. (1994). Social foraging: individual learning and cultural transmission of innovations. Behav. Ecol. 5, 35–43.
39. Barrera, J.P., Chong, L., Judy, K.N., and Blumstein, D.T. (2011). Reliability of public information: predators provide more information about risk than conspecifics. Anim. Behav. 81, 779–787.
Current Biology 26, 1195–1199, May 9, 2016 1199
Foraging Bumble Bees Weigh the Reliability of Personal and Social Information Highlights
Authors
d
Foraging bumble bees use the reliability of both social and personal information
Aimee S. Dunlap, Matthew E. Nielsen, Anna Dornhaus, Daniel R. Papaj
d
If social information is at all reliable, it is preferred over personal
Correspondence
d
Both personal and social information are successfully learned when reliable
[email protected] (A.S.D.), [email protected] (M.E.N.)
In Brief Dunlap et al. find that bumble bees use the reliability of information gained from both personal experience and other bees to make foraging decisions. They are especially sensitive to reliability of social information and trust others above themselves when information from other bees is reliable, even if personal information is more reliable.
Dunlap et al., 2016, Current Biology 26, 1195–1199 May 9, 2016 ª 2016 Elsevier Ltd. http://dx.doi.org/10.1016/j.cub.2016.03.009
Current Biology
Report Foraging Bumble Bees Weigh the Reliability of Personal and Social Information Aimee S. Dunlap,1,2,* Matthew E. Nielsen,1,* Anna Dornhaus,1 and Daniel R. Papaj1 1Department of Ecology and Evolutionary Biology, Center for Insect Science, University of Arizona, 1041 E Lowell Street, Tucson, AZ 85721, USA 2Department of Biology, University of Missouri-St. Louis, One University Boulevard, St. Louis, MO 63121, USA *Correspondence: [email protected] (A.S.D.), [email protected] (M.E.N.) http://dx.doi.org/10.1016/j.cub.2016.03.009
SUMMARY
Many animals, including insects, make decisions using both personally gathered information and social information derived from the behavior of other, usually conspecific, individuals [1]. Moreover, animals adjust use of social versus personal information appropriately under a variety of experimental conditions [2–5]. An important factor in how information is used is the information’s reliability, that is, how consistently the information is correlated with something of relevance in the environment [6]. The reliability of information determines which signals should be attended to during communication [6–9], which types of stimuli animals should learn about, and even whether learning should evolve [10, 11]. Here, we show that bumble bees (Bombus impatiens) account for the reliability of personally acquired information (which flower color was previously associated with reward) and social information (which flowers are chosen by other bees) in making foraging decisions; however, the two types of information are not treated equally. Bees prefer to use social information if it predicts a reward at all, but if social information becomes entirely unreliable, flower color will be used instead. This greater sensitivity to the reliability of social information, and avoidance of conspecifics in some cases, may reflect the specific ecological circumstances of bee foraging. Overall, the bees’ ability to make decisions based on both personally acquired and socially derived information, and the relative reliability of both, demonstrates a new level of sophistication and flexibility in animal, particularly insect, decision-making. RESULTS The natural environment encountered by pollinator species such as bees offers a multitude of cues that may be used to identify resources, all of which will vary in the reliability with which they indicate reward. Diverse environmental cues such as the color, shape, and other attributes of flowers can be used to distinguish species offering accessible, high-quality reward from less-
rewarding or non-rewarding species [12, 13]. However, several factors reduce reliability of these cues and thus personally gathered information: the reward of individual plants of the same species may vary due to microhabitat differences; some flowers may have been emptied by competing foragers, and some floral cues may be more difficult to detect and discriminate among than others, resulting in a change in the reliability of these cues from the point of view of the forager [12, 13]. Social information for pollinators can also vary in reliability. When a floral resource is abundant, conspecifics can provide a good indicator of reward presence; however, conspecifics may themselves make lessthan-perfect choices in resources. Even if a conspecific is visiting the best resource, that resource has now likely decreased in value because it has been at least partially exploited by the same conspecific providing the cue [14, 15]. An individual pollinator’s foraging success may thus depend on its ability to assess the relative reliability of the two forms of information and to respond appropriately. Nectar-foraging bumble bees provide an excellent system for studying the use of personal and social information. Bumble bees forage on flowers, an ephemeral resource, often in the company of other bees; both personal and social information can be used by bees to make floral choices (reviewed in [16]). In terms of personal information, bumble bees demonstrate individual trial-and-error learning, using a variety of floral cues associated with nectar reward, including color, odor, shape, texture, and spatial location [17]. In addition to personal learning of floral cues, bumble bees also use a variety of social cues to inform their flower choices. Social cues range from direct cues such as the sight of conspecifics on particular flowers or the odors left in a conspecific’s ‘‘footprints’’ to indirect cues such as floral scents carried into the nest [18–24]. Social information can even facilitate individual learning of floral cues by bumble bees [25– 27]. Finally, bumble bees show natural responses to artificially manipulated representations of floral and social cues (e.g., [25–29]), enabling us to control the reliability of floral (= personal) and social cues independently and precisely. To assess how the reliability of personal versus social information affects their use by bumble bees during foraging, we trained bees on arrays of artificial flowers in which we independently manipulated flower color (a floral cue providing personal information), the presence or absence of a pinned conspecific (a social cue providing social information), and the presence of a concealed sucrose reward (Figure 1). Each cue, floral or social, was made a fully reliable, moderately reliable, or unreliable indicator of the sucrose reward, in a fully factorial design (Table 1).
Current Biology 26, 1195–1199, May 9, 2016 ª 2016 Elsevier Ltd. 1195
Table 1. Predictions and Outcomes of Bee Preferences for Flowers Associated with Floral and Social Cues as a Function of Their Reliability Social Reliability 50% Floral Reliability
83%
100%
50% predict: neither predict: social predict: social result: neither U result: social U result: social U 83% predict: floral result: floral U 100% predict: floral result: floral U
predict: either predict: social result: social result: social U but not floral U predict: floral result: social X
predict: either result: social but not floral U
Checkmark (U) indicates that result met prediction; X indicates that result did not meet prediction.
Figure 1. Diagram of the Colored-Disc Artificial Flowers and Dried Pinned Bumble Bees Used as Social Cues The final testing array consisted of three of each of these types, for a total of 12 flowers. See also Figure S1.
Under these conditions, bees making perfect economic choices should use the most reliable information source, whether floral or social. They may also use the other source if it can provide any increase in total reliability, but they should never pay attention to unreliable sources of information. We tested these hypotheses by measuring the preferences of bumble bees on a non-rewarding array after training, finding that bees do not use unreliable information sources, personal or social, but favor social information over personal information so long as it has at least some reliability. Effect of Reliability on Information Use The preference during the test of bees for the floral versus social cue generally depended on their relative reliability during training as an indicator of sucrose reward (Table 1; Figure 1). In five of six treatments where one type of cue was more reliable than the other, the more reliable cue was preferred (Table 1). Nevertheless, bees did not respond equivalently to the reliability of the two types of cue. The response to the reliability of the social cue was strong and consistent. When the social cue was unreliable (reliability = 50%, i.e., a random chance of indicating food), bees never chose flowers bearing that cue more often than random (Table 1; Figure 2). In an extreme case, when social cues were unreliable (50%) and floral cue reliability moderate (83%), bees actually avoided the social cue (Figure 2A; t9 = 2.32, p = 0.02), which is not predicted in terms of reliability alone. On the other hand, when the social cue provided any reliable information (reliability R 83%), the response to the social cue was always significantly greater than 0.5 (Table 1; Figures 1B and 1C), meaning bees sought out flowers with conspecifics. Overall, these patterns gave rise to a strongly significant effect of social cue reliability on bee preference (repeated-measures ANOVA, F2,81 = 5.362, p < 0.0065; Table S1), as well as an interaction between social cue reliability and the focal cue type being measured (repeated-measures ANOVA, F2,81 = 17.026, p < 0.00001; Table S1). 1196 Current Biology 26, 1195–1199, May 9, 2016
In contrast, the bees’ response to the reliability of the floral cue was complex. When the floral cue was unreliable (50%), bees never chose flowers bearing that cue more often than random (Table 1; Figure 2), consistent with expectations. When the reliability of the floral cue was R83%, though, the response to the floral cue depended on the reliability of the social cue. If the social cue was unreliable (50%), bees responded significantly to a reliable floral cue (Figure 2A). However, when the social cue was reliable (83% or 100%), the floral cue was ignored, even when it was more reliable than the social cue (Figures 2B and 2C). Overall, the effect of floral cue reliability alone was not statistically significant (F2,81 = 0.937, p = 0.396) but interacted significantly with the effect of social cue reliability and the effect of focal cue type (social reliability 3 floral reliability 3 cue type interaction, F4,81 = 3.223, p = 0.0166; Table S1). Performance during Training We also examined how bees performed during the training phase by tabulating the number of rewarded visits across each of ten blocks of ten trials each. Bees showed improvement in terms of choosing rewarded flowers over blocks of trials (repeated-measures ANOVA, F9,729 = 5.84, p < 0.000001; Figure S2). Both social cue reliability and floral cue reliability had significant effects on how successful bees were at obtaining reward while foraging during training (F2,81 = 8.067, p = 0.0006 and F2,81 = 4.36, p = 0.016 for social and floral reliability, respectively; Figure 3). These results suggest that bees learned to use each of the cues. None of the interactions among block, floral cue reliability, and social cue reliability effects were statistically significant (Figure S2; Table S2). DISCUSSION We have shown here that bumble bees assess and react to the reliability of both personally acquired information and information acquired from conspecifics. To our knowledge, this is the first time that the reliability of personal and social information has been manipulated simultaneously and the relative impact on choice behavior assessed. Our study indicates that bees do not treat social and non-social information equivalently with respect to reliability: in our experiments, bees preferentially used social cues when these were reliable indicators of reward
Figure 2. Responses to Bee Presence and Trained Flower Color for All Reliability Treatment Combinations Left: all levels of floral cue reliability when social cue reliability = 50%. Middle: all levels of floral cue reliability when social cue reliability = 83%. Right: all levels of floral cue reliability when social cue reliability = 100%. Each point shows the mean ± SD for the ten bees tested in each treatment. Single asterisks indicate significant (p < 0.05) deviations from what is expected by chance (50%); double asterisks indicate points remaining statistically significant following Bonferroni corrections. See also Table S1.
yet avoided them when they were not. Social information was preferred to such an extent that when both the social and the floral cue were equally reliable, or even when the floral cue was more reliable but the social cues provided some information, bees used the social cue exclusively. The dominance of social information was not due to bees only being able to learn the social cue. During training, bees increased their performance for either
Figure 3. Visitation Rates during Training to Social and Floral Cues Based on the Reliability of Those Cues The main effects of social cue reliability and of floral clue reliability on overall performance during the training phase in terms of the proportion of rewarded landings (mean ± SE). For social reliability, each point is statistically significantly different from every other point (Fisher’s least significant difference [LSD], all p < 0.0019). For floral reliability, the mean number of rewarded landings per block when flowers are 100% reliable significantly differs from other levels (Fisher’s LSD, both p < 0.00008). See also Figure S2 and Table S2.
cue, particularly when that cue was reliable, independently of the reliability of the other cue. Thus, bumble bees were able to, and did, learn to use floral cues as well as social ones during training, and their differences in response to each type of cue in our later choice tests were not due to differences in information acquisition. Overall, social information should be more valuable than personal information when an individual is naive, uncertain, or less effective at gathering information than other individuals because it is less costly in time and energy to gather (e.g., [30–33]). Thus, the strong attraction to social cues in our study might reflect the fact that floral resources are ephemeral and must be found and exploited quickly. A conspecific bee at a novel flower type may commonly be a first and best indicator of an as-yet-untapped floral resource, and, indeed, closer tracking of changing environments is invoked as a benefit to using social information [34–36]. Consistent with this explanation are studies showing that bumble bees are more responsive to social cues when the floral resource at which conspecifics are found is a novel one [21, 26]. The strong preference for informative social cues might also support the common idea that social information is less costly to acquire and permits a forager to track a quickly changing environment without sampling resources directly [34, 36]. Although personal information may often be more reliable than social information in nature [37, 38], when social information is also reliable, this low cost would lead it to be preferred. Here, we have shown that bumble bees can evaluate the reliability of the cues they use when making foraging decisions and use this reliability to determine responses to cues, in this case deciding between social and personal information. A wide variety of other insects, fish, birds, and mammals have also been shown to discriminate between personal and social information, most often in a foraging context but also in other circumstances such as predator detection [2–5, 39]. The reliability of the personal and social information should affect these decisions as well. Nevertheless, we found that bees do treat social and personal information differently and are more sensitive to reliability of the former. Future studies will be needed to determine whether this difference in sensitivity is realized broadly across the spectrum of floral attributes (color, shape, odor, petal microtexture, etc.) and social contexts and whether a similar pattern is Current Biology 26, 1195–1199, May 9, 2016 1197
found in other animals engaging in other activities. Meanwhile, our results show that even relatively small-brained animals are able to make sophisticated decisions about which information to use based on its reliability. EXPERIMENTAL PROCEDURES In order to assess the effects of reliability of information on bee foraging, we constructed an array of artificial ‘‘flowers’’ in a small enclosed arena connected to a bumble bee nest. We then manipulated the reliability with which a floral cue (one of two colors of artificial flower) or a social cue (presence or absence of a pinned, dried conspecific bee) predicted a sucrose solution reward at a given flower. We defined the reliability of a cue as the conditional probability of sucrose solution being present at a flower, given the presence of the cue. We evaluated three levels of reliability for each cue: completely reliable (100%), moderately reliable (83%), and unreliable (50%) and all combinations thereof. We maintained a constant overall frequency of 50% rewarding flowers in each treatment, so the alternate cue was rewarded with the inverse frequency of the focal cue. We first allowed individual bees to interact with floral and social cues in their randomly assigned treatment combination and to collect sucrose solution for 100 flower visits. The solution was manually replenished in any emptied flowers immediately after the visit in order to maintain the statistical properties of reliability and reward, despite any potential revisits. After this experience, we gave bees a simple choice test, to evaluate their response to the floral and social cues. In the test, bees were allowed to search in an array in which flowers of each possible cue combination (each color without pinned bees and each color with pinned bees; Figure 1) were present with equal frequency. Bees were thus not forced to choose one cue at the expense of the other cue; they could potentially respond strongly to both. None of the test flowers offered sucrose solution reward. We estimated the bee’s response to the social cue as the proportion of landings on which a pinned bee was present. Similarly, we estimated the bee’s response to the floral cue as the proportion of landings that were on the trained color. See Supplemental Experimental Procedures for more details on the experimental design.
2. Kendal, R.L., Coolen, I., and Laland, K.N. (2004). The role of conformity in foraging when personal and social information conflict. Behav. Ecol. 15, 269–277. 3. Giraldeau, L.A., and Beauchamp, G. (1999). Food exploitation: searching for the optimal joining policy. Trends Ecol. Evol. 14, 102–106. 4. Webster, M.M., and Laland, K.N. (2011). Reproductive state affects reliance on public information in sticklebacks. Proc. Biol. Sci. 278, 619–627. 5. Rieucau, G., and Giraldeau, L.A. (2011). Exploring the costs and benefits of social information use: an appraisal of current experimental evidence. Philos. Trans. R. Soc. Lond. B Biol. Sci. 366, 949–957. 6. Searcy, W.A., and Nowicki, S. (2005). The Evolution of Animal Communication: Reliability and Deception in Signaling Systems (Princeton: Princeton University Press). 7. Bradbury, J.W., and Vehrencamp, S.L. (1998). Principles of Animal Communication (Sunderland: Sinauer Associates). 8. Blumstein, D.T., Verneyre, L., and Daniel, J.C. (2004). Reliability and the adaptive utility of discrimination among alarm callers. Proc. Biol. Sci. 271, 1851–1857. 9. Seyfarth, R.M., Cheney, D.L., Bergman, T., Fischer, J., Zuberbu¨hler, K., and Hammerschmidt, K. (2010). The central importance of information in studies of animal communication. Anim. Behav. 80, 3–8. 10. Dunlap, A.S., and Stephens, D.W. (2009). Components of change in the evolution of learning and unlearned preference. Proc. Biol. Sci. 276, 3201–3208. 11. Dunlap, A.S., and Stephens, D.W. (2014). Experimental evolution of prepared learning. Proc. Natl. Acad. Sci. USA 111, 11750–11755. 12. Raguso, R.A. (2004). Flowers as sensory billboards: progress towards an integrated understanding of floral advertisement. Curr. Opin. Plant Biol. 7, 434–440. 13. Leonard, A.S., Dornhaus, A., and Papaj, D.R. (2011). Forget-me-not: complex floral displays, inter-signal interactions, and pollinator cognition. Curr. Zool. 57, 215–244.
SUPPLEMENTAL INFORMATION
14. Giraldeau, L.A., Valone, T.J., and Templeton, J.J. (2002). Potential disadvantages of using socially acquired information. Philos. Trans. R. Soc. Lond. B Biol. Sci. 357, 1559–1566.
Supplemental Information includes Supplemental Experimental Procedures, two figures, and two tables and can be found with this article online at http://dx.doi.org/10.1016/j.cub.2016.03.009.
15. Hall, C.L., and Kramer, D.L. (2008). The economics of tracking a changing environment: competition and social information. Anim. Behav. 76, 1609– 1619.
AUTHOR CONTRIBUTIONS
16. Dukas, R. (2008). Evolutionary biology of insect learning. Annu. Rev. Entomol. 53, 145–160.
The experiment was designed by A.S.D., A.D., and D.R.P. Data were collected by A.S.D. Data were analyzed by A.S.D., D.R.P., and M.E.N. The paper was written by A.S.D., M.E.N., A.D., and D.R.P. ACKNOWLEDGMENTS We thank Brian Poulakidas, Zoe HarnEnz, Kevin Yee, and Eugene Choi for assistance in running the experiment. We also thank our four anonymous reviewers and the editor for their helpful comments and questions. This work was supported by an NIH-IRACDA-funded PERT fellowship to A.S.D. through the Center for Insect Science at the University of Arizona and by an NSF grant to D.R.P. and A.D. (IOS 0921280). Received: May 29, 2015 Revised: February 17, 2016 Accepted: March 2, 2016 Published: April 28, 2016 REFERENCES 1. Dall, S.R.X., Giraldeau, L.-A., Olsson, O., McNamara, J.M., and Stephens, D.W. (2005). Information and its use by animals in evolutionary ecology. Trends Ecol. Evol. 20, 187–193.
1198 Current Biology 26, 1195–1199, May 9, 2016
17. Laverty, T.M. (1994). Bumble bee learning and flower morphology. Anim. Behav. 47, 531–545. 18. Dornhaus, A., and Chittka, L. (2005). Bumble bees (Bombus terrestris) store both food and information in honeypots. Behav. Ecol. 16, 661–666. 19. Saleh, N., Scott, A.G., Bryning, G.P., and Chittka, L. (2007). Distinguishing signals and cues: bumblebees use general footprints to generate adaptive behaviour and flowers and nest. Arthropod-Plant Interact. 1, 119–127. 20. Baude, M., Dajoz, I., and Danchin, E. (2008). Inadvertent social information in foraging bumblebees: effects of flower distribution and implications for pollination. Anim. Behav. 76, 1863–1873. 21. Kawaguchi, L.G., Ohashi, K., and Toquenaga, Y. (2007). Contrasting responses of bumble bees to feeding conspecifics on their familiar and unfamiliar flowers. Proc. R. Soc. B 274, 2661–2667. 22. Kawaguchi, L.G., Ohashi, K., and Toquenaga, Y. (2006). Do bumble bees save time when choosing novel flowers by following conspecifics? Funct. Ecol. 20, 239–244. 23. Molet, M., Chittka, L., and Raine, N.E. (2009). How floral odours are learned inside the bumblebee (Bombus terrestris) nest. Naturwissenschaften 96, 213–219. 24. Jones, P.L., Ryan, M.J., and Chittkas, L. (2015). The influence of past experience with flower reward quality on social learning in bumblebees. Anim. Behav. 101, 11–18.
25. Worden, B.D., and Papaj, D.R. (2005). Flower choice copying in bumblebees. Biol. Lett. 1, 504–507.
33. Giraldeau, L.-A., and Dubois, F. (2008). Social foraging and the study of exploitative behavior. Adv. Stud. Behav. 38, 59–104.
26. Leadbeater, E., and Chittka, L. (2005). A new mode of information transfer in foraging bumblebees? Curr. Biol. 15, R447–R448.
34. Krebs, J.R., and Inman, A.J. (1992). Learning and foraging: individuals, groups, and populations. Am. Nat. 140 (Suppl 1 ), S63–S84.
27. Leadbeater, E., and Chittka, L. (2007). The dynamics of social learning in an insect model, the bumblebee (Bombus terrestris). Behav. Ecol. Sociobiol. 61, 1789–1796.
35. Giraldeau, L.A., and Caraco, T. (2000). Social Foraging Theory (Princeton: Princeton University Press).
28. Leadbeater, E., and Chittka, L. (2009). Bumble-bees learn the value of social cues through experience. Biol. Lett. 5, 310–312.
36. Galef, B.G., Jr., and Giraldeau, L.A. (2001). Social influences on foraging in vertebrates: causal mechanisms and adaptive functions. Anim. Behav. 61, 3–15.
29. Plowright, C.M.S., Ferguson, K.A., Jellen, S.L., Xu, V., Service, E.W., and Dookie, A.L. (2013). Bumblebees occupy: when foragers do and do not use the presence of others to first find food. Insectes Soc. 60, 517–524. 30. Laland, K.N. (2004). Social learning strategies. Learn. Behav. 32, 4–14.
37. Boyd, R., and Richerson, P.J. (1985). Culture and the Evolutionary Process (Chicago: University of Chicago Press).
31. Danchin, E., Giraldeau, L.A., Valone, T.J., and Wagner, R.H. (2004). Public information: from nosy neighbors to cultural evolution. Science 305, 487–491.
38. Boyd, R., and Richerson, P.J. (1988). An evolutionary model of social learning: the effects of spatial and temporal variation. In Social Learning: Psychological and Biological Perspectives, T.R. Zentall, and B.G.J. Galef, eds. (Hillsdale: Lawrence Erlbaum Associates), pp. 29–48.
32. Giraldeau, L.-A., Caraco, T., and Valone, T.J. (1994). Social foraging: individual learning and cultural transmission of innovations. Behav. Ecol. 5, 35–43.
39. Barrera, J.P., Chong, L., Judy, K.N., and Blumstein, D.T. (2011). Reliability of public information: predators provide more information about risk than conspecifics. Anim. Behav. 81, 779–787.
Current Biology 26, 1195–1199, May 9, 2016 1199