- Email: [email protected]
Food restriction and threat of predation affect visual pattern choices by flower-naïve bumblebees
Learning and Motivation 50 (2015) 3–10
Contents lists available at ScienceDirect
Learning and Motivation journal homepage: www.elsevier.com/locate/l&m
Food restriction and threat of predation affect visual pattern choices by flower-naïve bumblebees E.W. Service, C.M.S. Plowright ∗ School of Psychology, University of Ottawa, Ottawa, Ont. K1N 6N5, Canada
a r t i c l e
i n f o
Article history: Available online 9 March 2015
Keywords: Bumblebees Floral choice Carbon dioxide Social cognition Pattern recognition Predation threat
a b s t r a c t The aim of this study was to determine whether a preference by flower-naïve bumblebees could be created or enhanced by manipulating variables relevant to food collection and to defense against predation. In two experiments, colonies of bumblebees (Bombus impatiens) were deprived of pollen, exposed to CO2 , or neither. Choices of individual workers in a radial arm maze were monitored. In Experiment 1, both variables lead to a preference for corridors occupied by a conspecific bee. The effect was specific: no change in preference for corridors occupied by other objects (a coin and a piece of Styrofoam) was detected. In Experiment 2, radial and concentric patterns were used, both of which were unoccupied. Only pollen deprivation increased preference for radial stimuli, while CO2 had no discernible effect. Preferences for visual patterns by bees leaving their colony for the first time are modulated by variables that affect the internal state of the bees in problem-specific ways. © 2015 Elsevier Inc. All rights reserved.
The possibility that foraging bees use the presence of other insects on flowers to govern their own floral choices has received little if any support in the field. For instance, one survey of counts of floral visitors found on flower heads of sunflowers (Helianthus annuus L.) and onion flowers (Allium cepa L.) addressed the question of whether there would be attraction or avoidance of inflorescences that were already occupied by one or more individuals. The distribution of insects, including honeybees (Apis mellifera L.) and bumblebees (Bombus spp. Latreille), followed a Poisson distribution (Tepedino & Parker, 1981): there were neither more nor less flowers with just one forager than would be expected by chance. The notion that insects forage independently of other insects may, however, have been dismissed prematurely by field biologists. Recently, an examination of the details of social effects on foraging in the laboratory has revealed that bees use associative learning to take advantage of cues that predict important outcomes such as presence of floral reward (Avarguès-Weber & Chittka, 2014a; Dawson, Avarguès-Weber, Chittka, & Leadbeater, 2013; Leadbeater & Chittka, 2009) and presence of predators (Dawson & Chittka, 2014)—in general, social information is used by insects strategically (Grüter & Leadbeater, 2014). Moreover, workers that have just left their colony for the first time also show a significant preference for flowers that are already occupied by another forager (Kawaguchi, Ohashi, & Toquenaga, 2006; Leadbeater & Chittka, 2009), though this effect is most evident when the flowers are rare and the occupiers are large relative to the flowers (Plowright et al., 2013). Because of the difficulty in obtaining the pre-experimental histories of bees seen foraging in the field, these sorts of effects may have eluded detection in nature. In this paper, our focus is on the behaviour of “flower-naïve” bumblebee workers: bees that leave their colony for the first time, and as such, have had no prior experience with flowers. One view regarding the preferences for occupied flowers
∗ Corresponding author. E-mail address: [email protected] (C.M.S. Plowright). http://dx.doi.org/10.1016/j.lmot.2014.10.006 0023-9690/© 2015 Elsevier Inc. All rights reserved.
4
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
by flower-naïve bees is that they may help inexperienced bees to locate sources of food: a pattern with another bee on it is likely to be a flower. One difficulty with this interpretation is that it is post hoc. Indeed, the exact opposite argument can be made a priori: a flower that is occupied by a forager is likely to be empty or well on its way to being depleted, and so it ought not to be preferred but avoided. In other words, other foragers may possibly act as competitors rather than as informers (Baude, Danchin, Mugabo, & Dajoz, 2011). Another view is that the presence of an occupier on a flower has little to do with signalling resource availability. Field observations have suggested that interactions among bees on a flower are in fact aggressive (Kikuchi, 1963). Indeed, in the course of a prior study on pattern preferences of flower-naïve bees (Orbán & Plowright, 2013), we captured a few such interactions on film (Videos 1 and 2 in the Supplementary Materials): the behaviours of the bee landing on the flower seemed more directed at the occupier than at the flower, which suggests that the occupier might have been perceived as a predator or a competitor. In two experiments, we manipulated internal states of flower-naïve bees. To promote food finding behaviours, we manipulated the availability of pollen, which is needed for feeding to larvae. To promote aggressive or defensive behaviours, we exposed the bees to CO2 to simulate the presence of a mammalian predator. Predators of bumblebee colonies (Goulson, 2010) include mice (Mus domesticus), badgers (Meles meles L.) in Europe, and skunks (Mephitis mephitis Schreber) in NorthAmerica. Bumblebees perceive CO2 : they respond to mammalian breath and to currents of air containing 5 or 10% CO2 by hissing, which serves as an inter-specific defence signal (Kirchner & Röschard, 1999). If a preference for occupied stimuli is engaged when bees are food searching, then pollen deprivation should create or increase the preference. If the preference is engaged in situations that trigger aggressive behaviours, then exposure to CO2 should create or increase the preference. Experiment 1 tested these two predictions. In addition, by comparing the preference for stimuli occupied by another bee with the preference for stimuli occupied by non-organic objects (a coin, as in Dawson and Chittka (2012), and a piece of Styrofoam), we began to address the question of whether any preference for occupied stimuli was indeed social. Aggressive tendencies ought to be directed at other individuals and not at flowers themselves. Heightening an aggressive tendency ought not to increase a preference for a pattern that is more “floral” than another. To determine whether the effects of our variables were specific to social preferences, Experiment 2 examined their effects on preference for floral patterns with no occupiers. Given that radial patterns (illustrated in the legend of Fig. 3) are, by and large, preferred over concentric patterns (Orbán & Plowright, 2013) as they are thought to resemble flowers in nature (Lehrer, Horridge, Zhang, & Gadagkar, 1995), we examined the effect of food deprivation and CO2 exposure on relative choice of these patterns. Food deprivation ought, if anything, to increase preference for radial patterns, while CO2 should not. Given that larger bumblebees tend to invest themselves in foraging duties while smaller bees tend to the nest (Goulson, 2010), though task specialization is not as marked as in honeybees, we reasoned that the effect of food deprivation might interact with body mass. Methods Subjects Three commercial colonies of Bombus impatiens Cresson in plastic nest boxes (19.5 cm × 17.5 cm × 12 cm high) were supplied by Koppert Canada. Because the colonies were covered with an opaque lid, they received little light, as in nature. In Experiment 1, 48 bees from each of two colonies were used, for a total of 96. In Experiment 2, 60 bees from the third colony were used. All bees were tested the first time they left the colony: they had no pre-experimental experience outside of the nest box. In all conditions, to motivate the bees to exit the colony, the wick that absorbed sugar solution from a plastic bag beneath the nest box was capped for one or two days prior to testing. Apparatus The 12-corridor maze that we used, modelled on that of Lehrer et al. (1995), is diagrammed in Fig. 1. Photographs are shown in Plowright, Evans, Chew Leung, and Collin (2011, Fig. 1). It was constructed of grey Plexiglas® with a clear cover. From the central area (22 cm wide, 15 cm high), 12 corridors radiated outwards. The bees entered the maze via a screen tube that connected their colony exit hole to an entrance hole in the center of the floor of the maze. The exit was stoppered when the colony was not in use during the experiment: bees were not allowed to travel to and from the apparatus prior to testing. The entrance to the maze was gated during testing sessions: bees were let into the maze individually. The corridors were 6 cm wide at their entrance from the center of the maze and 15 cm long. The back walls were 13 cm wide. The maze was illuminated from above by high-frequency (>40 kHz) lighting equipment (three Sylvania Quicktronic T8 QHE4x32T8/112 light ballasts, each with four Sylvania model FO32/841/XP/SS/EC03 fluorescent light bulbs). Stimuli In both experiments, the stimuli were mounted with Velcro® on the back walls of the corridors in the maze, as shown in Fig. 1.
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
5
Fig. 1. Radial maze and connected hive box. Patterns were mounted vertically on the walls of the corridors in the positions marked with a red circle. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.) Source: Reproduced from Plowright et al. (2013), with kind permission from Springer Science and Business Media.
Experiment 1 Nine corridors in the maze were empty, that is, “Unoccupied”. The remaining three corridors enclosed a stimulus, so they were “Occupied”. In one of these three corridors, a dead pinned queen B. impatiens was attached to the back wall (“OccupiedConspecific”). Bumblebees are attracted to artificial flowers where a model bee or a dead bee have been pinned in much the same way they are attracted to artificial flowers where other bees are already foraging (Leadbeater & Chittka, 2007), though there is no claim here that motion cues are irrelevant in general (indeed, they are important in defensive behaviour of honeybees (Free, 1961) and in social learning of foraging behaviour by bumblebees (Avarguès-Weber & Chittka, 2014a,b). Accordingly, we used a dead pinned queen bee as in Plowright et al. (2013) to maximize the chances that it would be detected by virtue of its larger size. In nature, the times at which workers and queens forage within a season overlap and so workers could well encounter other queens. A Canadian dime (“Occupied-Coin”) the colour of steel was placed in the second corridor, and a piece of white Styrofoam® (“Occupied-Styrofoam”) of the same diameter as the coin (18 mm) was placed in the third corridor. Experiment 2 Two opposing corridors were blocked off and the remaining ten contained circular (6 cm diameter) stimuli: three radial patterns consisting of four black and four white wedges, and seven concentric patterns consisting of two black and two white circles. The patterns were printed on paper in black and white and then laminated. Design Experiment 1 For each of the following four conditions, 12 bees from each of the two colonies were used: (1) “Baseline”: the colony had ad libitum access to a ball of pollen paste placed beside the cocoons (pollen purchased from an apiarist, crushed and mixed with honey and water), though access to sugar solution was prevented for two days prior to testing so that the bees would be motivated to exit the colony. (2) “Pollen deprivation”: to increase the level of motivation to forage over and above the level in baseline, not only was there deprivation of sugar solution, but pollen deprivation was superimposed, with the pollen ball being removed for 24 h prior to testing. (3) “Strong CO2 ”: as with baseline, but with the addition of a steady stream into the maze of a mixture of CO2 and O2 (1:3 by volume) for the duration of the time spent in the maze by each bee and (4) “Pure CO2 ”: pure carbon dioxide gas (though the maze was not gas tight so there was some dilution). In the last two conditions, the gas was released directly into the maze through an aperture in the lid so as to ensure that individual bees were under
6
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
the influence of CO2 as they made their pattern choices. The rate of gas flow was 1.1–1.3 L/min. For the first colony used, the conditions were run in order 1–4, and for the second colony, the reverse order was used, so that any overall effect of experimental condition could not be attributed to colony aging. Three unoccupied corridors separated each of the three occupied corridors. The positions of the three types of occupied corridors (conspecific, coin and Styrofoam) were counterbalanced so that across 12 bees, they appeared equally often in each of the 12 corridors in the maze. Experiment 2 For each of the following three conditions, 20 bees were used: (1) baseline, (2) pure CO2 and (3) pollen deprivation. These conditions, which were run in order, are described above for Experiment 1. The positions of the patterns were counterbalanced so that across 20 bees in each condition, the three radial patterns appeared six times and the seven concentric patterns appeared fourteen times in each of the ten corridors used. Procedure The procedure was common to both experiments. Once a worker exited the colony and entered the maze for the first time (where the first time was also the last time), its first 20 choices were recorded as in Plowright et al. (2013). A choice was defined as making contact with the back wall of the corridor or the stimulus attached to it. A new choice was only recorded once the bee had crossed the threshold of the corridor and returned to the center of the maze (in other words, ‘dithering’ within a corridor did not count as multiple choices). Once 20 choices were made, the bee was removed from the apparatus and killed, after which the bees in Experiment 2 were weighted on a Mettler Toledo balance model PB 303-S. In the Pure CO2 condition, some bees (n = 7 in Experiment 1 and n = 4 in Experiment 2) became inactive for over 15 min with exposure to CO2 and did not complete 20 choices—they were excluded from the data set. Testing for each condition took place for 4–8 h per day, over the course of 1–3 days. The maze was cleaned with a damp cloth between subjects to minimize any visual evidence of a corridor having been visited. Though bumblebees can learn whether chemical “footprints” (i.e. scent-marks) predict reward or non-reward (Saleh & Chittka, 2006), in our case all patterns were non-rewarding and so there was no possibility of cued discriminative learning. Statistical analyses The data were treated as binomial proportions. Two analyses were undertaken in Experiment 1. First, the proportion of choices for the corridor occupied by the conspecific was compared across the four conditions. A logistic model, which specifies a binomial error term, was fit to the data using SPSS 21. Subsequently, each of the choice proportions was compared to a theoretical value of chance (1/12) using the replicated goodness-of-fit test (Sokal & Rohlf, 2012). The replicated goodnessof-fit test yields two G values, which are compared to 2 values in the tests of significance: GP (P for pooled) compares group choice frequency to that expected by chance, and GH (H for heterogeneity) tests for individual differences. In Experiment 2, a logistic model was used, with bee weights entered as a covariate. The G-test was used to determine whether the proportion of choices for the radial pattern differed from a theoretical value of chance (3/10) for each of the 3 conditions. Results Experiment 1: social preferences Comparisons across conditions Between-colony differences in the overall preference for the Occupied-Conspecific corridor were observed. Individuals from the second colony chose it 3.5 times out of 20 on average, which was significantly higher than the mean of 2.5 for the first colony (2 = 8.98, df = 1, p = .003). No interaction between colony and experimental condition, however, was detected (2 = 1.17, df = 3, p = .76), and so the data from both colonies were pooled. The effect of condition was significant (2 = 32.57, df = 3, p < .0001). Given that behaviour in the baseline condition was almost indistinguishable from chance (Fig. 2a and b), the question as to whether both the food restriction and the CO2 manipulations were effective in changing preference is addressed below in a comparison of those conditions with chance. Comparisons with theoretical value of chance The choice proportion for the Occupied-Conspecific corridor that would be expected by chance (Fig. 2a) was 8% (1/12). The group choice proportion in the Baseline condition was no different from chance (GP = 1.66, 1 df, p = .20), though individual differences were significant (GH = 37.90, df = 23, p = 0.03) with one outlier making 7 out of 20 choices of the OccupiedConspecific corridor while the rest made 0–4 such choices. In the Pollen deprivation condition (Fig. 2c), the preference of 14% was significantly higher than chance (GP = 18.01, 1 df, p < .0001). The same was true of the strong CO2 condition (GP = 14.59, 1 df, p < .0001) shown in Fig. 2d. The strongest effect was observed in the Pure CO2 condition (GP = 94.34, 1 df, p < .0001): Fig. 2e shows that the proportion of choices for the Occupied-Conspecific corridor was approximately twice as high as chance and there was a commensurate decrease in the
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
7
Fig. 2. Choice proportions, pooled for two colonies in Experiment 1. There were 9 unoccupied corridors and 3 occupied (with conspecific, Styrofoam or coin). In each condition, 12 bees per colony were given 20 choices each. (a) Theoretical proportions of chance. (b–e) Observed proportions under four conditions (baseline, pollen deprivation, strong CO2 and pure CO2 . * The choice proportion for the occupied-conspecific corridor was significantly greater than a chance value of 8%.
proportion of choices for the Unoccupied corridors. The increase in preference for occupied corridors was specific to the corridor occupied by the conspecific: little if any change can be seen in the proportion of choices for the corridors occupied by the coin or the Styrofoam. In all three experimental conditions, there was significant individual variation (GH ≥ 38.81, df = 23, p < .05). In the pure CO2 condition, two bees showed a preference for the Occupied-Conspecific corridors as high as 50% of choices. Experiment 2: preferences for floral patterns Comparisons across conditions A significant effect of condition was found on preference for the radial patterns, (2 = 7.93, df = 2, p = .019). Again, because the choice proportions in Baseline were at chance levels, the question of which treatments were effective in modifying behaviour are addressed below in comparisons between group choice proportions and chance levels. Body mass (Mean = .178 g, SD = .03 g), which was entered in the logistic model as a covariate, was not an important explanatory variable (2 = 2.20, df = 1, p = .14). The interaction between body mass and condition was not significant (2 = .41, df = 2, p = .81). Comparisons with theoretical value of chance In the baseline condition, the choice proportion for the radial patterns was not significantly different from a chance value of 30% (GP = 0.295, df = 1, p = .59). Fig. 3 shows that a preference emerged in the condition of pollen deprivation (GP = 13.14, df = 1, p = .0003): the choice proportion rose to 38%. No such preference, however, was detected in the pure CO2 condition (GP = .05, df = 1, p = .82). No individual differences were detected in any of the three conditions (GH ≤ 18.65, df = 19). Discussion In general, the results of our experiments show that pattern preferences by bumblebees leaving their nest for the first time are context specific: they are tied to variables relevant to different functions. As such this work fits well within an ecological approach to learning and motivation (Shettleworth, 2010) whereby specific situations trigger appropriate goalspecific sequences of unlearned and learned behaviour patterns. Preferences for unrewarding radial patterns over concentric patterns, which have been reported for flower-naïve bumblebees (Simonds & Plowright, 2004; Orbán & Plowright, 2013; Plowright, Simonds, & Butler, 2006) are thought to reflect preferences for floral shapes (Lehrer et al., 1995). Such preferences might help to guide inexperienced bees toward their first floral contact. If so, the preferences should depend on conditions of food deprivation. Here, we found no detectable preference when the bees had unlimited access to pollen, but a preference was created when pollen was taken away from the colony (Experiment 2). Administering CO2 , however, led to no discernible change in preference, though the manipulation of CO2 was anything but subtle: the concentration of CO2 could not have been higher. One caveat is that while there was ample room on the scale of measurement for an increase in preference, there was less room for a decrease, though one might have been envisaged given that exposure to CO2 in honeybees (A. mellifera) leads to a decrease in pollen collection (Ribbands, 1950).
8
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
Fig. 3. Mean choice frequencies of radial and concentric patterns in Experiment 2 with standard error bars. In each condition, 20 bees were given 20 choices in a maze where 3 corridors contained a radial pattern and 7 contained a concentric pattern. * Choice proportion significantly different from a chance value of 30%.
In colonies of bumblebees, division of labour is related to size variation, with larger workers carrying out a larger share of the foraging duties than smaller workers (Jandt & Dornhaus, 2009) and being more adept at these tasks (Spaethe & Weidenmüller, 2002). Current research on size variation in bumblebees has considered both its proximal causes (Couvillon, Jandt, Duong, & Dornhaus, 2010) and functional significance (Couvillon & Dornhaus, 2010; Peat, Tucker, & Goulson, 2005). Our argument that the preference for radial patterns is linked to food gathering would have been bolstered had larger bees been more affected by the manipulation of food availability than smaller bees, but no interaction of our experimental manipulation with body mass was obtained—with self-selected groups of bees that exited the colony rather than stayed in, the range of weights was presumably restricted. If a preference for a pattern occupied by another forager also helps to guide bees toward their first floral contact, then it too should depend on food deprivation. It did. Pollen deprivation, however, was not the only variable that affected preference. CO2 also had an effect, in both concentrations that we used (Experiment 1). Perhaps the preference for corridors occupied by a conspecific is a general stress response (see Muth, Scampini, & Leonard, 2015) to threatening situations that include both lack of food and the presence of predators. The function of the response to CO2 , however, is less clear than the function of a response to food deprivation. Here we suggest three possibilities: (1) The increase in attraction to an occupied stimulus may be the beginnings of an aggressive response, similar to that of individuals within colonies that hiss and buzz in response to the vibrations and sudden increase in CO2 levels characteristic of disruption by a predator (Kirchner & Röschard, 1999). One untested assumption underlying this interpretation is that the aggressive motivation triggered by the CO2 was not predator specific and would lead to an increased tendency to attack any organism in the vicinity. (2) Alternatively, it may be a response, in the context of danger, to what may be perceived as a signal of safety: an unlearned response analogous to the learned response documented by Dawson and Chittka (2014). (3) We cannot rule out the possibility that it may even be protective behaviour whereby one bee joins another to defend herself and her sister from a common threat. What does seem clear is that the response is selective: it was specific to corridors occupied by another insect, and not corridors occupied by a similarly sized coin or piece of Styrofoam. The specific physical characteristics of the insect that triggered approach behaviour remain to be determined. Perhaps dark colour is important, since dark cotton balls elicit more stinging in honeybees than white ones (Free, 1961). Indeed, black ping pong balls are more likely to be attacked by bumblebees (Bombus sonorus Say) than white ping pong balls (Visscher & Vetter, 1995). Rough texture may also be important (Free, 1961). The mechanisms by which our variables exerted their effects on choice remain to be determined. Pure CO2 is an anesthetic, and indeed a few bees in the pure CO2 condition in Experiment 2 did gradually became immobile and failed to complete 20 choices, perhaps just as much because of the decrease in O2 as the presence of CO2 . The effects in our study, however, were not simply decreases in motor activity, since choice of any corridor required the same approach toward the back wall and we obtained groups of equal size all completing 20 choices. The effects may be the product of increased attention to the relevant stimuli. They may reflect memory and decision making—indeed CO2 produces retrograde amnesia in honeybees (Beckmann, 1974), and individual decision-making in honeybees is influenced by variables that mimic predation, such as being shaken (Bateson, Desire, Gartside, & Wright, 2011). Having established both the effect of food deprivation on preference for radial stimuli, and the effects of CO2 exposure and food deprivation on preference for patterns occupied by other bees, future research should harness these variables in more ecologically realistic ways. In the future, gradations of pollen availability could be used instead of all-or-nothing. In addition, colonies at various stages of development could be used: the effect of pollen restriction is almost certainly specific to young colonies during periods of rapid colony growth when protein is in high demand for the feeding of larvae (Free, 1955). The CO2
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
9
concentration could also be brought down to levels that would be encountered in nature. We used concentrations of 25% and 100% CO2 , which vastly exceed the concentration in the colony [0.2–6% in honeybee hives (Nicolas & Sillans, 1989)], in the atmosphere [385 ppm in 2008 (Hansen et al., 2008) – up from 350 ppm in 1989 (Nicolas & Sillans, 1989)] or in mammalian breath [3.5% for humans (Kirchner & Röschard, 1999)]. In our study, the workers were also exposed to CO2 individually in the maze, whereas in nature the influx of CO2 would not likely be experienced at the foraging site but rather at the site of the colony. In future research, the two locations could be compared, though prolonged CO2 exposure in the colony can lead to aberrant behaviours such as expelling larvae (Pomeroy & Plowright, 1979) and also can affect the reproductive behaviour of the queen (Mackensen, 1947). Following the lead of Gould (1990), today bees are used as a model for understanding cognition (Menzel, 2012; Menzel & Brembs, 2007). Contemporary research topics in bee-cognition include visual categorization (Benard, Stach, & Giurfa, 2006; Dyer, 2012), timing (Boisvert & Sherry, 2006; Craig, Varnon, Sokolowski, Wells, & Abramson, 2014), spatial cognition (Brown & Demas, 1994; Cheng, 2000), numerical cognition (Pahl, Si, & Zhang, 2013), concept learning (Brown & Sayde, 2013; Giurfa, Zhang, Jenett, Menzel, & Srinivasan, 2001; Gould, 2002), picture-object correspondence (Thompson & Plowright, 2014) and problem solving (Mirwan & Kevan, 2014). Particularly relevant here is research on social learning and cognition (reviewed by Sherry & Strang, 2014), which has been neglected until recently (Dukas, 2008). Much of this cognition, however, is built upon some “pre-functional” behaviours (Hogan, 1994), that is, the building blocks of learning (Tinbergen, 1951). The behaviour that occurs prior to experience with the goals it is designed to attain remains poorly understood for bees. Prior research has documented some of the variables regarding the visual appearance of flowers that influence this pre-functional behaviour (reviewed by Orbán & Plowright, 2014), and this paper documents some of the variables that affect the internal state of the bees. Now that the ‘cognitive revolution’ has become mainstream in bee behaviour, there may be something to be gained by stepping back to consider the details of how bees do what they do when they are met with the world outside their colony for the first time. Acknowledgments This research was supported by a research grant to C.M.S.P. from the Natural Sciences and Engineering Research Council of Canada. We thank Daniel Dostie for his help with the graphics and Levente Orbán for his assistance with the videos. Vicki Xu and Emma Thompson provided constructive comments on the manuscript. We also thank Koppert Canada for their generosity in donating the bumblebee colonies for our research. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.lmot.2014.10.006. References Avarguès-Weber, A., & Chittka, L. (2014a). Observational conditioning in flower choice copying by bumblebees (Bombus terrestris): Influence of observer distance and demonstrator movement. PLOS ONE, 9, e88415. Avarguès-Weber, A., & Chittka, L. (2014b). Local enhancement or stimulus enhancement? Bumblebee social learning results in a specific pattern of flower preference. Animal Behaviour, 97, 185–191. Bateson, M., Desire, S., Gartside, S. E., & Wright, G. A. (2011). Agitated honeybees exhibit pessimistic cognitive biases. Current Biology, 21, 1070–1073. Baude, M., Danchin, É., Mugabo, M., & Dajoz, I. (2011). Conspecifics as informers and competitors: An experimental study in foraging bumble-bees. Proceedings of the Royal Society B, 278, 2806–2813. Beckmann, H. E. (1974). The damaging effect of supercooling, narcosis, and stress on memory of the honey bee. Journal of Comparative Physiology, 94, 249–266. Benard, J., Stach, S., & Giurfa, M. (2006). Categorization of visual stimuli in the honeybee Apis mellifera. Animal Cognition, 9, 257–270. Boisvert, M. J., & Sherry, D. F. (2006). Interval timing by an invertebrate, the bumble bee Bombus impatiens. Current Biology, 16, 1636–1640. Brown, M. F., & Demas, G. E. (1994). Evidence for spatial working memory in honeybees Apis mellifera. Journal of Comparative Psychology, 108, 344–352. Brown, M. F., & Sayde, J. M. (2013). Same/different discrimination by bumblebee colonies. Animal Cognition, 16, 117–125. Cheng, K. (2000). How honeybees find a place: Lessons from a simple mind. Animal Learning & Behavior, 28, 1–15. Couvillon, M. J., & Dornhaus, A. (2010). Small worker bumble bees (Bombus impatiens) are hardier against starvation than their larger sisters. Insectes Sociaux, 57, 193–197. Couvillon, M. J., Jandt, J. M., Duong, N. H. I., & Dornhaus, A. (2010). Ontogeny of worker body size distribution in bumble bee (Bombus impatiens) colonies. Ecological Entomology, 35, 424–435. Craig, D. P. A., Varnon, C. A., Sokolowski, M. B., Wells, H., & Abramson, C. I. (2014). An assessment of fixed interval timing in free-flying honey bees (Apis mellifera ligustica): An analysis of individual performance. PLOS ONE, 9, e101262. Dawson, E. H., Avarguès-Weber, A., Chittka, L., & Leadbeater, E. (2013). Learning by observation emerges from simple associations in an insect model. Current Biology, 23, 727–730. Dawson, E. H., & Chittka, L. (2012). Conspecific and heterospecific information use in bumblebees. PLOS ONE, 7, e31444. Dawson, E. H., & Chittka, L. (2014). Bumblebees (Bombus terrestris) use social information as an indicator of safety in dangerous environments. Proceedings of the Royal Society B, 281, 1–5. Dukas, R. (2008). Evolutionary biology of insect learning. Annual Review of Entomology, 53, 145–160. Dyer, A. G. (2012). The mysterious cognitive abilities of bees: Why models of visual processing need to consider experience and individual differences in animal performance. Journal of Experimental Biology, 215, 387–395. Free, J. B. (1955). The division of labour within bumblebee colonies. Insectes Sociaux, 2, 195–212. Free, J. B. (1961). The stimuli releasing the stinging response of honeybees. Animal Behaviour, 9, 193–196. Giurfa, M., Zhang, S., Jenett, A., Menzel, R., & Srinivasan, M. V. (2001). The concepts of ‘sameness’ and ‘difference’ in an insect. Nature, 410, 930–933.
10
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
Gould, J. L. (1990). Honey bee cognition. Cognition, 37, 83–103. Gould, J. L. (2002). Can honey bees create cognitive maps? In M. Bekoff, C. Allen, & G. M. Burghardt (Eds.), The cognitive animal (pp. 41–45). Cambridge, MA: MIT Press. Goulson, D. (2010). Bumblebees: Behaviour, ecology, and conservation (2nd ed.). New York: Oxford University Press., 317 pp. Grüter, C., & Leadbeater, E. (2014). Insights from insects about adaptive social information use. Trends in Ecology & Evolution, 29, 177–184. Hansen, J., Sato, M., Kharecha, P., Beerling, D., Berner, R., Masson-Delmotte, V., et al. (2008). Target atmospheric CO2 : Where should humanity aim? Open Atmospheric Science Journal, 2, 217–231. Hogan, J. A. (1994). Structure and development of behavior systems. Psychonomic Bulletin and Review, 1, 439–450. Jandt, J. M., & Dornhaus, A. (2009). Spatial organization and division of labour in the bumblebee Bombus impatiens. Animal Behaviour, 77, 641–651. Kawaguchi, L. G., Ohashi, K., & Toquenaga, Y. (2006). Do bumble bees save time when choosing novel flowers by following conspecifics? Functional Ecology, 20, 239–244. Kikuchi, T. (1963). Studies on the coaction among insects visiting flowers. III. Dominance relationship among flower-visiting flies, bees and butterflies. Science Reports of Tohoku University – Fourth Series: Biology, 29, 1–8. Kirchner, W. H., & Röschard, J. (1999). Hissing in bumblebees: An interspecies defence signal. Insectes Sociaux, 46, 239–243. Leadbeater, E., & Chittka, L. (2007). The dynamics of social learning in an insect model, the bumblebee (Bombus terrestris). Behavioral Ecology and Sociobiology, 61, 1789–1796. Leadbeater, E., & Chittka, L. (2009). Bumble-bees learn the value of social cues through experience. Biology Letters, 5, 310–312. Lehrer, M., Horridge, G. A., Zhang, S. W., & Gadagkar, R. (1995). Shape vision in bees: Innate preference for flower-like patterns. Philosophical Transactions of the Royal Society B, 347, 123–137. Mackensen, O. (1947). Effect of carbon dioxide on initial oviposition of artificially inseminated and virgin queen bees. Journal of Economic Entomology, 40, 344–349. Menzel, R. (2012). The honeybee as a model for understanding the basis of cognition. Nature Reviews Neuroscience, 13, 758–768. Menzel, R., & Brembs, B. (2007). Cognition in invertebrates. In J. H. Kaas (Ed.), Evolution of nervous systems, Vol. II: Evolution of nervous systems in invertebrates (pp. 403–422). Oxford: Academic Press. Mirwan, H. B., & Kevan, P. G. (2014). Problem solving by worker bumblebees Bombus impatiens (Hymenoptera: Apoidea). Animal Cognition, http://dx.doi.org/10.1007/s10071-014-0737-0 Muth, F., Scampini, A. V., & Leonard, A. S. (2015). The effects of acute stress on learning and memory in bumble bees. Learning and Motivation (in this issue). Nicolas, G., & Sillans, D. (1989). Immediate and latent effects of carbon dioxide on insects. Annual Review of Entomology, 34, 97–116. Orbán, L. L., & Plowright, C. M. S. (2013). The effect of flower-like and non-flower-like visual properties on choice of unrewarding patterns by bumblebees. Naturwissenschaften, 100, 621–631. Orbán, L. L., & Plowright, C. M. S. (2014). Getting to the start line: How bumblebees and honeybees are visually guided towards their first floral contact. Insectes Sociaux, http://dx.doi.org/10.1007/s00040-014-0366-2 Pahl, M., Si, A., & Zhang, S. (2013). Numerical cognition in bees and other insects. Frontiers in Psychology, 4, 162. Peat, J., Tucker, J., & Goulson, D. (2005). Does intraspecific size variation in bumblebees allow colonies to efficiently exploit different flowers? Ecological Entomology, 30, 176–181. Plowright, C. M. S., Evans, S. A., Chew Leung, J., & Collin, C. A. (2011). The preference for symmetry in flower naïve and not-so-naïve bumblebees. Learning and Motivation, 42, 76–83. Plowright, C. M. S., Ferguson, K. A., Jellen, S. L., Xu, V., Service, E. W., & Dookie, A. L. (2013). Bumblebees occupy: When foragers do and do not use the presence of others to first find food. Insectes Sociaux, 60, 517–524. Plowright, C. M. S., Simonds, V. M., & Butler, M. A. (2006). How bumblebees first find flowers: Habituation of visual pattern preferences, spontaneous recovery and dishabituation. Learning and Motivation, 37, 66–78. Pomeroy, N., & Plowright, R. C. (1979). Larval ejection following narcosis of bumble bees (Hymenoptera: Apidae). Journal of the Kansas Entomological Society, 52, 215–217. Ribbands, C. R. (1950). Changes in the behaviour of honey-bees following their recovery from anaesthesia. Journal of Experimental Biology, 27, 302–310. Saleh, N., & Chittka, L. (2006). The importance of experience in the interpretation of conspecific chemical signals. Behavioral Ecology and Sociobiology, 61, 215–220. Sherry, D. F., & Strang, C. G. (2014). Contrasting styles in cognition and behaviour in bumblebees and honeybees. Behavioural Processes, http://dx.doi.org/ 10.1016/j.beproc.2014.09.005 Shettleworth, S. J. (2010). Cognition, evolution, and behavior (2nd ed.). New York: Oxford University Press., 720 pp. Simonds, V., & Plowright, C. M. S. (2004). How do bumblebees first find flowers? Unlearned approach responses and habituation. Animal Behaviour, 67, 379–386. Sokal, R. R., & Rohlf, J. F. (2012). Biometry (4th ed.). New York: W.H. Freeman. Spaethe, J., & Weidenmüller, A. (2002). Size variation and foraging rate in bumblebees (Bombus terrestris). Insectes Sociaux, 49, 142–146. Tepedino, V. J., & Parker, F. D. (1981). The effect of floral occupancy on the foraging of flower-visiting insects. Psyche, 88, 321–330. Thompson, E. L., & Plowright, C. M. S. (2014). How images may or may not represent flowers: Picture-object correspondence in bumblebees (Bombus impatiens)? Animal Cognition, 17, 1031–1043. Tinbergen, N. (1951). The study of instinct. London: Oxford University Press. Visscher, K. P., & Vetter, R. S. (1995). Smoke and target color effects on defensive behavior in yellow jacket wasps and bumble bees (Hymenoptera: Vespidae, Apidae) with a description of an electronic attack monitor. Journal of Economic Entomology, 88, 579–583.
Contents lists available at ScienceDirect
Learning and Motivation journal homepage: www.elsevier.com/locate/l&m
Food restriction and threat of predation affect visual pattern choices by flower-naïve bumblebees E.W. Service, C.M.S. Plowright ∗ School of Psychology, University of Ottawa, Ottawa, Ont. K1N 6N5, Canada
a r t i c l e
i n f o
Article history: Available online 9 March 2015
Keywords: Bumblebees Floral choice Carbon dioxide Social cognition Pattern recognition Predation threat
a b s t r a c t The aim of this study was to determine whether a preference by flower-naïve bumblebees could be created or enhanced by manipulating variables relevant to food collection and to defense against predation. In two experiments, colonies of bumblebees (Bombus impatiens) were deprived of pollen, exposed to CO2 , or neither. Choices of individual workers in a radial arm maze were monitored. In Experiment 1, both variables lead to a preference for corridors occupied by a conspecific bee. The effect was specific: no change in preference for corridors occupied by other objects (a coin and a piece of Styrofoam) was detected. In Experiment 2, radial and concentric patterns were used, both of which were unoccupied. Only pollen deprivation increased preference for radial stimuli, while CO2 had no discernible effect. Preferences for visual patterns by bees leaving their colony for the first time are modulated by variables that affect the internal state of the bees in problem-specific ways. © 2015 Elsevier Inc. All rights reserved.
The possibility that foraging bees use the presence of other insects on flowers to govern their own floral choices has received little if any support in the field. For instance, one survey of counts of floral visitors found on flower heads of sunflowers (Helianthus annuus L.) and onion flowers (Allium cepa L.) addressed the question of whether there would be attraction or avoidance of inflorescences that were already occupied by one or more individuals. The distribution of insects, including honeybees (Apis mellifera L.) and bumblebees (Bombus spp. Latreille), followed a Poisson distribution (Tepedino & Parker, 1981): there were neither more nor less flowers with just one forager than would be expected by chance. The notion that insects forage independently of other insects may, however, have been dismissed prematurely by field biologists. Recently, an examination of the details of social effects on foraging in the laboratory has revealed that bees use associative learning to take advantage of cues that predict important outcomes such as presence of floral reward (Avarguès-Weber & Chittka, 2014a; Dawson, Avarguès-Weber, Chittka, & Leadbeater, 2013; Leadbeater & Chittka, 2009) and presence of predators (Dawson & Chittka, 2014)—in general, social information is used by insects strategically (Grüter & Leadbeater, 2014). Moreover, workers that have just left their colony for the first time also show a significant preference for flowers that are already occupied by another forager (Kawaguchi, Ohashi, & Toquenaga, 2006; Leadbeater & Chittka, 2009), though this effect is most evident when the flowers are rare and the occupiers are large relative to the flowers (Plowright et al., 2013). Because of the difficulty in obtaining the pre-experimental histories of bees seen foraging in the field, these sorts of effects may have eluded detection in nature. In this paper, our focus is on the behaviour of “flower-naïve” bumblebee workers: bees that leave their colony for the first time, and as such, have had no prior experience with flowers. One view regarding the preferences for occupied flowers
∗ Corresponding author. E-mail address: [email protected] (C.M.S. Plowright). http://dx.doi.org/10.1016/j.lmot.2014.10.006 0023-9690/© 2015 Elsevier Inc. All rights reserved.
4
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
by flower-naïve bees is that they may help inexperienced bees to locate sources of food: a pattern with another bee on it is likely to be a flower. One difficulty with this interpretation is that it is post hoc. Indeed, the exact opposite argument can be made a priori: a flower that is occupied by a forager is likely to be empty or well on its way to being depleted, and so it ought not to be preferred but avoided. In other words, other foragers may possibly act as competitors rather than as informers (Baude, Danchin, Mugabo, & Dajoz, 2011). Another view is that the presence of an occupier on a flower has little to do with signalling resource availability. Field observations have suggested that interactions among bees on a flower are in fact aggressive (Kikuchi, 1963). Indeed, in the course of a prior study on pattern preferences of flower-naïve bees (Orbán & Plowright, 2013), we captured a few such interactions on film (Videos 1 and 2 in the Supplementary Materials): the behaviours of the bee landing on the flower seemed more directed at the occupier than at the flower, which suggests that the occupier might have been perceived as a predator or a competitor. In two experiments, we manipulated internal states of flower-naïve bees. To promote food finding behaviours, we manipulated the availability of pollen, which is needed for feeding to larvae. To promote aggressive or defensive behaviours, we exposed the bees to CO2 to simulate the presence of a mammalian predator. Predators of bumblebee colonies (Goulson, 2010) include mice (Mus domesticus), badgers (Meles meles L.) in Europe, and skunks (Mephitis mephitis Schreber) in NorthAmerica. Bumblebees perceive CO2 : they respond to mammalian breath and to currents of air containing 5 or 10% CO2 by hissing, which serves as an inter-specific defence signal (Kirchner & Röschard, 1999). If a preference for occupied stimuli is engaged when bees are food searching, then pollen deprivation should create or increase the preference. If the preference is engaged in situations that trigger aggressive behaviours, then exposure to CO2 should create or increase the preference. Experiment 1 tested these two predictions. In addition, by comparing the preference for stimuli occupied by another bee with the preference for stimuli occupied by non-organic objects (a coin, as in Dawson and Chittka (2012), and a piece of Styrofoam), we began to address the question of whether any preference for occupied stimuli was indeed social. Aggressive tendencies ought to be directed at other individuals and not at flowers themselves. Heightening an aggressive tendency ought not to increase a preference for a pattern that is more “floral” than another. To determine whether the effects of our variables were specific to social preferences, Experiment 2 examined their effects on preference for floral patterns with no occupiers. Given that radial patterns (illustrated in the legend of Fig. 3) are, by and large, preferred over concentric patterns (Orbán & Plowright, 2013) as they are thought to resemble flowers in nature (Lehrer, Horridge, Zhang, & Gadagkar, 1995), we examined the effect of food deprivation and CO2 exposure on relative choice of these patterns. Food deprivation ought, if anything, to increase preference for radial patterns, while CO2 should not. Given that larger bumblebees tend to invest themselves in foraging duties while smaller bees tend to the nest (Goulson, 2010), though task specialization is not as marked as in honeybees, we reasoned that the effect of food deprivation might interact with body mass. Methods Subjects Three commercial colonies of Bombus impatiens Cresson in plastic nest boxes (19.5 cm × 17.5 cm × 12 cm high) were supplied by Koppert Canada. Because the colonies were covered with an opaque lid, they received little light, as in nature. In Experiment 1, 48 bees from each of two colonies were used, for a total of 96. In Experiment 2, 60 bees from the third colony were used. All bees were tested the first time they left the colony: they had no pre-experimental experience outside of the nest box. In all conditions, to motivate the bees to exit the colony, the wick that absorbed sugar solution from a plastic bag beneath the nest box was capped for one or two days prior to testing. Apparatus The 12-corridor maze that we used, modelled on that of Lehrer et al. (1995), is diagrammed in Fig. 1. Photographs are shown in Plowright, Evans, Chew Leung, and Collin (2011, Fig. 1). It was constructed of grey Plexiglas® with a clear cover. From the central area (22 cm wide, 15 cm high), 12 corridors radiated outwards. The bees entered the maze via a screen tube that connected their colony exit hole to an entrance hole in the center of the floor of the maze. The exit was stoppered when the colony was not in use during the experiment: bees were not allowed to travel to and from the apparatus prior to testing. The entrance to the maze was gated during testing sessions: bees were let into the maze individually. The corridors were 6 cm wide at their entrance from the center of the maze and 15 cm long. The back walls were 13 cm wide. The maze was illuminated from above by high-frequency (>40 kHz) lighting equipment (three Sylvania Quicktronic T8 QHE4x32T8/112 light ballasts, each with four Sylvania model FO32/841/XP/SS/EC03 fluorescent light bulbs). Stimuli In both experiments, the stimuli were mounted with Velcro® on the back walls of the corridors in the maze, as shown in Fig. 1.
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
5
Fig. 1. Radial maze and connected hive box. Patterns were mounted vertically on the walls of the corridors in the positions marked with a red circle. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.) Source: Reproduced from Plowright et al. (2013), with kind permission from Springer Science and Business Media.
Experiment 1 Nine corridors in the maze were empty, that is, “Unoccupied”. The remaining three corridors enclosed a stimulus, so they were “Occupied”. In one of these three corridors, a dead pinned queen B. impatiens was attached to the back wall (“OccupiedConspecific”). Bumblebees are attracted to artificial flowers where a model bee or a dead bee have been pinned in much the same way they are attracted to artificial flowers where other bees are already foraging (Leadbeater & Chittka, 2007), though there is no claim here that motion cues are irrelevant in general (indeed, they are important in defensive behaviour of honeybees (Free, 1961) and in social learning of foraging behaviour by bumblebees (Avarguès-Weber & Chittka, 2014a,b). Accordingly, we used a dead pinned queen bee as in Plowright et al. (2013) to maximize the chances that it would be detected by virtue of its larger size. In nature, the times at which workers and queens forage within a season overlap and so workers could well encounter other queens. A Canadian dime (“Occupied-Coin”) the colour of steel was placed in the second corridor, and a piece of white Styrofoam® (“Occupied-Styrofoam”) of the same diameter as the coin (18 mm) was placed in the third corridor. Experiment 2 Two opposing corridors were blocked off and the remaining ten contained circular (6 cm diameter) stimuli: three radial patterns consisting of four black and four white wedges, and seven concentric patterns consisting of two black and two white circles. The patterns were printed on paper in black and white and then laminated. Design Experiment 1 For each of the following four conditions, 12 bees from each of the two colonies were used: (1) “Baseline”: the colony had ad libitum access to a ball of pollen paste placed beside the cocoons (pollen purchased from an apiarist, crushed and mixed with honey and water), though access to sugar solution was prevented for two days prior to testing so that the bees would be motivated to exit the colony. (2) “Pollen deprivation”: to increase the level of motivation to forage over and above the level in baseline, not only was there deprivation of sugar solution, but pollen deprivation was superimposed, with the pollen ball being removed for 24 h prior to testing. (3) “Strong CO2 ”: as with baseline, but with the addition of a steady stream into the maze of a mixture of CO2 and O2 (1:3 by volume) for the duration of the time spent in the maze by each bee and (4) “Pure CO2 ”: pure carbon dioxide gas (though the maze was not gas tight so there was some dilution). In the last two conditions, the gas was released directly into the maze through an aperture in the lid so as to ensure that individual bees were under
6
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
the influence of CO2 as they made their pattern choices. The rate of gas flow was 1.1–1.3 L/min. For the first colony used, the conditions were run in order 1–4, and for the second colony, the reverse order was used, so that any overall effect of experimental condition could not be attributed to colony aging. Three unoccupied corridors separated each of the three occupied corridors. The positions of the three types of occupied corridors (conspecific, coin and Styrofoam) were counterbalanced so that across 12 bees, they appeared equally often in each of the 12 corridors in the maze. Experiment 2 For each of the following three conditions, 20 bees were used: (1) baseline, (2) pure CO2 and (3) pollen deprivation. These conditions, which were run in order, are described above for Experiment 1. The positions of the patterns were counterbalanced so that across 20 bees in each condition, the three radial patterns appeared six times and the seven concentric patterns appeared fourteen times in each of the ten corridors used. Procedure The procedure was common to both experiments. Once a worker exited the colony and entered the maze for the first time (where the first time was also the last time), its first 20 choices were recorded as in Plowright et al. (2013). A choice was defined as making contact with the back wall of the corridor or the stimulus attached to it. A new choice was only recorded once the bee had crossed the threshold of the corridor and returned to the center of the maze (in other words, ‘dithering’ within a corridor did not count as multiple choices). Once 20 choices were made, the bee was removed from the apparatus and killed, after which the bees in Experiment 2 were weighted on a Mettler Toledo balance model PB 303-S. In the Pure CO2 condition, some bees (n = 7 in Experiment 1 and n = 4 in Experiment 2) became inactive for over 15 min with exposure to CO2 and did not complete 20 choices—they were excluded from the data set. Testing for each condition took place for 4–8 h per day, over the course of 1–3 days. The maze was cleaned with a damp cloth between subjects to minimize any visual evidence of a corridor having been visited. Though bumblebees can learn whether chemical “footprints” (i.e. scent-marks) predict reward or non-reward (Saleh & Chittka, 2006), in our case all patterns were non-rewarding and so there was no possibility of cued discriminative learning. Statistical analyses The data were treated as binomial proportions. Two analyses were undertaken in Experiment 1. First, the proportion of choices for the corridor occupied by the conspecific was compared across the four conditions. A logistic model, which specifies a binomial error term, was fit to the data using SPSS 21. Subsequently, each of the choice proportions was compared to a theoretical value of chance (1/12) using the replicated goodness-of-fit test (Sokal & Rohlf, 2012). The replicated goodnessof-fit test yields two G values, which are compared to 2 values in the tests of significance: GP (P for pooled) compares group choice frequency to that expected by chance, and GH (H for heterogeneity) tests for individual differences. In Experiment 2, a logistic model was used, with bee weights entered as a covariate. The G-test was used to determine whether the proportion of choices for the radial pattern differed from a theoretical value of chance (3/10) for each of the 3 conditions. Results Experiment 1: social preferences Comparisons across conditions Between-colony differences in the overall preference for the Occupied-Conspecific corridor were observed. Individuals from the second colony chose it 3.5 times out of 20 on average, which was significantly higher than the mean of 2.5 for the first colony (2 = 8.98, df = 1, p = .003). No interaction between colony and experimental condition, however, was detected (2 = 1.17, df = 3, p = .76), and so the data from both colonies were pooled. The effect of condition was significant (2 = 32.57, df = 3, p < .0001). Given that behaviour in the baseline condition was almost indistinguishable from chance (Fig. 2a and b), the question as to whether both the food restriction and the CO2 manipulations were effective in changing preference is addressed below in a comparison of those conditions with chance. Comparisons with theoretical value of chance The choice proportion for the Occupied-Conspecific corridor that would be expected by chance (Fig. 2a) was 8% (1/12). The group choice proportion in the Baseline condition was no different from chance (GP = 1.66, 1 df, p = .20), though individual differences were significant (GH = 37.90, df = 23, p = 0.03) with one outlier making 7 out of 20 choices of the OccupiedConspecific corridor while the rest made 0–4 such choices. In the Pollen deprivation condition (Fig. 2c), the preference of 14% was significantly higher than chance (GP = 18.01, 1 df, p < .0001). The same was true of the strong CO2 condition (GP = 14.59, 1 df, p < .0001) shown in Fig. 2d. The strongest effect was observed in the Pure CO2 condition (GP = 94.34, 1 df, p < .0001): Fig. 2e shows that the proportion of choices for the Occupied-Conspecific corridor was approximately twice as high as chance and there was a commensurate decrease in the
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
7
Fig. 2. Choice proportions, pooled for two colonies in Experiment 1. There were 9 unoccupied corridors and 3 occupied (with conspecific, Styrofoam or coin). In each condition, 12 bees per colony were given 20 choices each. (a) Theoretical proportions of chance. (b–e) Observed proportions under four conditions (baseline, pollen deprivation, strong CO2 and pure CO2 . * The choice proportion for the occupied-conspecific corridor was significantly greater than a chance value of 8%.
proportion of choices for the Unoccupied corridors. The increase in preference for occupied corridors was specific to the corridor occupied by the conspecific: little if any change can be seen in the proportion of choices for the corridors occupied by the coin or the Styrofoam. In all three experimental conditions, there was significant individual variation (GH ≥ 38.81, df = 23, p < .05). In the pure CO2 condition, two bees showed a preference for the Occupied-Conspecific corridors as high as 50% of choices. Experiment 2: preferences for floral patterns Comparisons across conditions A significant effect of condition was found on preference for the radial patterns, (2 = 7.93, df = 2, p = .019). Again, because the choice proportions in Baseline were at chance levels, the question of which treatments were effective in modifying behaviour are addressed below in comparisons between group choice proportions and chance levels. Body mass (Mean = .178 g, SD = .03 g), which was entered in the logistic model as a covariate, was not an important explanatory variable (2 = 2.20, df = 1, p = .14). The interaction between body mass and condition was not significant (2 = .41, df = 2, p = .81). Comparisons with theoretical value of chance In the baseline condition, the choice proportion for the radial patterns was not significantly different from a chance value of 30% (GP = 0.295, df = 1, p = .59). Fig. 3 shows that a preference emerged in the condition of pollen deprivation (GP = 13.14, df = 1, p = .0003): the choice proportion rose to 38%. No such preference, however, was detected in the pure CO2 condition (GP = .05, df = 1, p = .82). No individual differences were detected in any of the three conditions (GH ≤ 18.65, df = 19). Discussion In general, the results of our experiments show that pattern preferences by bumblebees leaving their nest for the first time are context specific: they are tied to variables relevant to different functions. As such this work fits well within an ecological approach to learning and motivation (Shettleworth, 2010) whereby specific situations trigger appropriate goalspecific sequences of unlearned and learned behaviour patterns. Preferences for unrewarding radial patterns over concentric patterns, which have been reported for flower-naïve bumblebees (Simonds & Plowright, 2004; Orbán & Plowright, 2013; Plowright, Simonds, & Butler, 2006) are thought to reflect preferences for floral shapes (Lehrer et al., 1995). Such preferences might help to guide inexperienced bees toward their first floral contact. If so, the preferences should depend on conditions of food deprivation. Here, we found no detectable preference when the bees had unlimited access to pollen, but a preference was created when pollen was taken away from the colony (Experiment 2). Administering CO2 , however, led to no discernible change in preference, though the manipulation of CO2 was anything but subtle: the concentration of CO2 could not have been higher. One caveat is that while there was ample room on the scale of measurement for an increase in preference, there was less room for a decrease, though one might have been envisaged given that exposure to CO2 in honeybees (A. mellifera) leads to a decrease in pollen collection (Ribbands, 1950).
8
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
Fig. 3. Mean choice frequencies of radial and concentric patterns in Experiment 2 with standard error bars. In each condition, 20 bees were given 20 choices in a maze where 3 corridors contained a radial pattern and 7 contained a concentric pattern. * Choice proportion significantly different from a chance value of 30%.
In colonies of bumblebees, division of labour is related to size variation, with larger workers carrying out a larger share of the foraging duties than smaller workers (Jandt & Dornhaus, 2009) and being more adept at these tasks (Spaethe & Weidenmüller, 2002). Current research on size variation in bumblebees has considered both its proximal causes (Couvillon, Jandt, Duong, & Dornhaus, 2010) and functional significance (Couvillon & Dornhaus, 2010; Peat, Tucker, & Goulson, 2005). Our argument that the preference for radial patterns is linked to food gathering would have been bolstered had larger bees been more affected by the manipulation of food availability than smaller bees, but no interaction of our experimental manipulation with body mass was obtained—with self-selected groups of bees that exited the colony rather than stayed in, the range of weights was presumably restricted. If a preference for a pattern occupied by another forager also helps to guide bees toward their first floral contact, then it too should depend on food deprivation. It did. Pollen deprivation, however, was not the only variable that affected preference. CO2 also had an effect, in both concentrations that we used (Experiment 1). Perhaps the preference for corridors occupied by a conspecific is a general stress response (see Muth, Scampini, & Leonard, 2015) to threatening situations that include both lack of food and the presence of predators. The function of the response to CO2 , however, is less clear than the function of a response to food deprivation. Here we suggest three possibilities: (1) The increase in attraction to an occupied stimulus may be the beginnings of an aggressive response, similar to that of individuals within colonies that hiss and buzz in response to the vibrations and sudden increase in CO2 levels characteristic of disruption by a predator (Kirchner & Röschard, 1999). One untested assumption underlying this interpretation is that the aggressive motivation triggered by the CO2 was not predator specific and would lead to an increased tendency to attack any organism in the vicinity. (2) Alternatively, it may be a response, in the context of danger, to what may be perceived as a signal of safety: an unlearned response analogous to the learned response documented by Dawson and Chittka (2014). (3) We cannot rule out the possibility that it may even be protective behaviour whereby one bee joins another to defend herself and her sister from a common threat. What does seem clear is that the response is selective: it was specific to corridors occupied by another insect, and not corridors occupied by a similarly sized coin or piece of Styrofoam. The specific physical characteristics of the insect that triggered approach behaviour remain to be determined. Perhaps dark colour is important, since dark cotton balls elicit more stinging in honeybees than white ones (Free, 1961). Indeed, black ping pong balls are more likely to be attacked by bumblebees (Bombus sonorus Say) than white ping pong balls (Visscher & Vetter, 1995). Rough texture may also be important (Free, 1961). The mechanisms by which our variables exerted their effects on choice remain to be determined. Pure CO2 is an anesthetic, and indeed a few bees in the pure CO2 condition in Experiment 2 did gradually became immobile and failed to complete 20 choices, perhaps just as much because of the decrease in O2 as the presence of CO2 . The effects in our study, however, were not simply decreases in motor activity, since choice of any corridor required the same approach toward the back wall and we obtained groups of equal size all completing 20 choices. The effects may be the product of increased attention to the relevant stimuli. They may reflect memory and decision making—indeed CO2 produces retrograde amnesia in honeybees (Beckmann, 1974), and individual decision-making in honeybees is influenced by variables that mimic predation, such as being shaken (Bateson, Desire, Gartside, & Wright, 2011). Having established both the effect of food deprivation on preference for radial stimuli, and the effects of CO2 exposure and food deprivation on preference for patterns occupied by other bees, future research should harness these variables in more ecologically realistic ways. In the future, gradations of pollen availability could be used instead of all-or-nothing. In addition, colonies at various stages of development could be used: the effect of pollen restriction is almost certainly specific to young colonies during periods of rapid colony growth when protein is in high demand for the feeding of larvae (Free, 1955). The CO2
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
9
concentration could also be brought down to levels that would be encountered in nature. We used concentrations of 25% and 100% CO2 , which vastly exceed the concentration in the colony [0.2–6% in honeybee hives (Nicolas & Sillans, 1989)], in the atmosphere [385 ppm in 2008 (Hansen et al., 2008) – up from 350 ppm in 1989 (Nicolas & Sillans, 1989)] or in mammalian breath [3.5% for humans (Kirchner & Röschard, 1999)]. In our study, the workers were also exposed to CO2 individually in the maze, whereas in nature the influx of CO2 would not likely be experienced at the foraging site but rather at the site of the colony. In future research, the two locations could be compared, though prolonged CO2 exposure in the colony can lead to aberrant behaviours such as expelling larvae (Pomeroy & Plowright, 1979) and also can affect the reproductive behaviour of the queen (Mackensen, 1947). Following the lead of Gould (1990), today bees are used as a model for understanding cognition (Menzel, 2012; Menzel & Brembs, 2007). Contemporary research topics in bee-cognition include visual categorization (Benard, Stach, & Giurfa, 2006; Dyer, 2012), timing (Boisvert & Sherry, 2006; Craig, Varnon, Sokolowski, Wells, & Abramson, 2014), spatial cognition (Brown & Demas, 1994; Cheng, 2000), numerical cognition (Pahl, Si, & Zhang, 2013), concept learning (Brown & Sayde, 2013; Giurfa, Zhang, Jenett, Menzel, & Srinivasan, 2001; Gould, 2002), picture-object correspondence (Thompson & Plowright, 2014) and problem solving (Mirwan & Kevan, 2014). Particularly relevant here is research on social learning and cognition (reviewed by Sherry & Strang, 2014), which has been neglected until recently (Dukas, 2008). Much of this cognition, however, is built upon some “pre-functional” behaviours (Hogan, 1994), that is, the building blocks of learning (Tinbergen, 1951). The behaviour that occurs prior to experience with the goals it is designed to attain remains poorly understood for bees. Prior research has documented some of the variables regarding the visual appearance of flowers that influence this pre-functional behaviour (reviewed by Orbán & Plowright, 2014), and this paper documents some of the variables that affect the internal state of the bees. Now that the ‘cognitive revolution’ has become mainstream in bee behaviour, there may be something to be gained by stepping back to consider the details of how bees do what they do when they are met with the world outside their colony for the first time. Acknowledgments This research was supported by a research grant to C.M.S.P. from the Natural Sciences and Engineering Research Council of Canada. We thank Daniel Dostie for his help with the graphics and Levente Orbán for his assistance with the videos. Vicki Xu and Emma Thompson provided constructive comments on the manuscript. We also thank Koppert Canada for their generosity in donating the bumblebee colonies for our research. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.lmot.2014.10.006. References Avarguès-Weber, A., & Chittka, L. (2014a). Observational conditioning in flower choice copying by bumblebees (Bombus terrestris): Influence of observer distance and demonstrator movement. PLOS ONE, 9, e88415. Avarguès-Weber, A., & Chittka, L. (2014b). Local enhancement or stimulus enhancement? Bumblebee social learning results in a specific pattern of flower preference. Animal Behaviour, 97, 185–191. Bateson, M., Desire, S., Gartside, S. E., & Wright, G. A. (2011). Agitated honeybees exhibit pessimistic cognitive biases. Current Biology, 21, 1070–1073. Baude, M., Danchin, É., Mugabo, M., & Dajoz, I. (2011). Conspecifics as informers and competitors: An experimental study in foraging bumble-bees. Proceedings of the Royal Society B, 278, 2806–2813. Beckmann, H. E. (1974). The damaging effect of supercooling, narcosis, and stress on memory of the honey bee. Journal of Comparative Physiology, 94, 249–266. Benard, J., Stach, S., & Giurfa, M. (2006). Categorization of visual stimuli in the honeybee Apis mellifera. Animal Cognition, 9, 257–270. Boisvert, M. J., & Sherry, D. F. (2006). Interval timing by an invertebrate, the bumble bee Bombus impatiens. Current Biology, 16, 1636–1640. Brown, M. F., & Demas, G. E. (1994). Evidence for spatial working memory in honeybees Apis mellifera. Journal of Comparative Psychology, 108, 344–352. Brown, M. F., & Sayde, J. M. (2013). Same/different discrimination by bumblebee colonies. Animal Cognition, 16, 117–125. Cheng, K. (2000). How honeybees find a place: Lessons from a simple mind. Animal Learning & Behavior, 28, 1–15. Couvillon, M. J., & Dornhaus, A. (2010). Small worker bumble bees (Bombus impatiens) are hardier against starvation than their larger sisters. Insectes Sociaux, 57, 193–197. Couvillon, M. J., Jandt, J. M., Duong, N. H. I., & Dornhaus, A. (2010). Ontogeny of worker body size distribution in bumble bee (Bombus impatiens) colonies. Ecological Entomology, 35, 424–435. Craig, D. P. A., Varnon, C. A., Sokolowski, M. B., Wells, H., & Abramson, C. I. (2014). An assessment of fixed interval timing in free-flying honey bees (Apis mellifera ligustica): An analysis of individual performance. PLOS ONE, 9, e101262. Dawson, E. H., Avarguès-Weber, A., Chittka, L., & Leadbeater, E. (2013). Learning by observation emerges from simple associations in an insect model. Current Biology, 23, 727–730. Dawson, E. H., & Chittka, L. (2012). Conspecific and heterospecific information use in bumblebees. PLOS ONE, 7, e31444. Dawson, E. H., & Chittka, L. (2014). Bumblebees (Bombus terrestris) use social information as an indicator of safety in dangerous environments. Proceedings of the Royal Society B, 281, 1–5. Dukas, R. (2008). Evolutionary biology of insect learning. Annual Review of Entomology, 53, 145–160. Dyer, A. G. (2012). The mysterious cognitive abilities of bees: Why models of visual processing need to consider experience and individual differences in animal performance. Journal of Experimental Biology, 215, 387–395. Free, J. B. (1955). The division of labour within bumblebee colonies. Insectes Sociaux, 2, 195–212. Free, J. B. (1961). The stimuli releasing the stinging response of honeybees. Animal Behaviour, 9, 193–196. Giurfa, M., Zhang, S., Jenett, A., Menzel, R., & Srinivasan, M. V. (2001). The concepts of ‘sameness’ and ‘difference’ in an insect. Nature, 410, 930–933.
10
E.W. Service, C.M.S. Plowright / Learning and Motivation 50 (2015) 3–10
Gould, J. L. (1990). Honey bee cognition. Cognition, 37, 83–103. Gould, J. L. (2002). Can honey bees create cognitive maps? In M. Bekoff, C. Allen, & G. M. Burghardt (Eds.), The cognitive animal (pp. 41–45). Cambridge, MA: MIT Press. Goulson, D. (2010). Bumblebees: Behaviour, ecology, and conservation (2nd ed.). New York: Oxford University Press., 317 pp. Grüter, C., & Leadbeater, E. (2014). Insights from insects about adaptive social information use. Trends in Ecology & Evolution, 29, 177–184. Hansen, J., Sato, M., Kharecha, P., Beerling, D., Berner, R., Masson-Delmotte, V., et al. (2008). Target atmospheric CO2 : Where should humanity aim? Open Atmospheric Science Journal, 2, 217–231. Hogan, J. A. (1994). Structure and development of behavior systems. Psychonomic Bulletin and Review, 1, 439–450. Jandt, J. M., & Dornhaus, A. (2009). Spatial organization and division of labour in the bumblebee Bombus impatiens. Animal Behaviour, 77, 641–651. Kawaguchi, L. G., Ohashi, K., & Toquenaga, Y. (2006). Do bumble bees save time when choosing novel flowers by following conspecifics? Functional Ecology, 20, 239–244. Kikuchi, T. (1963). Studies on the coaction among insects visiting flowers. III. Dominance relationship among flower-visiting flies, bees and butterflies. Science Reports of Tohoku University – Fourth Series: Biology, 29, 1–8. Kirchner, W. H., & Röschard, J. (1999). Hissing in bumblebees: An interspecies defence signal. Insectes Sociaux, 46, 239–243. Leadbeater, E., & Chittka, L. (2007). The dynamics of social learning in an insect model, the bumblebee (Bombus terrestris). Behavioral Ecology and Sociobiology, 61, 1789–1796. Leadbeater, E., & Chittka, L. (2009). Bumble-bees learn the value of social cues through experience. Biology Letters, 5, 310–312. Lehrer, M., Horridge, G. A., Zhang, S. W., & Gadagkar, R. (1995). Shape vision in bees: Innate preference for flower-like patterns. Philosophical Transactions of the Royal Society B, 347, 123–137. Mackensen, O. (1947). Effect of carbon dioxide on initial oviposition of artificially inseminated and virgin queen bees. Journal of Economic Entomology, 40, 344–349. Menzel, R. (2012). The honeybee as a model for understanding the basis of cognition. Nature Reviews Neuroscience, 13, 758–768. Menzel, R., & Brembs, B. (2007). Cognition in invertebrates. In J. H. Kaas (Ed.), Evolution of nervous systems, Vol. II: Evolution of nervous systems in invertebrates (pp. 403–422). Oxford: Academic Press. Mirwan, H. B., & Kevan, P. G. (2014). Problem solving by worker bumblebees Bombus impatiens (Hymenoptera: Apoidea). Animal Cognition, http://dx.doi.org/10.1007/s10071-014-0737-0 Muth, F., Scampini, A. V., & Leonard, A. S. (2015). The effects of acute stress on learning and memory in bumble bees. Learning and Motivation (in this issue). Nicolas, G., & Sillans, D. (1989). Immediate and latent effects of carbon dioxide on insects. Annual Review of Entomology, 34, 97–116. Orbán, L. L., & Plowright, C. M. S. (2013). The effect of flower-like and non-flower-like visual properties on choice of unrewarding patterns by bumblebees. Naturwissenschaften, 100, 621–631. Orbán, L. L., & Plowright, C. M. S. (2014). Getting to the start line: How bumblebees and honeybees are visually guided towards their first floral contact. Insectes Sociaux, http://dx.doi.org/10.1007/s00040-014-0366-2 Pahl, M., Si, A., & Zhang, S. (2013). Numerical cognition in bees and other insects. Frontiers in Psychology, 4, 162. Peat, J., Tucker, J., & Goulson, D. (2005). Does intraspecific size variation in bumblebees allow colonies to efficiently exploit different flowers? Ecological Entomology, 30, 176–181. Plowright, C. M. S., Evans, S. A., Chew Leung, J., & Collin, C. A. (2011). The preference for symmetry in flower naïve and not-so-naïve bumblebees. Learning and Motivation, 42, 76–83. Plowright, C. M. S., Ferguson, K. A., Jellen, S. L., Xu, V., Service, E. W., & Dookie, A. L. (2013). Bumblebees occupy: When foragers do and do not use the presence of others to first find food. Insectes Sociaux, 60, 517–524. Plowright, C. M. S., Simonds, V. M., & Butler, M. A. (2006). How bumblebees first find flowers: Habituation of visual pattern preferences, spontaneous recovery and dishabituation. Learning and Motivation, 37, 66–78. Pomeroy, N., & Plowright, R. C. (1979). Larval ejection following narcosis of bumble bees (Hymenoptera: Apidae). Journal of the Kansas Entomological Society, 52, 215–217. Ribbands, C. R. (1950). Changes in the behaviour of honey-bees following their recovery from anaesthesia. Journal of Experimental Biology, 27, 302–310. Saleh, N., & Chittka, L. (2006). The importance of experience in the interpretation of conspecific chemical signals. Behavioral Ecology and Sociobiology, 61, 215–220. Sherry, D. F., & Strang, C. G. (2014). Contrasting styles in cognition and behaviour in bumblebees and honeybees. Behavioural Processes, http://dx.doi.org/ 10.1016/j.beproc.2014.09.005 Shettleworth, S. J. (2010). Cognition, evolution, and behavior (2nd ed.). New York: Oxford University Press., 720 pp. Simonds, V., & Plowright, C. M. S. (2004). How do bumblebees first find flowers? Unlearned approach responses and habituation. Animal Behaviour, 67, 379–386. Sokal, R. R., & Rohlf, J. F. (2012). Biometry (4th ed.). New York: W.H. Freeman. Spaethe, J., & Weidenmüller, A. (2002). Size variation and foraging rate in bumblebees (Bombus terrestris). Insectes Sociaux, 49, 142–146. Tepedino, V. J., & Parker, F. D. (1981). The effect of floral occupancy on the foraging of flower-visiting insects. Psyche, 88, 321–330. Thompson, E. L., & Plowright, C. M. S. (2014). How images may or may not represent flowers: Picture-object correspondence in bumblebees (Bombus impatiens)? Animal Cognition, 17, 1031–1043. Tinbergen, N. (1951). The study of instinct. London: Oxford University Press. Visscher, K. P., & Vetter, R. S. (1995). Smoke and target color effects on defensive behavior in yellow jacket wasps and bumble bees (Hymenoptera: Vespidae, Apidae) with a description of an electronic attack monitor. Journal of Economic Entomology, 88, 579–583.