- Email: [email protected]
Do mirrors reflect reality in agonistic encounters? A test of mutual cooperation in displays
Animal Behaviour 97 (2014) 63e67
Contents lists available at ScienceDirect
Animal Behaviour journal homepage: www.elsevier.com/locate/anbehav
Do mirrors reflect reality in agonistic encounters? A test of mutual cooperation in displays Robert W. Elwood a, *, Velizara Stoilova a, Amy McDonnell a, Ryan L. Earley b, Gareth Arnott a a b
School of Biological Sciences, Queen's University, Belfast, U.K. Department of Biological Sciences, University of Alabama, Tuscaloosa, AL, U.S.A.
a r t i c l e i n f o Article history: Received 11 June 2014 Initial acceptance 7 July 2014 Final acceptance 21 July 2014 Published online MS. number: 14-00477R Keywords: contest behaviour information exchange lateral displays lateralization mirror test
Animals frequently engage in mutual displays that may allow or at least help decisions about the outcome of agonistic encounters with mutual benefit to the opponents. In fish these often involve lateral displays, with previous studies finding evidence of population-level lateralization with a marked preference for showing the right side and using the right eye. Because both opponents tend to show this preference a head to tail configuration is formed and is used extensively during the display phase. Here we tested the significance of these lateral displays by comparing displays to a mirror with those to a real opponent behind a transparent barrier. The frequency of displays was lower to a mirror but the individual displays were of greater duration indicating a slower pace of the interaction with a mirror. This suggests that fish respond to initiatives of real opponents but as mirror images do not initiate moves the focal fish only moves when it is ready to change position. However, lateralization was still found with mirrors, indicating that the right-side bias is a feature of the individual and not of the interaction between opponents. We discuss implications for ideas about the evolution of mutual cooperation and information exchange in contests, as well as the utility of the use of mirrors in the study of aggression in fish. © 2014 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
The outcomes of animal contests ensure the unequal distribution of vital resources (Arnott & Elwood, 2008) and thus drive evolution (Elwood & Arnott, 2012). The manner by which the outcome is determined is the subject of much debate, particularly concerning the degree to which animals gather information about the opponent (Arnott & Elwood, 2009a; Elwood & Arnott, 2012, 2013; Fawcett & Mowles, 2013; Mesterton-Gibbons & Heap, 2014). Some game theory models stress the importance of the contestant only monitoring its own state, termed self-assessment (Arnott & Elwood, 2009a; Taylor & Elwood, 2003), e.g. the energetic war of attrition (Payne & Pagel, 1996, 1997) and the cumulative assessment models (Payne, 1998). Others, such as the sequential assessment model (SAM; Enquist & Leimar, 1983), are said to involve mutual assessment, and stress the importance of gathering information about the opponent (Enquist & Leimar, 1983). With SAM there has been a presumption that the information about an opponent is compared with information about
* Correspondence: R. W. Elwood, School of Biology & Biochemistry, The Queen's University of Belfast, 97 Lisburn Road, Belfast BT9 7BL, U.K. E-mail address: [email protected] (R. W. Elwood).
oneself (Elwood & Arnott, 2013; Fawcett & Mowles, 2013). However, it has been noted that simple systems might exist that do not require the cognitive ability of comparison of two values (Elwood & Arnott, 2012). Irrespective of the cognitive process, this apparent ability to compare one's own ability with that estimated for the opponent has a distinct advantage over self-assessment models because it should enable an animal to quit a contest as soon as it perceives that it is likely to lose. Because this might occur at an early display stage, before the contest has escalated, the animal could avoid the cost of a fight it would inevitably lose (Arnott & Elwood, 2009a). The winner would also gain because it could get the resource more quickly and without paying the cost of escalation (see Mesterton-Gibbons & Heap, 2014 for relative costs of mutual and self-assessment for individuals of different strengths). This leads to the idea that animals might cooperate in order to exchange information (Arnott, Ashton, & Elwood, 2011; Earley, 2010), enabling the contest to be resolved with minimum cost to both contestants (Arnott & Elwood, 2009a). This mutual benefit (West, Griffin, & Gardner, 2007) has been suggested as an important factor in the evolution of cooperative ritualized aggressive displays (Arnott et al., 2011). Such displays typically precede escalated physical contact; examples include the mutual vocal displays
http://dx.doi.org/10.1016/j.anbehav.2014.07.028 0003-3472/© 2014 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
64
R. W. Elwood et al. / Animal Behaviour 97 (2014) 63e67
occurring between male red deer, Cervus elaphus (Clutton-Brock & Albon, 1979) and fallow deer, Dama dama (Jennings, Elwood, Carlin, Hayden, & Gammell, 2012). These ungulates also engage in a conspicuous lateral visual display during contests, termed the parallel walk, whereby the deer walk in the same direction a short distance apart, often adopting a stiff-legged gait (Jennings & Gammell, 2013). This lateral display has been interpreted as providing a means for males to assess the competitive ability of opponents (e.g. Clutton-Brock, Albon, Gibson, & Guinness, 1979; Jennings & Gammell, 2013), and contests may be settled during these early display stages (Jennings, 2012). Fish also commonly show their flanks in early stages of a contest, engaging in so-called lateral displays, again typically interpreted as a means of allowing each contestant to observe the physical attributes of its opponent (Arnott & Elwood, 2009b, 2009c; Enquist, Leimar, Ljungberg, Mallner, & Segerdahl, 1990; Hurd, 1997). Fish, however, can align in two ways during lateral displays, with their heads either facing in the same direction (head to head) or in opposite directions (head to tail). Fish might also preferentially show one side to their opponent (Bisazza & de Santi, 2003; Reddon & Balshine, 2010). For example, competing convict cichlids, Amatitlania nigrofasciata, more commonly show their right than their left flank which results in the head to tail configuration being far more common than the head to head configuration (Arnott et al., 2011). It has been speculated that population-level lateralization of displays enables coordination of these agonistic interactions (Ghirlanda, Frasnelli, & Vallortigara, 2009). One way to test whether the head to tail posture is important in the coordination of displays is to prevent key features of the mutual display by the use of mirror images. Mirrors are frequently used as a substitute for live opponents in studies on aggression in a range of species (e.g. crayfish, Procambarus clarkii, May & Mercier, 2007; Japanese quail, Coturnix japonica, Hirschenhauser, Wittek, Johnston, & Mostl, 2008), and this approach is particularly popular with fish aggression studies (e.g. Balzarini, Taborsky, Wanner, Koch, & Frommen, 2014; Earley, Hsu, & Wolf, 2000; Tinbergen, 1951; Verbeek, Iwamoto, & Murakami, 2007; Wilson, de Boer, Arnott, & Grimmer, 2011). They are popular because each fish used is a focal fish so fewer animals might be required and pseudoreplication is avoided, and because mirrors provoke strong aggressive responses. Further, the use of mirrors avoids welfare problems that arise from two animals being placed together when one may harm the other (Elwood, 1991). However, the suitability of mirror-elicited behaviour as a means of predicting contest performance or provoking the same behavioural and physiological responses as real contests is beginning to be questioned. Mirror images fail to elicit the same brain gene expression (Desjardins & Fernald, 2010) and the same hormonal responses (Oliveira, Carneiro, & Carneiro, 2005; but see Dijkstra, Schaafsma, Hofmann, & Groothuis, 2012) as live opponents. Balzarini et al. (2014) compared the displays of three cichlid fish species to a mirror and to a real opponent behind a clear barrier. Only one species showed positive correlations between the two situations for a number of different aspects of displays, which again casts doubt on the validity of using mirrors to elicit responses that would be informative for predicting behaviour against live opponents. A key problem with using a mirror is that the reflection will not allow a head to tail posture and this might account for differences between displays to real opponents and mirror images (Arnott et al., 2011). A fish might turn to assume the head to tail posture but the reflection will turn simultaneously and the head to head alignment will persist. Thus, only the head to head posture is attained with a mirror and the common head to tail posture will be missing. This key element of mutual display, which appears to be
mediated by population-level lateralization, will be disrupted because the reflection fails to cooperate in the display (Arnott et al., 2011). Thus if fish repeatedly attempt to attain the head to tail posture we predict that with a mirror their responses will speed up resulting in more individual displays but with a shorter duration of each lateral display. Alternatively, a mirror might not elicit the same response as a real opponent because real opponents initiate moves that elicit responses from the focal fish (e.g. Van Dyk & Evans, 2008). Mirror images, of course, do not initiate moves and thus a fish displaying to a reflection will only generate its own moves rather than responding to new moves of the opponent. We might envisage situations in which a focal fish waits for the real opponent to make a move that it will then counter. This leads to an alternative prediction that, with a mirror, the focal fish will not change position so frequently leading to fewer but longer individual displays to a mirror. In this study we used convict cichlids to examine left lateral displays, right lateral displays and frontal displays of focal fish to mirrors and real opponents. In this species both sexes show territorial defence and lateralization of lateral displays with a preference for the right side (Arnott et al., 2011). We recorded the number of each display, the total duration of displays and the median duration of the displays in the two situations. This will reveal whether the fish faced with a mirror image speed up behavioural changes in an attempt to achieve a head to tail posture or slow down behavioural changes because the mirror does not initiate moves and therefore no counter displays are required. Further, if the mirror image fails to cooperate then the laterality noted with a real opponent might break down. Thus we also compared the laterality of displays to real opponents and mirror images. METHODS Twenty-six female, size-matched convict cichlids were obtained from a local supplier (Grosvenor Tropicals, Belfast, U.K.) in batches of six or seven and kept in individual glass tanks measuring 30 20 cm and 20 cm high with approximately 2 cm depth of gravel. A controlled artificial 12:12 h light:dark cycle was in place, the water aerated and kept to a depth of 15 cm and the temperature maintained at approximately 27 C. Tanks were aligned end to end (Fig. 1) with opaque partitions visually isolating the fish outside of test sessions. Fish were fed every other day with flake food and on the experimentation day they were fed after observations had taken place. The fish were maintained isolated in their tanks for 1e2 weeks before experimentation, thus ameliorating any
Figure 1. Experimental set-up. The two holding tanks were identical in condition and content and the grey opaque partition situated between the tanks was removable either to expose the focal fish to a real opponent or to allow insertion of a mirror. The fish coloured black is the focal fish in the procedure.
R. W. Elwood et al. / Animal Behaviour 97 (2014) 63e67
behavioural effects of prior winning or losing experiences (reviewed in Hsu, Earley, & Wolf, 2006). Each subject was tested twice, once against a mirror and once against a real opponent, in a random order and with a gap of 10e15 min between tests. When observations against the real opponents were carried out the opaque divider was removed from between two tanks and the focal fish was filmed for 30 min. When a mirror was used the divider was removed and immediately replaced with a 20 20 cm mirror and the focal fish was filmed for 30 min. Each focal fish was exposed to the stimulus fish in the tank to the right except for the last in the row of tanks, which was moved so that it could see the first fish as the stimulus 24 h after moving. Thus all focal fish/stimulus fish combinations were novel and pseudoreplication was avoided. The laboratory was isolated from disturbance during filming. Behavioural Assessment The films were then observed and behavioural displays recorded using Observer v. 3.0 software (Noldus Technology, Wageningen, The Netherlands). Activities recorded were right lateral displays, left lateral displays and frontal displays. A right lateral display was recorded if the fish was displaying its right lateral side at 45 or less to the glass at the end of the tank closest to the mirror/opponent, or a left lateral display if the left side was shown at 45 or less and a frontal display was recorded if the fish was head on to the glass (>45 ). Ethical Note After discussion with the local Home Office veterinary inspector it was decided that there was no likelihood of fish being harmed from the intended procedure and thus no licence was required. Maintenance of fish in individual tanks meant that they did not require capture and transfer to other tanks for the specific tests and thus reduced stress and damage. We confirm that no harm occurred during the experiment and the fish were passed for use in another noninvasive study. Statistical Analyses For each type of display (left, right and frontal) we obtained the number used by each fish, the total duration for each display for each fish and the median duration of each display for each fish. The data were not normally distributed and the nonparametric Wilcoxon matched-pairs signed-ranks test for nonindependent data was used to compare responses to real opponents and mirror images. The same test was used to compare number, total duration and median durations of left and right lateral displays. Spearman rank correlations were used to examine relations between mirrors and real opponents for each display component. All analyses were carried out using StatView (SAS Institute Inc., Cary, NC, U.S.A.). RESULTS Comparison of Displays towards Real Opponent and Mirror Image Real opponents elicited more displays than did mirror images for left lateral displays (real left frequency median ¼ 139.5, mirror left frequency median ¼ 57.0, z ¼ 3.34, P < 0.001), right lateral displays (real right frequency median ¼ 145.5, mirror right frequency median ¼ 76.0, z ¼ 3.75, P < 0.001) and frontal displays (real frontal frequency median ¼ 117.5, mirror frontal frequency median ¼ 90.0, z ¼ 2.25, P ¼ 0.025).
65
Real opponents also elicited a greater total duration of displays than did mirror images for left lateral displays (real left duration median ¼ 339.7 s, mirror left duration median ¼ 224.6 s, z ¼ 2.96, P ¼ 0.003) but not for right lateral displays (real right duration median ¼ 442.9 s, mirror right duration mean ¼ 354.6 s, z ¼ 1.69, P ¼ 0.091) or frontal displays (real frontal duration median ¼ 322.1 s, mirror frontal duration median ¼ 347.9 s, z ¼ 0.292, P ¼ 0.770). The median duration of displays did not differ for left lateral displays to a real opponent or a mirror image (real left duration median ¼ 2.56 s, mirror left duration median ¼ 3.11 s, z ¼ 1.59, P ¼ 0.112). However, for right lateral displays, the median duration to a mirror was greater than to an opponent (real right duration median ¼ 3.26 s, mirror right duration median ¼ 4.22 s, z ¼ 2.35, P ¼ 0.019). There was no significant difference between the median duration of frontal displays when displaying against a real opponent or a mirror image (real frontal duration median ¼ 2.59 s, mirror frontal duration median ¼ 3.23 s, z ¼ 1.6, P ¼ 0.107). There were significant positive correlations of various display measures between mirror images and real opponents, these being found for the number of left lateral (rs ¼ 0.54, P ¼ 0.007), right lateral (rs ¼ 0.48, P ¼ 0.016) but not frontal displays (rs ¼ 0.30, P ¼ 0.138). Positive correlations were found for the total duration of left lateral (rs ¼ 0.69, P < 0.001), right lateral (rs ¼ 0.59, P ¼ 0.003) and frontal displays (rs ¼ 0.59, P ¼ 0.003). However, there were no significant correlations for the median display durations for left lateral (rs ¼ 0.24, P ¼ 0.247), right lateral (rs ¼ 0.08, P ¼ 0.71) or frontal displays (rs ¼ 0.32, P ¼ 0.10). Examination of Lateralization to Real Opponents and Mirrors There was a small but nevertheless significant right-side bias for the number of lateral displays to a real opponent (left frequency median ¼ 139.5, right frequency median ¼ 145.5, z ¼ 2.92, P ¼ 0.004) but not to a mirror image (left frequency median ¼ 57.0, right frequency median ¼ 76.0, z ¼ 1.53, P ¼ 0.126). There was also a longer total duration displaying the right side as opposed to the left to a real opponent (left duration median ¼ 339.7 s, right duration median ¼ 442.9 s, z ¼ 2.86, P ¼ 0.004) and also when looking at a mirror image (left duration median ¼ 224.6 s, right duration median ¼ 354.6 s, z ¼ 2.88, P ¼ 0.004). However, the median duration of displays to a real opponent did not differ between left and right (left duration median ¼ 2.56 s, right duration median ¼ 3.26 s, z ¼ 2.40, P ¼ 0.164), but was significantly longer with right than left displays to a mirror image (left duration median ¼ 3.11 s, right duration median ¼ 4.22 s, z ¼ 2.02, P ¼ 0.044). DISCUSSION Two contrasting hypotheses, each with clear predictions, were investigated in the current study. First, it was hypothesized that if fish repeatedly attempt to attain the head to tail posture then in the mirror treatment their responses will speed up resulting in more individual displays but a shorter duration of each lateral display. There was no support for this prediction. Second, it was suggested that the mirror might not elicit the same response as a real opponent since mirror images do not initiate moves to which the focal fish would respond. Thus, in the mirror treatment, the focal fish should not change position as frequently, resulting in fewer displays but longer individual displays to a mirror. There was firm support for this hypothesis. Real opponents elicited more frequent right and left lateral displays and frontal displays than did a mirror, consistent with the results of Balzarini et al. (2014) for Neolamprologus pulcher. Real opponents also elicited a greater total duration of left lateral displays
66
R. W. Elwood et al. / Animal Behaviour 97 (2014) 63e67
although not for right lateral or frontal displays. Further, the median duration of right lateral displays was greater towards a mirror than towards a real opponent. This last result is important because it indicates a slower change from one display type to another when a mirror image is seen and thus supports the idea that the fish react to behavioural change in real opponents and hence speed up transitions between displays. That the difference between conditions was found for the right lateral display but not for the left might indicate that the fish is particularly motivated to show the right side or use the right eye and that a mirror image does not induce the same behavioural changes as a real opponent. While the median duration of left lateral and frontal displays did not differ significantly between mirror and real opponent both were in the direction of longer displays to the mirror. Obviously, the mirror does not initiate behavioural change and thus fails to reflect reality when compared with a real opponent. We thus question the validity of using mirrors as a proxy for a live opponent and note that the slower pace of the mirror contest might explain findings of different brain gene expression (Desjardins & Fernald, 2010) and hormonal responses (Oliveira et al., 2005) when mirror and real opponents are compared. The observed change in contest dynamics, of fewer and longer displays, when facing a mirror image raises questions about how this would affect other aggressive activities. One could speculate that in this abnormal scenario individuals may abandon lateral displays for more escalated aggressive activities, such as biting. A study in lizards provides some support for this assertion with the finding that male Jacky dragons, Amphibolurus muricatus, facing an interactive video playback that mirrored their own aggressive behaviour had higher attack rates than in treatments that incorporated more appropriate video responses to the signal given by the test individuals (e.g. submission following an aggressive display; Van Dyk & Evans, 2008). In the current study, various display measures were positively correlated between mirror images and real opponents, including the number of left and right lateral displays, and the total durations of right and left lateral and frontal displays. Similarly, for N. pulcher, Balzarini et al. (2014) found correlations between aggressive behaviour directed towards a mirror and a real opponent. These findings are perhaps best viewed as highlighting the consistency of aggressiveness within individuals, an aspect of personality that is well recognized in fish (e.g. Wilson et al., 2011). Thus, in the context of personality research, the use of mirrors to assay aggressiveness is still a technique with merit, particularly because it avoids the need to face a real opponent, with the potential ethical concerns that such encounters may entail (e.g. Elwood, 1991). However, despite the correlations, there are still important differences in behaviour when comparing real and mirror image interactions, as revealed in the current study and that of Balzarini et al. (2014). In addition to the differences in frequencies and durations noted above there was no significant correlation between the two conditions for the number of frontal displays or the median duration of any type of display. Thus, we suggest that while the mirror test provides a somewhat crude measure of aggression, it fails to enable fish to interact in the appropriate way and is therefore not applicable as a method when researchers are interested in examining contest behaviour or contest resolution (Arnott & Elwood, 2009a; Earley et al., 2000; Elwood & Arnott, 2012). The present study confirms previous work (Arnott et al., 2011) demonstrating population-level lateralization, with a bias to present the right flank and/or use the right eye during aggressive lateral displays to an opponent beyond a transparent partition. We have demonstrated here that a right-side bias is also evident when displaying to a mirror, in terms of both total duration and median duration. With a real opponent this laterality of displays naturally results in the head to tail posture and thus is suggestive of
cooperation between opponents (Arnott et al., 2011). That is, laterality and cooperation were thought to be linked. However, our present results show that laterality is not driven by the focal fish seeing the opponent's right side and adopting the right-side posture in return, to achieve the head to tail configuration, because that configuration does not occur with the mirror. Clearly, the right-side bias is a feature of the individual and not of the interaction between opponents. Thus our previous suggestion (Arnott et al., 2011) that laterality is an important component of possible cooperation in information exchange must now be questioned because laterality still occurs with an uncooperative mirror image. Because lateralization of these agonistic interactions is so clear we suspect that showing the right side and using the right eye to view an opponent confers an advantage, particularly as a right-eye preference during aggressive encounters has been found by other studies in fish (e.g. Bisazza & de Santi, 2003; Reddon & Balshine, 2010). The use of visual information acquired via the right eye has been implicated in decision making (Rogers, 2002; Rogers, Vallortigara, & Andrew, 2013), thus lending further support to the idea that lateralization may play an important role in contest behaviour. Additionally, while results from the current study suggest mutual cooperation in displays may not have been the selective force for such lateralization, it could still emerge as a consequence, and when facing a real opponent this positioning may be a key feature of the interaction. Indeed, lateralization has been associated with predictable social behaviour, with a bias in the direction in which individuals approach each other, an important aspect of social interactions (e.g. Baraud, Buytet, Bec, & Blois-Heulin, 2009). An alternative explanation for lateralized displays towards mirror images could be that selection has acted so strongly on lateralization as a mechanism to engage in cooperative displays and reduce contest costs that the strategy remains fixed regardless of opponent orientation (or cooperative tendency). Individuals would rarely, if ever, encounter under natural circumstances the type of stimulus that a mirror image provides. Studies in fallow deer have confirmed the importance of lateralization during contests, with Jennings (2012) observing a rightsided bias to terminate parallel walks, interpreted as evidence of hemispheric specialization for processing information about the opponent during the visual display phase. However, recently, Jennings (2014a) found no evidence that more lateralized individuals (in terms of the decision to terminate the parallel walk) benefitted from reduced fight costs during escalated physical fighting, and contrary to initial predictions there was a positive relationship between laterality and time spent in escalated fighting. This suggests that the population-level lateralization observed during lateral displays in fish and deer may serve to facilitate this early phase of the contest, rather than more escalated phases. Adding to the complex picture of lateralization and aggressive behaviour in fallow deer, Jennings (2014b) found that weakly lateralized individuals were more likely to conduct third-party interventions of ongoing fights. However, individuals that showed lateralization for right-eye use were less likely to be targeted by an intervening male, an important benefit because targeted males are highly likely to be defeated by the intervener (Jennings, Carlin, & Gammell, 2009). It was also the case that less lateralized individuals achieved the greatest number of copulations. These studies highlight that the fitness effects of lateralization are context dependent (Rogers et al., 2013). The majority of studies on lateralization and aggression have focused on escalated attacks, documenting a left-eyeeright-hemisphere bias (Vallortigara & Rogers, 2005). Our studies and those of Jennings (see above) are the first to explore the role of lateralization during ritualized display phases of the contest. In fish the use of mirrors to disrupt the normal orientation of contestants has clear
R. W. Elwood et al. / Animal Behaviour 97 (2014) 63e67
effects on the frequency and durations of specific displays and thus mirrors do not reflect reality. Acknowledgments We are grateful to Gillian Riddell for technical assistance. We also thank two anonymous referees for their excellent suggestions for improving the manuscript. References Arnott, G., & Elwood, R. W. (2008). Information gathering and decision making about resource value in animal contests. Animal Behaviour, 76, 529e542. Arnott, G., & Elwood, R. W. (2009a). Assessment of fighting ability in animal contests. Animal Behaviour, 77, 991e1004. Arnott, G., & Elwood, R. W. (2009b). Gender differences in aggressive behaviour in convict cichlids. Animal Behaviour, 78, 1221e1227. Arnott, G., & Elwood, R. W. (2009c). Probing aggressive motivation in a cichlid fish. Biology Letters, 5, 762e764. Arnott, G., Ashton, C., & Elwood, R. W. (2011). Lateralization of lateral displays in convict cichlids. Biology Letters, 7, 683e685. Balzarini, V., Taborsky, M., Wanner, S., Koch, F., & Frommen, J. G. (2014). Mirror, mirror on the wall: the predictive value of mirror tests for measuring aggression in fish. Behavioral Ecology and Sociobiology, 68, 871e878. Baraud, I., Buytet, B., Bec, P., & Blois-Heulin, C. (2009). Social laterality and ‘transversality’ in two species of mangabeys: influence of rank and implication for hemispheric specialization. Behavioural Brain Research, 198, 449e458. Bisazza, A., & de Santi, A. (2003). Lateralization of aggression in fish. Behavioural Brain Research, 141, 131e136. Clutton-Brock, T. H., & Albon, S. D. (1979). The roaring of red deer and the evolution of honest advertisement. Behaviour, 69, 145e170. Clutton-Brock, T. H., Albon, S. D., Gibson, R. M., & Guinness, F. E. (1979). The logical stag: adaptive aspects of fighting in red deer. Animal Behaviour, 27, 211e255. Desjardins, J. K., & Fernald, R. D. (2010). What do fish make of mirror images? Biology Letters, 6, 744e747. Dijkstra, P. D., Schaafsma, S. M., Hofmann, H. A., & Groothuis, T. G. G. (2012). ‘Winner effect’ without winning: unresolved social conflicts increase the probability of winning a subsequent contest in a cichlid fish. Physiology & Behavior, 105, 489e492. Earley, R. L. (2010). Social eavesdropping and the evolution of conditional cooperation and cheating strategies. Philosophical Transactions of the Royal Society of London, Series B, 365, 2675e2686. Earley, R. L., Hsu, Y., & Wolf, L. L. (2000). The use of standard aggression testing methods to predict combat behaviour and contest outcome in Rivulus marmoratus dyads (Teleostei: Cyprinodontidae). Ethology, 106, 743e761. Elwood, R. W. (1991). Ethical implications of studies on infanticide and maternal aggression in rodents. Animal Behaviour, 42, 841e849. Elwood, R. W., & Arnott, G. (2012). Understanding how animals fight with Lloyd Morgan's canon. Animal Behaviour, 84, 1095e1102. Elwood, R. W., & Arnott, G. (2013). Assessments in contests are frequently assumed to be complex when simple explanations will suffice. Animal Behaviour, 86, e8ee12. Enquist, M., & Leimar, O. (1983). Evolution of fighting behaviour: decision rules and assessment of relative strength. Journal of Theoretical Biology, 102, 387e410. Enquist, M., Leimar, O., Ljungberg, T., Mallner, Y., & Segerdahl, N. (1990). A test of the sequential assessment game: fighting in the cichlid fish Nannacara anomala. Animal Behaviour, 40, 1e14. Fawcett, T. W., & Mowles, S. L. (2013). Assessments of fighting ability need not be cognitively complex. Animal Behaviour, 86, e1ee17. Ghirlanda, S., Frasnelli, E., & Vallortigara, G. (2009). Intraspecific competition and coordination in the evolution of lateralization. Philosophical Transactions of the Royal Society of London, Series B, 364, 861e866.
67
Hirschenhauser, K., Wittek, M., Johnston, P., & Mostl, E. (2008). Social context rather than behavioral output or winning modulates post-conflict testosterone responses in Japanese quail (Coturnix japonica). Physiology and Behaviour, 95, 457e463. Hsu, Y. Y., Earley, R. L., & Wolf, L. L. (2006). Modulation of aggressive behaviour by fighting experience: mechanisms and contest outcomes. Biological Reviews, 81, 33e74. http://dx.doi.org/10.1017/S146479310500686X. Hurd, P. L. (1997). Cooperative signalling between opponents in fish fights. Animal Behaviour, 54, 1309e1315. Jennings, D. J. (2012). Right-sided bias in fallow deer terminating parallel walks: evidence for lateralization during a lateral display. Animal Behaviour, 83, 1427e1432. Jennings, D. J. (2014a). Information gathering during contests: the relationship between lateralisation and contestant behaviour during fallow deer fights. Behavioural Processes, 103, 278e282. Jennings, D. J. (2014b). Limited evidence that visual lateralization is associated with fitness in rutting male fallow deer. Animal Behaviour, 92, 85e91. Jennings, D. J., Carlin, C. M., & Gammell, M. P. (2009). A winner effect supports third party intervention behaviour during fallow deer, Dama dama, fights. Animal Behaviour, 77, 343e348. Jennings, D. J., Elwood, R. W., Carlin, C. M., Hayden, T. J., & Gammell, M. P. (2012). Vocal rate as an assessment process during fallow deer contests. Behavioural Processes, 91, 152e158. Jennings, D. J., & Gammell, M. P. (2013). Ungulate contest behaviour. In I. C. W. Hardy, & M. Briffa (Eds.), Animal contests (pp. 304e320). Cambridge, U.K.: Cambridge University Press. May, H. Y., & Mercier, A. J. (2007). Duration of socialization influences responses to a mirror: responses of dominant and subordinate crayfish diverge with time of pairing. Journal of Experimental Biology, 210, 4428e4436. Mesterton-Gibbons, M., & Heap, S. M. (2014). Variation between self- and mutual assessment in animal contests. American Naturalist, 183, 199e213. Oliveira, R. F., Carneiro, L. A., & Carneiro, A. V. M. (2005). No hormonal response in tied fights. Nature, 437, 207e208. Payne, R. J. H. (1998). Gradually escalating fights and displays: the cumulative assessment model. Animal Behaviour, 56, 651e662. Payne, R. J. H., & Pagel, M. (1996). Escalation and time costs in displays of endurance. Journal of Theoretical Biology, 183, 185e193. Payne, R. J. H., & Pagel, M. (1997). Why do animals repeat displays? Animal Behaviour, 54, 109e119. Reddon, A. R., & Balshine, S. (2010). Lateralization in response to social stimuli in a cooperatively breeding cichlid fish. Behavioural Processes, 85, 68e71. Rogers, L. J. (2002). Advantages and disadvantages of lateralization. In L. J. Rogers, & R. J. Andrew (Eds.), Comparative vertebrate lateralization (pp. 126e153). Cambridge, U.K.: Cambridge University Press. Rogers, L. J., Vallortigara, C., & Andrew, R. J. (2013). Divided brains: The biology and behaviour of brain asymmetries. Cambridge, U.K.: Cambridge University Press. Taylor, P. W., & Elwood, R. W. (2003). The mismeasure of animal contests. Animal Behaviour, 65, 1195e1202. Tinbergen, N. (1951). The study of instinct. New York: Clarendon/Oxford University Press. Vallortigara, G., & Rogers, L. J. (2005). Survival with an asymmetrical brain: advantages and disadvantages of cerebral lateralization. Behavioral and Brain Sciences, 28, 575e633. Van Dyk, D. A., & Evans, C. S. (2008). Opponent assessment in lizards: examining the effect of aggressive and submissive signals. Behavioural Ecology, 19, 895e901. Verbeek, P., Iwamoto, T., & Murakami, N. (2007). Differences in aggression between wild-type and domesticated fighting fish are context dependent. Animal Behaviour, 73, 75e83. West, S. A., Griffin, A. S., & Gardner, A. (2007). Social semantics: altruism, cooperation, mutualism, strong reciprocity and group selection. Journal of Evolutionary Biology, 20, 415e432. Wilson, A. J., de Boer, M., Arnott, G., & Grimmer, A. (2011). Integrating personality research and animal contest theory: aggressiveness in the green swordtail Xiphophorus helleri. PLoS One, 6, e28024.
Contents lists available at ScienceDirect
Animal Behaviour journal homepage: www.elsevier.com/locate/anbehav
Do mirrors reflect reality in agonistic encounters? A test of mutual cooperation in displays Robert W. Elwood a, *, Velizara Stoilova a, Amy McDonnell a, Ryan L. Earley b, Gareth Arnott a a b
School of Biological Sciences, Queen's University, Belfast, U.K. Department of Biological Sciences, University of Alabama, Tuscaloosa, AL, U.S.A.
a r t i c l e i n f o Article history: Received 11 June 2014 Initial acceptance 7 July 2014 Final acceptance 21 July 2014 Published online MS. number: 14-00477R Keywords: contest behaviour information exchange lateral displays lateralization mirror test
Animals frequently engage in mutual displays that may allow or at least help decisions about the outcome of agonistic encounters with mutual benefit to the opponents. In fish these often involve lateral displays, with previous studies finding evidence of population-level lateralization with a marked preference for showing the right side and using the right eye. Because both opponents tend to show this preference a head to tail configuration is formed and is used extensively during the display phase. Here we tested the significance of these lateral displays by comparing displays to a mirror with those to a real opponent behind a transparent barrier. The frequency of displays was lower to a mirror but the individual displays were of greater duration indicating a slower pace of the interaction with a mirror. This suggests that fish respond to initiatives of real opponents but as mirror images do not initiate moves the focal fish only moves when it is ready to change position. However, lateralization was still found with mirrors, indicating that the right-side bias is a feature of the individual and not of the interaction between opponents. We discuss implications for ideas about the evolution of mutual cooperation and information exchange in contests, as well as the utility of the use of mirrors in the study of aggression in fish. © 2014 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
The outcomes of animal contests ensure the unequal distribution of vital resources (Arnott & Elwood, 2008) and thus drive evolution (Elwood & Arnott, 2012). The manner by which the outcome is determined is the subject of much debate, particularly concerning the degree to which animals gather information about the opponent (Arnott & Elwood, 2009a; Elwood & Arnott, 2012, 2013; Fawcett & Mowles, 2013; Mesterton-Gibbons & Heap, 2014). Some game theory models stress the importance of the contestant only monitoring its own state, termed self-assessment (Arnott & Elwood, 2009a; Taylor & Elwood, 2003), e.g. the energetic war of attrition (Payne & Pagel, 1996, 1997) and the cumulative assessment models (Payne, 1998). Others, such as the sequential assessment model (SAM; Enquist & Leimar, 1983), are said to involve mutual assessment, and stress the importance of gathering information about the opponent (Enquist & Leimar, 1983). With SAM there has been a presumption that the information about an opponent is compared with information about
* Correspondence: R. W. Elwood, School of Biology & Biochemistry, The Queen's University of Belfast, 97 Lisburn Road, Belfast BT9 7BL, U.K. E-mail address: [email protected] (R. W. Elwood).
oneself (Elwood & Arnott, 2013; Fawcett & Mowles, 2013). However, it has been noted that simple systems might exist that do not require the cognitive ability of comparison of two values (Elwood & Arnott, 2012). Irrespective of the cognitive process, this apparent ability to compare one's own ability with that estimated for the opponent has a distinct advantage over self-assessment models because it should enable an animal to quit a contest as soon as it perceives that it is likely to lose. Because this might occur at an early display stage, before the contest has escalated, the animal could avoid the cost of a fight it would inevitably lose (Arnott & Elwood, 2009a). The winner would also gain because it could get the resource more quickly and without paying the cost of escalation (see Mesterton-Gibbons & Heap, 2014 for relative costs of mutual and self-assessment for individuals of different strengths). This leads to the idea that animals might cooperate in order to exchange information (Arnott, Ashton, & Elwood, 2011; Earley, 2010), enabling the contest to be resolved with minimum cost to both contestants (Arnott & Elwood, 2009a). This mutual benefit (West, Griffin, & Gardner, 2007) has been suggested as an important factor in the evolution of cooperative ritualized aggressive displays (Arnott et al., 2011). Such displays typically precede escalated physical contact; examples include the mutual vocal displays
http://dx.doi.org/10.1016/j.anbehav.2014.07.028 0003-3472/© 2014 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
64
R. W. Elwood et al. / Animal Behaviour 97 (2014) 63e67
occurring between male red deer, Cervus elaphus (Clutton-Brock & Albon, 1979) and fallow deer, Dama dama (Jennings, Elwood, Carlin, Hayden, & Gammell, 2012). These ungulates also engage in a conspicuous lateral visual display during contests, termed the parallel walk, whereby the deer walk in the same direction a short distance apart, often adopting a stiff-legged gait (Jennings & Gammell, 2013). This lateral display has been interpreted as providing a means for males to assess the competitive ability of opponents (e.g. Clutton-Brock, Albon, Gibson, & Guinness, 1979; Jennings & Gammell, 2013), and contests may be settled during these early display stages (Jennings, 2012). Fish also commonly show their flanks in early stages of a contest, engaging in so-called lateral displays, again typically interpreted as a means of allowing each contestant to observe the physical attributes of its opponent (Arnott & Elwood, 2009b, 2009c; Enquist, Leimar, Ljungberg, Mallner, & Segerdahl, 1990; Hurd, 1997). Fish, however, can align in two ways during lateral displays, with their heads either facing in the same direction (head to head) or in opposite directions (head to tail). Fish might also preferentially show one side to their opponent (Bisazza & de Santi, 2003; Reddon & Balshine, 2010). For example, competing convict cichlids, Amatitlania nigrofasciata, more commonly show their right than their left flank which results in the head to tail configuration being far more common than the head to head configuration (Arnott et al., 2011). It has been speculated that population-level lateralization of displays enables coordination of these agonistic interactions (Ghirlanda, Frasnelli, & Vallortigara, 2009). One way to test whether the head to tail posture is important in the coordination of displays is to prevent key features of the mutual display by the use of mirror images. Mirrors are frequently used as a substitute for live opponents in studies on aggression in a range of species (e.g. crayfish, Procambarus clarkii, May & Mercier, 2007; Japanese quail, Coturnix japonica, Hirschenhauser, Wittek, Johnston, & Mostl, 2008), and this approach is particularly popular with fish aggression studies (e.g. Balzarini, Taborsky, Wanner, Koch, & Frommen, 2014; Earley, Hsu, & Wolf, 2000; Tinbergen, 1951; Verbeek, Iwamoto, & Murakami, 2007; Wilson, de Boer, Arnott, & Grimmer, 2011). They are popular because each fish used is a focal fish so fewer animals might be required and pseudoreplication is avoided, and because mirrors provoke strong aggressive responses. Further, the use of mirrors avoids welfare problems that arise from two animals being placed together when one may harm the other (Elwood, 1991). However, the suitability of mirror-elicited behaviour as a means of predicting contest performance or provoking the same behavioural and physiological responses as real contests is beginning to be questioned. Mirror images fail to elicit the same brain gene expression (Desjardins & Fernald, 2010) and the same hormonal responses (Oliveira, Carneiro, & Carneiro, 2005; but see Dijkstra, Schaafsma, Hofmann, & Groothuis, 2012) as live opponents. Balzarini et al. (2014) compared the displays of three cichlid fish species to a mirror and to a real opponent behind a clear barrier. Only one species showed positive correlations between the two situations for a number of different aspects of displays, which again casts doubt on the validity of using mirrors to elicit responses that would be informative for predicting behaviour against live opponents. A key problem with using a mirror is that the reflection will not allow a head to tail posture and this might account for differences between displays to real opponents and mirror images (Arnott et al., 2011). A fish might turn to assume the head to tail posture but the reflection will turn simultaneously and the head to head alignment will persist. Thus, only the head to head posture is attained with a mirror and the common head to tail posture will be missing. This key element of mutual display, which appears to be
mediated by population-level lateralization, will be disrupted because the reflection fails to cooperate in the display (Arnott et al., 2011). Thus if fish repeatedly attempt to attain the head to tail posture we predict that with a mirror their responses will speed up resulting in more individual displays but with a shorter duration of each lateral display. Alternatively, a mirror might not elicit the same response as a real opponent because real opponents initiate moves that elicit responses from the focal fish (e.g. Van Dyk & Evans, 2008). Mirror images, of course, do not initiate moves and thus a fish displaying to a reflection will only generate its own moves rather than responding to new moves of the opponent. We might envisage situations in which a focal fish waits for the real opponent to make a move that it will then counter. This leads to an alternative prediction that, with a mirror, the focal fish will not change position so frequently leading to fewer but longer individual displays to a mirror. In this study we used convict cichlids to examine left lateral displays, right lateral displays and frontal displays of focal fish to mirrors and real opponents. In this species both sexes show territorial defence and lateralization of lateral displays with a preference for the right side (Arnott et al., 2011). We recorded the number of each display, the total duration of displays and the median duration of the displays in the two situations. This will reveal whether the fish faced with a mirror image speed up behavioural changes in an attempt to achieve a head to tail posture or slow down behavioural changes because the mirror does not initiate moves and therefore no counter displays are required. Further, if the mirror image fails to cooperate then the laterality noted with a real opponent might break down. Thus we also compared the laterality of displays to real opponents and mirror images. METHODS Twenty-six female, size-matched convict cichlids were obtained from a local supplier (Grosvenor Tropicals, Belfast, U.K.) in batches of six or seven and kept in individual glass tanks measuring 30 20 cm and 20 cm high with approximately 2 cm depth of gravel. A controlled artificial 12:12 h light:dark cycle was in place, the water aerated and kept to a depth of 15 cm and the temperature maintained at approximately 27 C. Tanks were aligned end to end (Fig. 1) with opaque partitions visually isolating the fish outside of test sessions. Fish were fed every other day with flake food and on the experimentation day they were fed after observations had taken place. The fish were maintained isolated in their tanks for 1e2 weeks before experimentation, thus ameliorating any
Figure 1. Experimental set-up. The two holding tanks were identical in condition and content and the grey opaque partition situated between the tanks was removable either to expose the focal fish to a real opponent or to allow insertion of a mirror. The fish coloured black is the focal fish in the procedure.
R. W. Elwood et al. / Animal Behaviour 97 (2014) 63e67
behavioural effects of prior winning or losing experiences (reviewed in Hsu, Earley, & Wolf, 2006). Each subject was tested twice, once against a mirror and once against a real opponent, in a random order and with a gap of 10e15 min between tests. When observations against the real opponents were carried out the opaque divider was removed from between two tanks and the focal fish was filmed for 30 min. When a mirror was used the divider was removed and immediately replaced with a 20 20 cm mirror and the focal fish was filmed for 30 min. Each focal fish was exposed to the stimulus fish in the tank to the right except for the last in the row of tanks, which was moved so that it could see the first fish as the stimulus 24 h after moving. Thus all focal fish/stimulus fish combinations were novel and pseudoreplication was avoided. The laboratory was isolated from disturbance during filming. Behavioural Assessment The films were then observed and behavioural displays recorded using Observer v. 3.0 software (Noldus Technology, Wageningen, The Netherlands). Activities recorded were right lateral displays, left lateral displays and frontal displays. A right lateral display was recorded if the fish was displaying its right lateral side at 45 or less to the glass at the end of the tank closest to the mirror/opponent, or a left lateral display if the left side was shown at 45 or less and a frontal display was recorded if the fish was head on to the glass (>45 ). Ethical Note After discussion with the local Home Office veterinary inspector it was decided that there was no likelihood of fish being harmed from the intended procedure and thus no licence was required. Maintenance of fish in individual tanks meant that they did not require capture and transfer to other tanks for the specific tests and thus reduced stress and damage. We confirm that no harm occurred during the experiment and the fish were passed for use in another noninvasive study. Statistical Analyses For each type of display (left, right and frontal) we obtained the number used by each fish, the total duration for each display for each fish and the median duration of each display for each fish. The data were not normally distributed and the nonparametric Wilcoxon matched-pairs signed-ranks test for nonindependent data was used to compare responses to real opponents and mirror images. The same test was used to compare number, total duration and median durations of left and right lateral displays. Spearman rank correlations were used to examine relations between mirrors and real opponents for each display component. All analyses were carried out using StatView (SAS Institute Inc., Cary, NC, U.S.A.). RESULTS Comparison of Displays towards Real Opponent and Mirror Image Real opponents elicited more displays than did mirror images for left lateral displays (real left frequency median ¼ 139.5, mirror left frequency median ¼ 57.0, z ¼ 3.34, P < 0.001), right lateral displays (real right frequency median ¼ 145.5, mirror right frequency median ¼ 76.0, z ¼ 3.75, P < 0.001) and frontal displays (real frontal frequency median ¼ 117.5, mirror frontal frequency median ¼ 90.0, z ¼ 2.25, P ¼ 0.025).
65
Real opponents also elicited a greater total duration of displays than did mirror images for left lateral displays (real left duration median ¼ 339.7 s, mirror left duration median ¼ 224.6 s, z ¼ 2.96, P ¼ 0.003) but not for right lateral displays (real right duration median ¼ 442.9 s, mirror right duration mean ¼ 354.6 s, z ¼ 1.69, P ¼ 0.091) or frontal displays (real frontal duration median ¼ 322.1 s, mirror frontal duration median ¼ 347.9 s, z ¼ 0.292, P ¼ 0.770). The median duration of displays did not differ for left lateral displays to a real opponent or a mirror image (real left duration median ¼ 2.56 s, mirror left duration median ¼ 3.11 s, z ¼ 1.59, P ¼ 0.112). However, for right lateral displays, the median duration to a mirror was greater than to an opponent (real right duration median ¼ 3.26 s, mirror right duration median ¼ 4.22 s, z ¼ 2.35, P ¼ 0.019). There was no significant difference between the median duration of frontal displays when displaying against a real opponent or a mirror image (real frontal duration median ¼ 2.59 s, mirror frontal duration median ¼ 3.23 s, z ¼ 1.6, P ¼ 0.107). There were significant positive correlations of various display measures between mirror images and real opponents, these being found for the number of left lateral (rs ¼ 0.54, P ¼ 0.007), right lateral (rs ¼ 0.48, P ¼ 0.016) but not frontal displays (rs ¼ 0.30, P ¼ 0.138). Positive correlations were found for the total duration of left lateral (rs ¼ 0.69, P < 0.001), right lateral (rs ¼ 0.59, P ¼ 0.003) and frontal displays (rs ¼ 0.59, P ¼ 0.003). However, there were no significant correlations for the median display durations for left lateral (rs ¼ 0.24, P ¼ 0.247), right lateral (rs ¼ 0.08, P ¼ 0.71) or frontal displays (rs ¼ 0.32, P ¼ 0.10). Examination of Lateralization to Real Opponents and Mirrors There was a small but nevertheless significant right-side bias for the number of lateral displays to a real opponent (left frequency median ¼ 139.5, right frequency median ¼ 145.5, z ¼ 2.92, P ¼ 0.004) but not to a mirror image (left frequency median ¼ 57.0, right frequency median ¼ 76.0, z ¼ 1.53, P ¼ 0.126). There was also a longer total duration displaying the right side as opposed to the left to a real opponent (left duration median ¼ 339.7 s, right duration median ¼ 442.9 s, z ¼ 2.86, P ¼ 0.004) and also when looking at a mirror image (left duration median ¼ 224.6 s, right duration median ¼ 354.6 s, z ¼ 2.88, P ¼ 0.004). However, the median duration of displays to a real opponent did not differ between left and right (left duration median ¼ 2.56 s, right duration median ¼ 3.26 s, z ¼ 2.40, P ¼ 0.164), but was significantly longer with right than left displays to a mirror image (left duration median ¼ 3.11 s, right duration median ¼ 4.22 s, z ¼ 2.02, P ¼ 0.044). DISCUSSION Two contrasting hypotheses, each with clear predictions, were investigated in the current study. First, it was hypothesized that if fish repeatedly attempt to attain the head to tail posture then in the mirror treatment their responses will speed up resulting in more individual displays but a shorter duration of each lateral display. There was no support for this prediction. Second, it was suggested that the mirror might not elicit the same response as a real opponent since mirror images do not initiate moves to which the focal fish would respond. Thus, in the mirror treatment, the focal fish should not change position as frequently, resulting in fewer displays but longer individual displays to a mirror. There was firm support for this hypothesis. Real opponents elicited more frequent right and left lateral displays and frontal displays than did a mirror, consistent with the results of Balzarini et al. (2014) for Neolamprologus pulcher. Real opponents also elicited a greater total duration of left lateral displays
66
R. W. Elwood et al. / Animal Behaviour 97 (2014) 63e67
although not for right lateral or frontal displays. Further, the median duration of right lateral displays was greater towards a mirror than towards a real opponent. This last result is important because it indicates a slower change from one display type to another when a mirror image is seen and thus supports the idea that the fish react to behavioural change in real opponents and hence speed up transitions between displays. That the difference between conditions was found for the right lateral display but not for the left might indicate that the fish is particularly motivated to show the right side or use the right eye and that a mirror image does not induce the same behavioural changes as a real opponent. While the median duration of left lateral and frontal displays did not differ significantly between mirror and real opponent both were in the direction of longer displays to the mirror. Obviously, the mirror does not initiate behavioural change and thus fails to reflect reality when compared with a real opponent. We thus question the validity of using mirrors as a proxy for a live opponent and note that the slower pace of the mirror contest might explain findings of different brain gene expression (Desjardins & Fernald, 2010) and hormonal responses (Oliveira et al., 2005) when mirror and real opponents are compared. The observed change in contest dynamics, of fewer and longer displays, when facing a mirror image raises questions about how this would affect other aggressive activities. One could speculate that in this abnormal scenario individuals may abandon lateral displays for more escalated aggressive activities, such as biting. A study in lizards provides some support for this assertion with the finding that male Jacky dragons, Amphibolurus muricatus, facing an interactive video playback that mirrored their own aggressive behaviour had higher attack rates than in treatments that incorporated more appropriate video responses to the signal given by the test individuals (e.g. submission following an aggressive display; Van Dyk & Evans, 2008). In the current study, various display measures were positively correlated between mirror images and real opponents, including the number of left and right lateral displays, and the total durations of right and left lateral and frontal displays. Similarly, for N. pulcher, Balzarini et al. (2014) found correlations between aggressive behaviour directed towards a mirror and a real opponent. These findings are perhaps best viewed as highlighting the consistency of aggressiveness within individuals, an aspect of personality that is well recognized in fish (e.g. Wilson et al., 2011). Thus, in the context of personality research, the use of mirrors to assay aggressiveness is still a technique with merit, particularly because it avoids the need to face a real opponent, with the potential ethical concerns that such encounters may entail (e.g. Elwood, 1991). However, despite the correlations, there are still important differences in behaviour when comparing real and mirror image interactions, as revealed in the current study and that of Balzarini et al. (2014). In addition to the differences in frequencies and durations noted above there was no significant correlation between the two conditions for the number of frontal displays or the median duration of any type of display. Thus, we suggest that while the mirror test provides a somewhat crude measure of aggression, it fails to enable fish to interact in the appropriate way and is therefore not applicable as a method when researchers are interested in examining contest behaviour or contest resolution (Arnott & Elwood, 2009a; Earley et al., 2000; Elwood & Arnott, 2012). The present study confirms previous work (Arnott et al., 2011) demonstrating population-level lateralization, with a bias to present the right flank and/or use the right eye during aggressive lateral displays to an opponent beyond a transparent partition. We have demonstrated here that a right-side bias is also evident when displaying to a mirror, in terms of both total duration and median duration. With a real opponent this laterality of displays naturally results in the head to tail posture and thus is suggestive of
cooperation between opponents (Arnott et al., 2011). That is, laterality and cooperation were thought to be linked. However, our present results show that laterality is not driven by the focal fish seeing the opponent's right side and adopting the right-side posture in return, to achieve the head to tail configuration, because that configuration does not occur with the mirror. Clearly, the right-side bias is a feature of the individual and not of the interaction between opponents. Thus our previous suggestion (Arnott et al., 2011) that laterality is an important component of possible cooperation in information exchange must now be questioned because laterality still occurs with an uncooperative mirror image. Because lateralization of these agonistic interactions is so clear we suspect that showing the right side and using the right eye to view an opponent confers an advantage, particularly as a right-eye preference during aggressive encounters has been found by other studies in fish (e.g. Bisazza & de Santi, 2003; Reddon & Balshine, 2010). The use of visual information acquired via the right eye has been implicated in decision making (Rogers, 2002; Rogers, Vallortigara, & Andrew, 2013), thus lending further support to the idea that lateralization may play an important role in contest behaviour. Additionally, while results from the current study suggest mutual cooperation in displays may not have been the selective force for such lateralization, it could still emerge as a consequence, and when facing a real opponent this positioning may be a key feature of the interaction. Indeed, lateralization has been associated with predictable social behaviour, with a bias in the direction in which individuals approach each other, an important aspect of social interactions (e.g. Baraud, Buytet, Bec, & Blois-Heulin, 2009). An alternative explanation for lateralized displays towards mirror images could be that selection has acted so strongly on lateralization as a mechanism to engage in cooperative displays and reduce contest costs that the strategy remains fixed regardless of opponent orientation (or cooperative tendency). Individuals would rarely, if ever, encounter under natural circumstances the type of stimulus that a mirror image provides. Studies in fallow deer have confirmed the importance of lateralization during contests, with Jennings (2012) observing a rightsided bias to terminate parallel walks, interpreted as evidence of hemispheric specialization for processing information about the opponent during the visual display phase. However, recently, Jennings (2014a) found no evidence that more lateralized individuals (in terms of the decision to terminate the parallel walk) benefitted from reduced fight costs during escalated physical fighting, and contrary to initial predictions there was a positive relationship between laterality and time spent in escalated fighting. This suggests that the population-level lateralization observed during lateral displays in fish and deer may serve to facilitate this early phase of the contest, rather than more escalated phases. Adding to the complex picture of lateralization and aggressive behaviour in fallow deer, Jennings (2014b) found that weakly lateralized individuals were more likely to conduct third-party interventions of ongoing fights. However, individuals that showed lateralization for right-eye use were less likely to be targeted by an intervening male, an important benefit because targeted males are highly likely to be defeated by the intervener (Jennings, Carlin, & Gammell, 2009). It was also the case that less lateralized individuals achieved the greatest number of copulations. These studies highlight that the fitness effects of lateralization are context dependent (Rogers et al., 2013). The majority of studies on lateralization and aggression have focused on escalated attacks, documenting a left-eyeeright-hemisphere bias (Vallortigara & Rogers, 2005). Our studies and those of Jennings (see above) are the first to explore the role of lateralization during ritualized display phases of the contest. In fish the use of mirrors to disrupt the normal orientation of contestants has clear
R. W. Elwood et al. / Animal Behaviour 97 (2014) 63e67
effects on the frequency and durations of specific displays and thus mirrors do not reflect reality. Acknowledgments We are grateful to Gillian Riddell for technical assistance. We also thank two anonymous referees for their excellent suggestions for improving the manuscript. References Arnott, G., & Elwood, R. W. (2008). Information gathering and decision making about resource value in animal contests. Animal Behaviour, 76, 529e542. Arnott, G., & Elwood, R. W. (2009a). Assessment of fighting ability in animal contests. Animal Behaviour, 77, 991e1004. Arnott, G., & Elwood, R. W. (2009b). Gender differences in aggressive behaviour in convict cichlids. Animal Behaviour, 78, 1221e1227. Arnott, G., & Elwood, R. W. (2009c). Probing aggressive motivation in a cichlid fish. Biology Letters, 5, 762e764. Arnott, G., Ashton, C., & Elwood, R. W. (2011). Lateralization of lateral displays in convict cichlids. Biology Letters, 7, 683e685. Balzarini, V., Taborsky, M., Wanner, S., Koch, F., & Frommen, J. G. (2014). Mirror, mirror on the wall: the predictive value of mirror tests for measuring aggression in fish. Behavioral Ecology and Sociobiology, 68, 871e878. Baraud, I., Buytet, B., Bec, P., & Blois-Heulin, C. (2009). Social laterality and ‘transversality’ in two species of mangabeys: influence of rank and implication for hemispheric specialization. Behavioural Brain Research, 198, 449e458. Bisazza, A., & de Santi, A. (2003). Lateralization of aggression in fish. Behavioural Brain Research, 141, 131e136. Clutton-Brock, T. H., & Albon, S. D. (1979). The roaring of red deer and the evolution of honest advertisement. Behaviour, 69, 145e170. Clutton-Brock, T. H., Albon, S. D., Gibson, R. M., & Guinness, F. E. (1979). The logical stag: adaptive aspects of fighting in red deer. Animal Behaviour, 27, 211e255. Desjardins, J. K., & Fernald, R. D. (2010). What do fish make of mirror images? Biology Letters, 6, 744e747. Dijkstra, P. D., Schaafsma, S. M., Hofmann, H. A., & Groothuis, T. G. G. (2012). ‘Winner effect’ without winning: unresolved social conflicts increase the probability of winning a subsequent contest in a cichlid fish. Physiology & Behavior, 105, 489e492. Earley, R. L. (2010). Social eavesdropping and the evolution of conditional cooperation and cheating strategies. Philosophical Transactions of the Royal Society of London, Series B, 365, 2675e2686. Earley, R. L., Hsu, Y., & Wolf, L. L. (2000). The use of standard aggression testing methods to predict combat behaviour and contest outcome in Rivulus marmoratus dyads (Teleostei: Cyprinodontidae). Ethology, 106, 743e761. Elwood, R. W. (1991). Ethical implications of studies on infanticide and maternal aggression in rodents. Animal Behaviour, 42, 841e849. Elwood, R. W., & Arnott, G. (2012). Understanding how animals fight with Lloyd Morgan's canon. Animal Behaviour, 84, 1095e1102. Elwood, R. W., & Arnott, G. (2013). Assessments in contests are frequently assumed to be complex when simple explanations will suffice. Animal Behaviour, 86, e8ee12. Enquist, M., & Leimar, O. (1983). Evolution of fighting behaviour: decision rules and assessment of relative strength. Journal of Theoretical Biology, 102, 387e410. Enquist, M., Leimar, O., Ljungberg, T., Mallner, Y., & Segerdahl, N. (1990). A test of the sequential assessment game: fighting in the cichlid fish Nannacara anomala. Animal Behaviour, 40, 1e14. Fawcett, T. W., & Mowles, S. L. (2013). Assessments of fighting ability need not be cognitively complex. Animal Behaviour, 86, e1ee17. Ghirlanda, S., Frasnelli, E., & Vallortigara, G. (2009). Intraspecific competition and coordination in the evolution of lateralization. Philosophical Transactions of the Royal Society of London, Series B, 364, 861e866.
67
Hirschenhauser, K., Wittek, M., Johnston, P., & Mostl, E. (2008). Social context rather than behavioral output or winning modulates post-conflict testosterone responses in Japanese quail (Coturnix japonica). Physiology and Behaviour, 95, 457e463. Hsu, Y. Y., Earley, R. L., & Wolf, L. L. (2006). Modulation of aggressive behaviour by fighting experience: mechanisms and contest outcomes. Biological Reviews, 81, 33e74. http://dx.doi.org/10.1017/S146479310500686X. Hurd, P. L. (1997). Cooperative signalling between opponents in fish fights. Animal Behaviour, 54, 1309e1315. Jennings, D. J. (2012). Right-sided bias in fallow deer terminating parallel walks: evidence for lateralization during a lateral display. Animal Behaviour, 83, 1427e1432. Jennings, D. J. (2014a). Information gathering during contests: the relationship between lateralisation and contestant behaviour during fallow deer fights. Behavioural Processes, 103, 278e282. Jennings, D. J. (2014b). Limited evidence that visual lateralization is associated with fitness in rutting male fallow deer. Animal Behaviour, 92, 85e91. Jennings, D. J., Carlin, C. M., & Gammell, M. P. (2009). A winner effect supports third party intervention behaviour during fallow deer, Dama dama, fights. Animal Behaviour, 77, 343e348. Jennings, D. J., Elwood, R. W., Carlin, C. M., Hayden, T. J., & Gammell, M. P. (2012). Vocal rate as an assessment process during fallow deer contests. Behavioural Processes, 91, 152e158. Jennings, D. J., & Gammell, M. P. (2013). Ungulate contest behaviour. In I. C. W. Hardy, & M. Briffa (Eds.), Animal contests (pp. 304e320). Cambridge, U.K.: Cambridge University Press. May, H. Y., & Mercier, A. J. (2007). Duration of socialization influences responses to a mirror: responses of dominant and subordinate crayfish diverge with time of pairing. Journal of Experimental Biology, 210, 4428e4436. Mesterton-Gibbons, M., & Heap, S. M. (2014). Variation between self- and mutual assessment in animal contests. American Naturalist, 183, 199e213. Oliveira, R. F., Carneiro, L. A., & Carneiro, A. V. M. (2005). No hormonal response in tied fights. Nature, 437, 207e208. Payne, R. J. H. (1998). Gradually escalating fights and displays: the cumulative assessment model. Animal Behaviour, 56, 651e662. Payne, R. J. H., & Pagel, M. (1996). Escalation and time costs in displays of endurance. Journal of Theoretical Biology, 183, 185e193. Payne, R. J. H., & Pagel, M. (1997). Why do animals repeat displays? Animal Behaviour, 54, 109e119. Reddon, A. R., & Balshine, S. (2010). Lateralization in response to social stimuli in a cooperatively breeding cichlid fish. Behavioural Processes, 85, 68e71. Rogers, L. J. (2002). Advantages and disadvantages of lateralization. In L. J. Rogers, & R. J. Andrew (Eds.), Comparative vertebrate lateralization (pp. 126e153). Cambridge, U.K.: Cambridge University Press. Rogers, L. J., Vallortigara, C., & Andrew, R. J. (2013). Divided brains: The biology and behaviour of brain asymmetries. Cambridge, U.K.: Cambridge University Press. Taylor, P. W., & Elwood, R. W. (2003). The mismeasure of animal contests. Animal Behaviour, 65, 1195e1202. Tinbergen, N. (1951). The study of instinct. New York: Clarendon/Oxford University Press. Vallortigara, G., & Rogers, L. J. (2005). Survival with an asymmetrical brain: advantages and disadvantages of cerebral lateralization. Behavioral and Brain Sciences, 28, 575e633. Van Dyk, D. A., & Evans, C. S. (2008). Opponent assessment in lizards: examining the effect of aggressive and submissive signals. Behavioural Ecology, 19, 895e901. Verbeek, P., Iwamoto, T., & Murakami, N. (2007). Differences in aggression between wild-type and domesticated fighting fish are context dependent. Animal Behaviour, 73, 75e83. West, S. A., Griffin, A. S., & Gardner, A. (2007). Social semantics: altruism, cooperation, mutualism, strong reciprocity and group selection. Journal of Evolutionary Biology, 20, 415e432. Wilson, A. J., de Boer, M., Arnott, G., & Grimmer, A. (2011). Integrating personality research and animal contest theory: aggressiveness in the green swordtail Xiphophorus helleri. PLoS One, 6, e28024.