Does the presence of an odd individual affect group choice?

Does the presence of an odd individual affect group choice? Abstract Group formation (shoaling) with conspecifics is common in fishes and provides several antipredator benefits, such as improved food and predator detection. However, coral reef fishes often form mixed-species shoals, which can generate costs for some group members. For example, individuals that stand out from a group are more likely to be targeted by predators according to the oddity effect. Consequently, the presence of an odd fish might reduce the risk of predation to other group members. Alternatively, an odd individual might attract predators and increase predation risk for the group as a whole. We examined three co-occurring species of coral reef fishes using 2-choice tests to investigate: 1) whether individuals chose to associate with conspecifics over heterospecifics (i.e. the oddity effect); and 2) whether individuals associate with or avoid shoals containing an odd individual under conditions of low- and high-predation pressure. One species actively avoided associating with heterospecifics, consistent with the oddity effect. In contrast, 2 species exhibited no preference for heterospecifics versus conspecifics, perhaps due to less pronounced phenotypic differences between species pairs resulting in a lower relative risk of being odd. None of the 3 species showed either active avoidance or preference for shoals containing an odd individual, irrespective of predation pressure. In instances where the oddity effect is apparent (one species in our study), lower individual predation risk from associating with an odd fish might be negated by greater predation risk to the group as a whole. INTRODUCTION Group formation is a widespread behavior in terrestrial and aquatic animals (Krause and Ruxton 2002). In fishes, for example, individuals of the same species often aggregate to form unstructured groups (shoals) or groups that move in a coordinated fashion (schools) (Pitcher and Parrish 1993). Such aggregations provide advantages to individual members, such as improved food detection and protection from predators (Landeau and Terborgh 1986; Pitcher and Parrish 1993). Antipredator benefits arise from phenomena such as the selfish herd (individuals reduce their predation risk by putting other conspecifics between themselves and predators; Hamilton 1971), the “dilution effect” (the probability of being attacked is reduced to 1/n in a group of n individuals; Foster and Treherne 1981; Wrona and Dixon 1991), the “confusion effect” (predators have difficulties focusing on a single prey in a moving group; Welty 1934), and the “increased vigilance” or “many eyes” hypotheses (groups composed of many individuals are better at detecting approaching predators; Ward et al. 2011). However, being part of a group also has shortcomings. One notable disadvantage occurs when an individual conspicuously differs from other group members and thus becomes a preferred target of predators. This is known as the “oddity effect” (Mueller 1971; Milinski 1977; Landeau and Terborgh 1986). Fish shoals are dynamic, with individuals constantly reassessing the costs and benefits of membership, resulting in decisions to join, stay, or leave the group (Pitcher and Parrish 1993). Given the common occurrence of the oddity effect (reviewed in Krause et al. 2000; Krause and Ruxton 2002; Rodgers et al. 2015), selection should favor individuals that carefully evaluate the risk of associating with a particular shoal. To avoid appearing odd, fishes tend to join homogeneous shoals by clustering with conspecifics that are similar in size (Theodorakis 1989; Ranta et al. 1992b; Rodgers et al. 2015) or coloration (Krause et al. 1996; McRobert and Bradner 1998; Bradner and McRobert 2001; Gómez-Laplaza 2005). Studies to date have focused largely on shoaling behavior among conspecifics whereas comparatively little work exists on assortment by species and species-related differential risks for members of mixed-species shoals (Ranta et al. 1994; Krause et al. 1996; Ward et al. 2002). This is surprising given that countless fish species live in sympatry and form mixed-species shoals, particularly in species-rich habitats such as coral reefs (e.g. Ehrlich and Ehrlich 1973; Wolf 1985; Overholtzer and Motta 2000). Evolutionary explanations for interspecific group formation include antipredatory vigilance benefits (Barnard et al. 1982), reduced competition for the rare species while maintaining the antipredator benefits of large group-size (Allan 1986), and reduced predator detection for cryptic species joining conspicuous ones (Tosh et al. 2007). However, as with single-species groups, the costs and benefits of group formation are likely to be asymmetric for members of mixed-species shoals (Hobson 1963; Tosh et al. 2007). In a mixed-species shoal, according to the oddity effect, the less-abundant species should suffer from increased predation risk assuming that species are equally conspicuous. Indeed, several studies have documented that fish preferentially associate with conspecifics over heterospecifics when given a choice (sensu Krause et al. 2000). By extension, it is possible that individuals of the common species benefit from reduced predation risk via the presence of an odd individual. Thus, an individual of the common species might benefit from joining a shoal with an odd-group member present. In contrast, groups containing an odd individual might attract predators and experience greater overall predation risk at the expense of both species (Landeau and Terborgh 1986). In this case, individuals should always prefer uniform over heterogeneous shoals. These opposing predictions have never been experimentally tested. Here, we examined whether 3 species of coral reef fishes that often shoal together in nature prefer to 1) associate with con versus heterospecifics, and 2) associate or avoid shoals containing an odd individual. MATERIALS AND METHODS Fish collection and husbandry Experiments were conducted in August-September 2015 at the Lizard Island Research Station (14°40’S; 145°28’E), Queensland, Australia. We studied 2 species of Pomacentridae (damselfishes): Neopomacentrus azysron and Chromis viridis; and 1 species of Caesionidae (fusilier): Caesio teres (Supplementary Figure S1). These species are abundant on the Great Barrier Reef and often form mixed-species shoals in the wild (Figure 1). Ninety fish per species were collected on reefs surrounding Lizard Island with a monofilament barrier net and hand-nets, and transported in 20-L buckets to the LIRS aquarium facility within 1 h of capture. Fishes were then housed by species in glass aquaria (38 × 94 × 38 cm, W × L × H; stocking density of 20 to 25 fish per aquarium or ~1.0 g L−1) with air stones and flow-through water pumped directly from the reef. Cut PVC pipes were provided as shelter. The water temperature was 23 °C (±1 °C; actual variation), oxygen saturation was maintained above 95% (monitored with a FireStingO2 Optical Oxygen Meter; PyroScience GmbH, Aachen, Germany), and the light cycle was c. 12:12 h light:dark. We fed fishes ad libitum with tropical fish flakes (ORCA, Taiwan) twice daily and syphoned uneaten food from the bottom of the tank daily. Eighty fish per species were used in the experiments: 60 as focal individuals and 20 to form shoals. A shoal contained 10 fishes, a size at which shoal members were found to be significantly less vulnerable to predation than either solitary individuals or individuals in smaller groups of 2 to 4 (Landeau and Terborgh 1986). Familiarity among individuals is known to influence assortative grouping (Warburton and Lees 1996; Killen et al. 2017); therefore, fish used to form stimulus shoals were housed separately from the focal fish (see the description of housing tanks above). Once tested, focal fish were used to form additional stimulus shoals, beyond the 2 previously mentioned, which were used in trials involving focal heterospecifics. Focal fish were similar in size to those comprising the shoals (1.0 cm range) to avoid any confounding effect of size (Krause et al. 1996). Fishes ranged between 3.0 and 5.0 cm fork length (FL); based on size, C. viridis and N. azysron were juveniles and sub-adults, and C. teres were juveniles. Figure 1 View largeDownload slide The three study species swimming in mixed-species groups in the wild: (a) Chromis viridis with Caesio teres, (b) Neopomacentrus azysron with C. viridis, and (c) N. azysron with C. teres. Figure 1 View largeDownload slide The three study species swimming in mixed-species groups in the wild: (a) Chromis viridis with Caesio teres, (b) Neopomacentrus azysron with C. viridis, and (c) N. azysron with C. teres. Fish collections were authorized by the Great Barrier Reef Marine Parks Authority (permit G14/36625.1 delivered to the LIRS). Lateralization test We used a detour test to assess behavioral lateralization (see Bisazza et al. 1997; Roche et al. 2013) and selected focal individuals with low lateralization scores (absolute lateralization index [LA] below 40) to reduce biases due to individual directional preferences. Fish with a LA greater than 40 (considered significantly lateralized based on a binomial test) were used to form shoals. Experimental setup The experimental setup resembled the apparatus used in previous tests of the oddity effect in fishes (e.g. Krause and Godin 1994; Ward and Krause 2001). Trials were conducted in a glass aquarium (38 × 88 × 38 cm, W × L × H) with 3 opaque sides. A smaller aquarium (26 × 38 × 20 cm, W × L × H) was placed adjacent to the transparent side: this tank was either empty or contained a common predator of our study species, the peacock grouper (Cephalopolis cyanostigma). The predatory stimulus was visual in nature; there was no exchange of water between the tanks containing the predator and the focal fish. Four C. cyanostigma (range = 25.5–30.0 cm total length) were used in the study; they were kept in individual, flow-through tanks (45 × 68 × 39 cm, W × L × H), fed frozen pilchard twice daily, and randomly assigned to sequences of 5 consecutive trials (there were never more than 5 trials per day involving a predator). Opposite the tank for the predator, inside the main aquarium, 2 stimulus shoals were placed in small aquaria (17.7 × 17.7 × 20.2 cm, W × L × H) built of transparent acrylic (Figure 2). These aquaria contained holes to allow water exchanges since chemical cues are known to influence group choice (Ward et al. 2002; Wisenden et al. 2003; Ward and Currie 2013). Figure 2 View largeDownload slide The experimental arena. A focal fish (in blue) was introduced in a removable prism at a distance of 30 cm from 2 smaller aquaria (in gray), each containing one shoal of ten fish separated by an opaque partition. The association zone (in red) corresponds to 3 times the body length of the focal fish. The focal fish could smell and see the 2 shoals. A smaller aquarium at the opposite end was either empty (control, treatments 1 and 2) or contained a predator (peacock grouper; treatment 3). There was no exchange of water between the tank containing the predator and the focal fish; hence, the focal fish could see but not smell the predator in treatment 3. Figure 2 View largeDownload slide The experimental arena. A focal fish (in blue) was introduced in a removable prism at a distance of 30 cm from 2 smaller aquaria (in gray), each containing one shoal of ten fish separated by an opaque partition. The association zone (in red) corresponds to 3 times the body length of the focal fish. The focal fish could smell and see the 2 shoals. A smaller aquarium at the opposite end was either empty (control, treatments 1 and 2) or contained a predator (peacock grouper; treatment 3). There was no exchange of water between the tank containing the predator and the focal fish; hence, the focal fish could see but not smell the predator in treatment 3. Fish comprising the shoals were introduced in the test arena 10 min before each trial. The acrylic tanks containing the stimulus shoals were designed such that they could easily be moved to and from the test arena to minimize stress to the fish from netting and air exposure. Five minutes prior to the trial, a focal fish was introduced in the center of the test zone inside a bottomless acrylic rectangular container with holes so water could circulate through the walls (Figure 2). Focal fish were gently netted in an adjacent tank and transferred to the test arena with minimal air exposure (<5 s). An opaque partition between the test arena and the adjacent aquarium prevented fish from seeing the aquarium when it was empty or contained a predator (Figure 2). This barrier was removed 3 min before the start of a trial. The trial began when the acrylic container was removed by the experimenter, allowing the focal fish to move freely in the arena. Experimental treatments Different focal fish were tested under 4 conditions (60 fish were used per species, resulting in a sample size of 15 independent fish per combination of condition and species). Focal fish were never used more than once. The control consisted of a choice between 2 shoals of conspecifics. Treatment 1 was a choice between a shoal of conspecifics and a shoal of heterospecifics. Treatment 2 was a choice between a shoal of conspecifics and a shoal of conspecifics containing an odd individual (i.e. a heterospecific). Treatment 3 was the same as treatment 2, but with a predator placed in the adjacent tank. The experiments were designed so each species was used once as focal individuals and once as the odd species. Species pairs were determined randomly: N. azysron was tested with C. viridis (odd species), C. viridis with Caesio teres (odd species), and Caesio teres with N. azysron (odd species). The experiment was conducted inside a closed room of the Sir John Proud Aquarium Facility at the LIRS. Each trial lasted 10 min during which the behavior of the focal fish was recorded with a GoPro Hero 3+ (GoPro, San Mateo USA) positioned above the aquarium. The stock lens of the camera was replaced with a commercially available non-distortive lens (4.14 mm f/3.0 86° HFOV 5MP GP41430; Peau Productions Inc., San Diego, USA). The experimenter remotely controlled the camera with an Ipad® (Apple Inc., California, USA) from behind a wet lab bench, at a distance greater than 3 m to avoid disturbing the fish during the experiments. We tested 10 focal fish per day. Trials for one species in the 4 conditions were completed in 6 days. The identity of the shoals and their position in the test arena were randomly assigned for each trial using a random number generator in Microsoft Excel. Every evening, 5 fish were randomly chosen and transferred between stimulus shoals such that different shoals were used every day. Different odd individuals were used every day. Behavioral experiments were approved by the Queensland Government’s Department of Agriculture and Fisheries Animal Ethics Committee (permit CA 2015-06-868). Mortality was less than 2% and fishes were released on the reef, near their site of capture, at the end of the study. Video analysis We recorded the movements of focal fishes using the behavior logging software ODLog (Macropod Software, Yarraville, VIC, Australia), which logs keystrokes representing events or states. Cropped videos were coded (by DGR), enabling blind analysis by a single observer (FQ) (Clark et al. 2016). A focal fish was considered to associate with a shoal when it was within 3 body-lengths in front of a shoal compartment (c. 14 cm for C. viridis and Caesio teres; c. 11 cm for N. azysron) (Ward et al. 2002; Gómez-Laplaza 2005; Gomez-Laplaza 2009). A fish was not considered in association with a shoal if any part of its body was on the line drawn between the zones in front of the 2 shoals (Figure 2). We recorded the total time that the focal fish spent associating with each shoal and the number of times it moved between shoals (Gomez-Laplaza 2009). Statistical analysis One C. teres and one C. viridis were excluded from treatment 3 because of video recording issues. We compared the absolute time spent by focal fish associating with either shoal in the different conditions using paired t-tests. We used a Wilcoxon signed-rank test when the difference between association times in each zone was non-normally distributed. We compared the percentage of time that fish spent outside the 3 BL association zone and the number of times they switched shoals in the different conditions using Kruskal–Wallis tests. We did not correct for multiple hypothesis testing but interpret our results conservatively (Moran 2003; Nakagawa 2004). Statistical analyses were done in R v3.2.2 (R Core Team 2016). RESULTS Individuals from all three species spent most of their time in the association zone and readily moved between shoals (Table 1, Supplementary Figure S2). Control fishes did not display any preference for a particular shoal of conspecifics, indicating no inherent bias in our experimental setup (all Ps > 0.35, Table 2). N. azysron preferred to shoal with conspecifics over heterospecifics (s = 0, P < 0.001; Figure 3a) but C. viridis and C. teres exhibited no preference for either group (Table 2, Figure 3b and c). None of the 3 species displayed active avoidance or preference for shoals containing an odd individual, regardless of whether a predator was present or not (Table 2, Figure 3). Table 1 Mean (± one standard deviation) association time with a shoal (any of the 2 shoals) and number of times focal fishes switched shoal for the 3 species examined Species Association time (%) No. of switches Neopomacentrus azysron 87.38 ± 11.26 27.93 ±19.61 Caesio teres 87.88 ± 8.74 13.24 ± 8.00 Chromis viridis 93.81 ± 7.55 14.66 ± 13.76 Species Association time (%) No. of switches Neopomacentrus azysron 87.38 ± 11.26 27.93 ±19.61 Caesio teres 87.88 ± 8.74 13.24 ± 8.00 Chromis viridis 93.81 ± 7.55 14.66 ± 13.76 View Large Table 1 Mean (± one standard deviation) association time with a shoal (any of the 2 shoals) and number of times focal fishes switched shoal for the 3 species examined Species Association time (%) No. of switches Neopomacentrus azysron 87.38 ± 11.26 27.93 ±19.61 Caesio teres 87.88 ± 8.74 13.24 ± 8.00 Chromis viridis 93.81 ± 7.55 14.66 ± 13.76 Species Association time (%) No. of switches Neopomacentrus azysron 87.38 ± 11.26 27.93 ±19.61 Caesio teres 87.88 ± 8.74 13.24 ± 8.00 Chromis viridis 93.81 ± 7.55 14.66 ± 13.76 View Large Table 2 Results of the paired t-tests (paired) and Wilcoxon signed-rank tests (wilcox) to compare the association time between 2 shoals Species Treatment test df t/v value P value Neopomacentrus azysron control paired 14 0.648 0.505 treatment 1 wilcox 14 0 <0.001 treatment 2 paired 14 1.653 0.121 treatment 3 paired 14 0.894 0.387 Caesio teres control paired 14 0.939 0.364 treatment 1 paired 14 0.552 0.590 treatment 2 paired 14 −0.462 0.651 treatment 3 paired 13 2.099 0.056 Chromis viridis control wilcox 14 68 0.670 treatment 1 paired 14 −0.941 0.362 treatment 2 paired 13 0.597 0.561 treatment 3 paired 14 −0.349 0.732 Species Treatment test df t/v value P value Neopomacentrus azysron control paired 14 0.648 0.505 treatment 1 wilcox 14 0 <0.001 treatment 2 paired 14 1.653 0.121 treatment 3 paired 14 0.894 0.387 Caesio teres control paired 14 0.939 0.364 treatment 1 paired 14 0.552 0.590 treatment 2 paired 14 −0.462 0.651 treatment 3 paired 13 2.099 0.056 Chromis viridis control wilcox 14 68 0.670 treatment 1 paired 14 −0.941 0.362 treatment 2 paired 13 0.597 0.561 treatment 3 paired 14 −0.349 0.732 Three species were tested in 4 conditions: the control was a choice between 2 shoals of conspecifics; treatment 1 was a choice between a shoal of conspecifics and a shoal of heterospecifics; treatment 2 was a choice between a shoal of conspecifics and a shoal of conspecifics containing an odd individual (i.e. a heterospecific); treatment 3 was the same as treatment 2 but with a predator placed in an adjacent tank. View Large Table 2 Results of the paired t-tests (paired) and Wilcoxon signed-rank tests (wilcox) to compare the association time between 2 shoals Species Treatment test df t/v value P value Neopomacentrus azysron control paired 14 0.648 0.505 treatment 1 wilcox 14 0 <0.001 treatment 2 paired 14 1.653 0.121 treatment 3 paired 14 0.894 0.387 Caesio teres control paired 14 0.939 0.364 treatment 1 paired 14 0.552 0.590 treatment 2 paired 14 −0.462 0.651 treatment 3 paired 13 2.099 0.056 Chromis viridis control wilcox 14 68 0.670 treatment 1 paired 14 −0.941 0.362 treatment 2 paired 13 0.597 0.561 treatment 3 paired 14 −0.349 0.732 Species Treatment test df t/v value P value Neopomacentrus azysron control paired 14 0.648 0.505 treatment 1 wilcox 14 0 <0.001 treatment 2 paired 14 1.653 0.121 treatment 3 paired 14 0.894 0.387 Caesio teres control paired 14 0.939 0.364 treatment 1 paired 14 0.552 0.590 treatment 2 paired 14 −0.462 0.651 treatment 3 paired 13 2.099 0.056 Chromis viridis control wilcox 14 68 0.670 treatment 1 paired 14 −0.941 0.362 treatment 2 paired 13 0.597 0.561 treatment 3 paired 14 −0.349 0.732 Three species were tested in 4 conditions: the control was a choice between 2 shoals of conspecifics; treatment 1 was a choice between a shoal of conspecifics and a shoal of heterospecifics; treatment 2 was a choice between a shoal of conspecifics and a shoal of conspecifics containing an odd individual (i.e. a heterospecific); treatment 3 was the same as treatment 2 but with a predator placed in an adjacent tank. View Large Figure 3 View largeDownload slide Box-plots showing the difference in the association time (Δtassociation; in seconds) of focal fishes with shoals in four conditions: two shoals of conspecifics (control); a shoal of conspecifics and a shoal of heterospecifics (treatment 1; T1); a shoal of conspecifics and a shoal of conspecifics containing a heterospecific (treatment 2; T2); a shoal of conspecifics and a shoal of conspecifics containing a heterospecific in the presence of a predator (treatment 3; T3). Differences are calculated as: Δtassociation = t(left shoal) − t(right shoal) (control) or Δtassociation = t(shoal containing heterospecific individual(s)) − t(shoal of conspecifics) (treatments 1, 2, and 3). A Δtassociation of zero indicates that fish associated equally with both shoals. Empty circles represent individual data points (i.e. focal individuals). Figure 3 View largeDownload slide Box-plots showing the difference in the association time (Δtassociation; in seconds) of focal fishes with shoals in four conditions: two shoals of conspecifics (control); a shoal of conspecifics and a shoal of heterospecifics (treatment 1; T1); a shoal of conspecifics and a shoal of conspecifics containing a heterospecific (treatment 2; T2); a shoal of conspecifics and a shoal of conspecifics containing a heterospecific in the presence of a predator (treatment 3; T3). Differences are calculated as: Δtassociation = t(left shoal) − t(right shoal) (control) or Δtassociation = t(shoal containing heterospecific individual(s)) − t(shoal of conspecifics) (treatments 1, 2, and 3). A Δtassociation of zero indicates that fish associated equally with both shoals. Empty circles represent individual data points (i.e. focal individuals). DISCUSSION The oddity effect in mixed-species shoals We found evidence for the oddity effect in one of the 3 pairs of coral reef fishes tested: N. azysron was the only species that preferred shoaling with conspecifics over heterospecifics (Figure 3a). Two reasons might explain the occurrence of the oddity effect in this species. First, darkish blue-gray N. azysron appear highly conspicuous in a group of pale blue-green C. viridis. Second, in the wild, N. azysron typically escape in the water column when threatened (Hoggett 2016), whereas C. viridis dart for refuge among coral branches (Myers and Lieske 2004), potentially leaving the former isolated and vulnerable. Similar differences in escape strategies among members of mixed-species shoals have previously been documented in field and semi-natural experiments (e.g. Wolf 1985; Allan and Pitcher 1986). The other 2 species tested (C. viridis and C. teres) showed no obvious preference for associating with hetero or conspecifics (Figure 3b and c). Although the oddity effect is predicted to operate strongly in small groups such as those tested in our study (Landeau and Terborgh 1986), its influence on assortative behavior also depends on the type and degree of asymmetry between group members (Krause and Ruxton 2002; Rodgers et al. 2015). For example, though large individuals preferentially associate with similar-sized fish, small-bodied fish sometimes exhibit no preference for groups of different-size individuals (Svensson et al. 2000; Rodgers et al. 2011) or even prefer larger bodied conspecifics (Jones et al. 2010). Such non-assortment might result because: 1) small individuals benefit from joining larger ones if the latter are more profitable to predators (assuming similar handling costs) and hence at greater risk of predation (Peuhkuri 1997; Rodgers et al. 2015); or 2) factors such as aggression and competition override the influence of predation on group choice (Krause and Godin 1994; Jones et al. 2010). Thus, the lack of an oddity effect in 2 of the species pairs tested might occur because these fishes often shoal together in nature and blend relatively well, such that the perceived risk of joining a group of heterospecifics is low. Importantly, we did not test all possible combinations of species pairs in our study and note that different species combinations could have led to different outcomes. Which species comprises the alternative shoal in a 2-choice test appears to have a strong influence on whether fish exhibit a preference for conspecifics or not. For example, threespine sticklebacks (Gasterosteus aculeatus) prefer conspecifics when presented with an alternative shoal of bitterlings (Rhodeus sericeus) but not with an alternative shoal of roach (Rutilus rutilus) (Keenleyside 1955). Similarly, Krause et al. (2000) noted evidence for the oddity effect when G. aculeatus are given the choice between conspecifics and sympatric blackspotted sticklebacks (Gasterosteus wheatlandi) (FitzGerald and Morrissette 1992) but not when given a choice between conspecifics and the European minnow (Phoxinus phoxinus) (Barber et al. 1998). In our study, C. viridis exhibited no preference when given a choice to join a group of conspecifics over a group of C. teres, which are relatively similar in colouration. Had we tested the opposite scenario and presented C. teres with a shoal of C. teres and a shoal of C. viridis, focal individuals might have exhibited choices consistent with the oddity effect. This is because, similar to N. azysron, C. teres occupies upper levels of the water column on the reef and would be left isolated when C. viridis retreats to shelter in response to a threat. Our experiments were not designed to test this hypothesis explicitly, but future studies should explore the interesting possibility that behavioral differences in predator evasion might mediate assortative grouping in addition to heterogeneity in size, coloration, and species identity (see Eshel 1978; Wolf 1985; Allan and Pitcher 1986). Group choice in the presence of an odd individual Early experiments on predator preference and hunting success indicate that predatory fishes selectively target odd-group members (Milinski 1977; Landeau and Terborgh 1986; Theodorakis 1989). Given the vulnerability of odd individuals to predation and the ability of many fishes to identify odd-group members (including themselves), fish might benefit from associating with vulnerable individuals to lower their own predation risk. For example, under high-predation risk, armored brook sticklebacks (Culaea inconstans) prefer to associate with more vulnerable, non-armored fathead minnows (Pimephales promelas) despite this choice resulting in higher competition for food (Mathis and Chivers 2003). Here, we predicted that fish would either 1) join a shoal with an odd individual to reduce their own predation risk, or 2) avoid a shoal with an odd individual because all group members might suffer from increased predation risk. There is evidence that the presence of odd individuals in a shoal increases not only their own vulnerability to predation, but also that of the common phenotype: Landeau and Terborgh (1986) observed that the presence of 1 or 2 odd individuals in a shoal increases predatory attacks on the group by almost 3-fold and the incidence of capture by almost 5-fold. Interestingly, none of the 3 species we tested was attracted or repelled by a shoal containing an odd individual, including the species that exhibited a significant preference for conspecifics, consistent with the oddity effect. This behavior was unaffected by an increase in predation pressure (Figure 3). We note that experimentally simulating increased predation risk sometimes alters fish behavior and group choice (e.g. Krause 1994; Krause and Godin 1994; Mathis and Chivers 2003) but not always (e.g. Pitcher et al. 1986; Theodorakis 1989; Ranta et al. 1992a; Peuhkuri 1997). As such, even when the oddity effect is apparent, our results suggest that any benefit of lower predation risk from associating with an odd fish might be negated by elevated predation risk on the group as a whole. CONCLUSION We tested the oddity effect in three co-occurring species of coral reef fishes, exposing focal individuals to con and heterospecific visual and olfactory stimuli simultaneously (see Ward et al. 2002). Our results are consistent with previous studies suggesting that the oddity effect is not a universal phenomenon in mixed-species shoals (sensu Krause et al. 2000), and that the choice of species pairs appears to be important in eliciting a preference for conspecifics. We hypothesize that behavioral differences in predator evasion might be an important but overlooked trait affecting assortative grouping, in addition to factors such as size, coloration, harassment, and competition. Finally, our results extend the findings of early predation experiments on the oddity effect (Landeau and Terborgh 1986) by suggesting that the potential for reduced individual predation risk from joining a group with an odd member might be offset by an elevated predation risk to the entire group. Since size differences appear to be more consistently associated with the oddity effect than species differences (Krause et al. 2000), future studies should test the generality of this hypothesis with intraspecific group-choice experiments based on size differences—for example, by using a stimulus shoal comprising a large odd individual in a group of small conspecifics. FUNDING This work was supported funded by a Swiss National Science Foundation grant to R.B. (FN 31003A-153067/1) and a Fonds de Recherche du Québec Nature et Technologies postdoctoral fellowship to D.G.R. We thank Patricia Backwell for the idea that led to this study, the staff of the Lizard Island Research Station for logistical support, and members of the Eco-Ethology Lab at the University of Neuchâtel for comments on the manuscript. Data accessibility: The data and analysis script for this study are archived in the repository figshare (Quattrini et al. 2018; https://doi.org/10.6084/m9.figshare.3079942.v1) following best practices (Roche et al. 2015) and were made available to the editors upon initial submission. REFERENCES Allan JR , Pitcher TJ . 1986 . Species segregation during predator evasion in cyprinid fish shoals . Freshwat Biol . 16 : 653 – 659 . Google Scholar CrossRef Search ADS Barber I , Downey L , Braithwaite V . 1998 . Parasitism, oddity and the mechanism of shoal choice . J Fish Biol . 53 : 1365 – 1368 . Google Scholar CrossRef Search ADS Barnard C , Thompson D , Stephens H . 1982 . Time budgets, feeding efficiency and flock dynamics in mixed species flocks of lapwings, golden plovers and gulls . Behaviour . 80 : 44 – 69 . Google Scholar CrossRef Search ADS Bisazza A , Pignatti R , Vallortigara G . 1997 . Detour tests reveal task- and stimulus-specific behavioral lateralization in mosquitofish (Gambusia holbrooki) . Behav Brain Res . 89 : 237 – 242 . Google Scholar CrossRef Search ADS PubMed Bradner J , McRobert S . 2001 . The effect of shoal size on patterns of body colour segregation in mollies . J Fish Biol . 59 : 960 – 967 . Google Scholar CrossRef Search ADS Clark TD , Binning SA , Raby GD , Speers-Roesch B , Sundin J , Jutfelt F , Roche DG . 2016 . Scientific misconduct: the elephant in the lab. A response to parker et al . Trends Ecol Evol . 31 : 899 – 900 . Google Scholar CrossRef Search ADS PubMed Ehrlich PR , Ehrlich AH . 1973 . Coevolution: heterotypic schooling in caribbean reef fishes . Am Nat . 107 : 157 – 160 . Google Scholar CrossRef Search ADS Eshel I . 1978 . On a prey-predator nonzero-sum game and the evolution of gregarious behavior of evasive prey . Am Nat . 112 : 787 – 795 . Google Scholar CrossRef Search ADS FitzGerald G , Morrissette J . 1992 . Kin recognition and choice of shoal mates by threespine sticklebacks . Ethol Ecol Evol . 4 : 273 – 283 . Google Scholar CrossRef Search ADS Foster WA , Treherne JE . 1981 . Evidence for the dilution effect in the selfish herd from fish predation on a marine insect . Nature . 293 : 466 – 467 . Google Scholar CrossRef Search ADS Gómez-Laplaza LM . 2005 . The influence of social status on shoaling preferences in the freshwater angelfish (Pterophyllum scalare) . Behaviour . 142 : 827 – 844 . Google Scholar CrossRef Search ADS Gómez-Laplaza LM . 2009 . Recent social environment affects colour-assortative shoaling in juvenile angelfish (Pterophyllum scalare) . Behav Processes . 82 : 39 – 44 . Google Scholar CrossRef Search ADS PubMed Hamilton WD . 1971 . Geometry for the selfish herd . J Theor Biol . 31 : 295 – 311 . Google Scholar CrossRef Search ADS PubMed Hobson ES . 1963 . Selective feeding by the gafftopsail pompano, Trachinotus rhodopus (Gill), in mixed schools of herring and anchovies in the Gulf of California . Copeia . 1963 : 595 – 596 . Google Scholar CrossRef Search ADS Hoggett A . 2016 . Lizard Island Field Guide-yellowtail demoiselle . Available from: http://lifgaustralianmuseumnetau/Grouphtml?groupId=xfVEM2Tn&hierarchyId=PVWrQCLG (accessed 10 April 2016). Jones KA , Croft DP , Ramnarine IW , Godin JGJ . 2010 . Size-assortative shoaling in the guppy (Poecilia reticulata): the role of active choice . Ethology . 116 : 147 – 154 . Google Scholar CrossRef Search ADS Keenleyside MH . 1955 . Some aspects of the schooling behaviour of fish . Behaviour . 8 : 183 – 247 . Google Scholar CrossRef Search ADS Killen SS , Marras S , Nadler L , Domenici P . 2017 . The role of physiological traits in assortment among and within fish shoals . Philos Trans R Soc B . 372 : 20160233 . Google Scholar CrossRef Search ADS Krause J . 1994 . The influence of food competition and predation risk on size-assortative shoaling in juvenile chub (Leuciscus cephalus) . Ethology . 96 : 105 – 116 . Google Scholar CrossRef Search ADS Krause J , Butlin RK , Peuhkuri N , Pritchard VL . 2000 . The social organization of fish shoals: a test of the predictive power of laboratory experiments for the field . Biol Rev Camb Philos Soc . 75 : 477 – 501 . Google Scholar CrossRef Search ADS PubMed Krause J , Godin JGJ . 1994 . Shoal choice in the banded killifish (Fundulus diaphanus, Teleostei, Cyprinodontidae): effects of predation risk, fish size, species composition and size of shoals . Ethology . 98 : 128 – 136 . Google Scholar CrossRef Search ADS Krause J , Godin JGJ , Brown D . 1996 . Size-assortativeness in multi-species fish shoals . J Fish Biol . 49 : 221 – 225 . Google Scholar CrossRef Search ADS Krause J , Ruxton GD . 2002 . Living in groups . Oxford (UK): Oxford University Press . Landeau L , Terborgh J . 1986 . Oddity and the ‘confusion effect’ in predation . Anim Behav . 34 : 1372 – 1380 . Google Scholar CrossRef Search ADS Mathis A , Chivers DP . 2003 . Overriding the oddity effect in mixed-species aggregations: group choice by armored and nonarmored prey . Behav Ecol . 14 : 334 – 339 . Google Scholar CrossRef Search ADS McRobert SP , Bradner J . 1998 . The influence of body coloration on shoaling preferences in fish . Anim Behav . 56 : 611 – 615 . Google Scholar CrossRef Search ADS PubMed Milinski M . 1977 . Experiments on the selection by predators against spatial oddity of their prey . Ethology . 43 : 311 – 325 . Moran MD . 2003 . Arguments for rejecting the sequential Bonferroni in ecological studies . Oikos . 100 : 403 – 405 . Google Scholar CrossRef Search ADS Mueller HC . 1971 . Oddity and specific searching image more important than conspicuousness in prey selection . Nature . 233 : 345 – 346 . Google Scholar CrossRef Search ADS PubMed Myers R , Lieske E . 2004 . Coral reef guide red sea . London (UK) : Collins Publishers . Nakagawa S . 2004 . A farewell to Bonferroni: the problems of low statistical power and publication bias . Behav Ecol . 15 : 1044 – 1045 . Google Scholar CrossRef Search ADS Overholtzer KL , Motta PJ . 2000 . Effects of mixed-species foraging groups on the feeding and aggression of juvenile parrotfishes . Environ Biol Fishes . 58 : 345 – 354 . Google Scholar CrossRef Search ADS Peuhkuri N . 1997 . Size-assortative shoaling in fish: the effect of oddity on foraging behaviour . Anim Behav . 54 : 271 – 278 . Google Scholar CrossRef Search ADS PubMed Pitcher TJ , Magurran AE , Allan JR . 1986 . Size-segregative behaviour in minnow shoals . J Fish Biol . 29 : 83 – 95 . Google Scholar CrossRef Search ADS Pitcher TJ , Parrish JK . 1993 . Functions of shoaling behaviour in teleosts . In: Pitcher TJ , editor. Behaviour of teleost fishes . 2nd ed . London (UK): Chapman & Hall . p. 363 – 439 . Quattrini F , Bshary R , Roche DG . 2018 . Does the presence of an odd individual affect group choice ? Figshare Fileset . doi: 10.6084/m9.figshare.3079942.v1 R Core Team . 2016 . R: a language and environment for statistical computing . Vienna (Austria) : R Foundation for Statistical Computing . Available from: https://www.R-project.org/ (accessed 23 February 2016). Ranta E , Juvonen SK , Peuhkuri N . 1992a . Further evidence for size-assortative schooling in sticklebacks . J Fish Biol . 41 : 627 – 630 . Google Scholar CrossRef Search ADS Ranta E , Lindström K , Peuhkuri N . 1992b . Size matters when three-spined sticklebacks go to school . Anim Behav . 43 : 160 – 162 . Google Scholar CrossRef Search ADS Ranta E , Peuhkuri N , Laurila A . 1994 . A theoretical exploration of antipredatory and foraging factors promoting phenotype-assorted fish schools . Ecoscience . 1 : 99 – 106 . Google Scholar CrossRef Search ADS Roche DG , Binning SA , Strong LE , Davies JN , Jennions MD . 2013 . Increased behavioural lateralization in parasitized coral reef fish . Behav Ecol Sociobiol . 67 : 1339 – 1344 . Google Scholar CrossRef Search ADS Roche DG , Kruuk LE , Lanfear R , Binning SA . 2015 . Public data archiving in ecology and evolution: how well are we doing ? PLoS Biol . 13 : e1002295 . Google Scholar CrossRef Search ADS PubMed Rodgers GM , Downing B , Morrell LJ . 2015 . Prey body size mediates the predation risk associated with being “odd” . Behav Ecol . 26 : 242 – 246 . Google Scholar CrossRef Search ADS Rodgers GM , Ward JR , Askwith B , Morrell LJ . 2011 . Balancing the dilution and oddity effects: decisions depend on body size . PLoS One . 6 : e14819 . Google Scholar CrossRef Search ADS PubMed Svensson PA , Barber I , Forsgren E . 2000 . Shoaling behaviour of the two-spotted goby . J Fish Biol . 56 : 1477 – 1487 . Google Scholar CrossRef Search ADS Theodorakis CW . 1989 . Size segregation and the effects of oddity on predation risk in minnow schools . Anim Behav . 38 : 496 – 502 . Google Scholar CrossRef Search ADS Tosh CR , Jackson AL , Ruxton GD . 2007 . Individuals from different-looking animal species may group together to confuse shared predators: simulations with artificial neural networks . Proc R Soc Lond B . 274 : 827 – 832 . Google Scholar CrossRef Search ADS Warburton K , Lees N . 1996 . Species discrimination in guppies: learned responses to visual cues . Anim Behav . 52 : 371 – 378 . Google Scholar CrossRef Search ADS Ward AJ , Axford S , Krause J . 2002 . Mixed-species shoaling in fish: the sensory mechanisms and costs of shoal choice . Behav Ecol Sociobiol . 52 : 182 – 187 . Google Scholar CrossRef Search ADS Ward AJW , Currie S . 2013 . Shoaling fish can size-assort by chemical cues alone . Behav Ecol Sociobiol . 67 : 667 – 673 . Google Scholar CrossRef Search ADS Ward AJ , Herbert-Read JE , Sumpter DJ , Krause J . 2011 . Fast and accurate decisions through collective vigilance in fish shoals . Proc Natl Acad Sci USA . 108 : 2312 – 2315 . Google Scholar CrossRef Search ADS PubMed Ward AJ , Krause J . 2001 . Body length assortative shoaling in the European minnow, Phoxinus phoxinus . Anim Behav . 62 : 617 – 621 . Google Scholar CrossRef Search ADS Welty JC . 1934 . Experiments in group behavior of fishes . Physiol Zool . 7 : 85 – 128 . Google Scholar CrossRef Search ADS Wisenden BD , Pollock MS , Tremaine RJ , Webb JM , Wismer ME , Chivers DP . 2003 . Synergistic interactions between chemical alarm cues and the presence of conspecific and heterospecific fish shoals . Behav Ecol Sociobiol . 54 : 485 – 490 . Google Scholar CrossRef Search ADS Wolf NG . 1985 . Odd fish abandon mixed-species groups when threatened . Behav Ecol Sociobiol . 17 : 47 – 52 . Google Scholar CrossRef Search ADS Wrona FJ , Dixon RJ . 1991 . Group size and predation risk: a field analysis of encounter and dilution effects . Am Nat . 137 : 186 – 201 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Behavioral Ecology Oxford University Press

Does the presence of an odd individual affect group choice?

Loading next page...
1
 
/lp/ou_press/does-the-presence-of-an-odd-individual-affect-group-choice-huyWqxDHvN
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1045-2249
eISSN
1465-7279
D.O.I.
10.1093/beheco/ary062
Publisher site
See Article on Publisher Site

Abstract

Abstract Group formation (shoaling) with conspecifics is common in fishes and provides several antipredator benefits, such as improved food and predator detection. However, coral reef fishes often form mixed-species shoals, which can generate costs for some group members. For example, individuals that stand out from a group are more likely to be targeted by predators according to the oddity effect. Consequently, the presence of an odd fish might reduce the risk of predation to other group members. Alternatively, an odd individual might attract predators and increase predation risk for the group as a whole. We examined three co-occurring species of coral reef fishes using 2-choice tests to investigate: 1) whether individuals chose to associate with conspecifics over heterospecifics (i.e. the oddity effect); and 2) whether individuals associate with or avoid shoals containing an odd individual under conditions of low- and high-predation pressure. One species actively avoided associating with heterospecifics, consistent with the oddity effect. In contrast, 2 species exhibited no preference for heterospecifics versus conspecifics, perhaps due to less pronounced phenotypic differences between species pairs resulting in a lower relative risk of being odd. None of the 3 species showed either active avoidance or preference for shoals containing an odd individual, irrespective of predation pressure. In instances where the oddity effect is apparent (one species in our study), lower individual predation risk from associating with an odd fish might be negated by greater predation risk to the group as a whole. INTRODUCTION Group formation is a widespread behavior in terrestrial and aquatic animals (Krause and Ruxton 2002). In fishes, for example, individuals of the same species often aggregate to form unstructured groups (shoals) or groups that move in a coordinated fashion (schools) (Pitcher and Parrish 1993). Such aggregations provide advantages to individual members, such as improved food detection and protection from predators (Landeau and Terborgh 1986; Pitcher and Parrish 1993). Antipredator benefits arise from phenomena such as the selfish herd (individuals reduce their predation risk by putting other conspecifics between themselves and predators; Hamilton 1971), the “dilution effect” (the probability of being attacked is reduced to 1/n in a group of n individuals; Foster and Treherne 1981; Wrona and Dixon 1991), the “confusion effect” (predators have difficulties focusing on a single prey in a moving group; Welty 1934), and the “increased vigilance” or “many eyes” hypotheses (groups composed of many individuals are better at detecting approaching predators; Ward et al. 2011). However, being part of a group also has shortcomings. One notable disadvantage occurs when an individual conspicuously differs from other group members and thus becomes a preferred target of predators. This is known as the “oddity effect” (Mueller 1971; Milinski 1977; Landeau and Terborgh 1986). Fish shoals are dynamic, with individuals constantly reassessing the costs and benefits of membership, resulting in decisions to join, stay, or leave the group (Pitcher and Parrish 1993). Given the common occurrence of the oddity effect (reviewed in Krause et al. 2000; Krause and Ruxton 2002; Rodgers et al. 2015), selection should favor individuals that carefully evaluate the risk of associating with a particular shoal. To avoid appearing odd, fishes tend to join homogeneous shoals by clustering with conspecifics that are similar in size (Theodorakis 1989; Ranta et al. 1992b; Rodgers et al. 2015) or coloration (Krause et al. 1996; McRobert and Bradner 1998; Bradner and McRobert 2001; Gómez-Laplaza 2005). Studies to date have focused largely on shoaling behavior among conspecifics whereas comparatively little work exists on assortment by species and species-related differential risks for members of mixed-species shoals (Ranta et al. 1994; Krause et al. 1996; Ward et al. 2002). This is surprising given that countless fish species live in sympatry and form mixed-species shoals, particularly in species-rich habitats such as coral reefs (e.g. Ehrlich and Ehrlich 1973; Wolf 1985; Overholtzer and Motta 2000). Evolutionary explanations for interspecific group formation include antipredatory vigilance benefits (Barnard et al. 1982), reduced competition for the rare species while maintaining the antipredator benefits of large group-size (Allan 1986), and reduced predator detection for cryptic species joining conspicuous ones (Tosh et al. 2007). However, as with single-species groups, the costs and benefits of group formation are likely to be asymmetric for members of mixed-species shoals (Hobson 1963; Tosh et al. 2007). In a mixed-species shoal, according to the oddity effect, the less-abundant species should suffer from increased predation risk assuming that species are equally conspicuous. Indeed, several studies have documented that fish preferentially associate with conspecifics over heterospecifics when given a choice (sensu Krause et al. 2000). By extension, it is possible that individuals of the common species benefit from reduced predation risk via the presence of an odd individual. Thus, an individual of the common species might benefit from joining a shoal with an odd-group member present. In contrast, groups containing an odd individual might attract predators and experience greater overall predation risk at the expense of both species (Landeau and Terborgh 1986). In this case, individuals should always prefer uniform over heterogeneous shoals. These opposing predictions have never been experimentally tested. Here, we examined whether 3 species of coral reef fishes that often shoal together in nature prefer to 1) associate with con versus heterospecifics, and 2) associate or avoid shoals containing an odd individual. MATERIALS AND METHODS Fish collection and husbandry Experiments were conducted in August-September 2015 at the Lizard Island Research Station (14°40’S; 145°28’E), Queensland, Australia. We studied 2 species of Pomacentridae (damselfishes): Neopomacentrus azysron and Chromis viridis; and 1 species of Caesionidae (fusilier): Caesio teres (Supplementary Figure S1). These species are abundant on the Great Barrier Reef and often form mixed-species shoals in the wild (Figure 1). Ninety fish per species were collected on reefs surrounding Lizard Island with a monofilament barrier net and hand-nets, and transported in 20-L buckets to the LIRS aquarium facility within 1 h of capture. Fishes were then housed by species in glass aquaria (38 × 94 × 38 cm, W × L × H; stocking density of 20 to 25 fish per aquarium or ~1.0 g L−1) with air stones and flow-through water pumped directly from the reef. Cut PVC pipes were provided as shelter. The water temperature was 23 °C (±1 °C; actual variation), oxygen saturation was maintained above 95% (monitored with a FireStingO2 Optical Oxygen Meter; PyroScience GmbH, Aachen, Germany), and the light cycle was c. 12:12 h light:dark. We fed fishes ad libitum with tropical fish flakes (ORCA, Taiwan) twice daily and syphoned uneaten food from the bottom of the tank daily. Eighty fish per species were used in the experiments: 60 as focal individuals and 20 to form shoals. A shoal contained 10 fishes, a size at which shoal members were found to be significantly less vulnerable to predation than either solitary individuals or individuals in smaller groups of 2 to 4 (Landeau and Terborgh 1986). Familiarity among individuals is known to influence assortative grouping (Warburton and Lees 1996; Killen et al. 2017); therefore, fish used to form stimulus shoals were housed separately from the focal fish (see the description of housing tanks above). Once tested, focal fish were used to form additional stimulus shoals, beyond the 2 previously mentioned, which were used in trials involving focal heterospecifics. Focal fish were similar in size to those comprising the shoals (1.0 cm range) to avoid any confounding effect of size (Krause et al. 1996). Fishes ranged between 3.0 and 5.0 cm fork length (FL); based on size, C. viridis and N. azysron were juveniles and sub-adults, and C. teres were juveniles. Figure 1 View largeDownload slide The three study species swimming in mixed-species groups in the wild: (a) Chromis viridis with Caesio teres, (b) Neopomacentrus azysron with C. viridis, and (c) N. azysron with C. teres. Figure 1 View largeDownload slide The three study species swimming in mixed-species groups in the wild: (a) Chromis viridis with Caesio teres, (b) Neopomacentrus azysron with C. viridis, and (c) N. azysron with C. teres. Fish collections were authorized by the Great Barrier Reef Marine Parks Authority (permit G14/36625.1 delivered to the LIRS). Lateralization test We used a detour test to assess behavioral lateralization (see Bisazza et al. 1997; Roche et al. 2013) and selected focal individuals with low lateralization scores (absolute lateralization index [LA] below 40) to reduce biases due to individual directional preferences. Fish with a LA greater than 40 (considered significantly lateralized based on a binomial test) were used to form shoals. Experimental setup The experimental setup resembled the apparatus used in previous tests of the oddity effect in fishes (e.g. Krause and Godin 1994; Ward and Krause 2001). Trials were conducted in a glass aquarium (38 × 88 × 38 cm, W × L × H) with 3 opaque sides. A smaller aquarium (26 × 38 × 20 cm, W × L × H) was placed adjacent to the transparent side: this tank was either empty or contained a common predator of our study species, the peacock grouper (Cephalopolis cyanostigma). The predatory stimulus was visual in nature; there was no exchange of water between the tanks containing the predator and the focal fish. Four C. cyanostigma (range = 25.5–30.0 cm total length) were used in the study; they were kept in individual, flow-through tanks (45 × 68 × 39 cm, W × L × H), fed frozen pilchard twice daily, and randomly assigned to sequences of 5 consecutive trials (there were never more than 5 trials per day involving a predator). Opposite the tank for the predator, inside the main aquarium, 2 stimulus shoals were placed in small aquaria (17.7 × 17.7 × 20.2 cm, W × L × H) built of transparent acrylic (Figure 2). These aquaria contained holes to allow water exchanges since chemical cues are known to influence group choice (Ward et al. 2002; Wisenden et al. 2003; Ward and Currie 2013). Figure 2 View largeDownload slide The experimental arena. A focal fish (in blue) was introduced in a removable prism at a distance of 30 cm from 2 smaller aquaria (in gray), each containing one shoal of ten fish separated by an opaque partition. The association zone (in red) corresponds to 3 times the body length of the focal fish. The focal fish could smell and see the 2 shoals. A smaller aquarium at the opposite end was either empty (control, treatments 1 and 2) or contained a predator (peacock grouper; treatment 3). There was no exchange of water between the tank containing the predator and the focal fish; hence, the focal fish could see but not smell the predator in treatment 3. Figure 2 View largeDownload slide The experimental arena. A focal fish (in blue) was introduced in a removable prism at a distance of 30 cm from 2 smaller aquaria (in gray), each containing one shoal of ten fish separated by an opaque partition. The association zone (in red) corresponds to 3 times the body length of the focal fish. The focal fish could smell and see the 2 shoals. A smaller aquarium at the opposite end was either empty (control, treatments 1 and 2) or contained a predator (peacock grouper; treatment 3). There was no exchange of water between the tank containing the predator and the focal fish; hence, the focal fish could see but not smell the predator in treatment 3. Fish comprising the shoals were introduced in the test arena 10 min before each trial. The acrylic tanks containing the stimulus shoals were designed such that they could easily be moved to and from the test arena to minimize stress to the fish from netting and air exposure. Five minutes prior to the trial, a focal fish was introduced in the center of the test zone inside a bottomless acrylic rectangular container with holes so water could circulate through the walls (Figure 2). Focal fish were gently netted in an adjacent tank and transferred to the test arena with minimal air exposure (<5 s). An opaque partition between the test arena and the adjacent aquarium prevented fish from seeing the aquarium when it was empty or contained a predator (Figure 2). This barrier was removed 3 min before the start of a trial. The trial began when the acrylic container was removed by the experimenter, allowing the focal fish to move freely in the arena. Experimental treatments Different focal fish were tested under 4 conditions (60 fish were used per species, resulting in a sample size of 15 independent fish per combination of condition and species). Focal fish were never used more than once. The control consisted of a choice between 2 shoals of conspecifics. Treatment 1 was a choice between a shoal of conspecifics and a shoal of heterospecifics. Treatment 2 was a choice between a shoal of conspecifics and a shoal of conspecifics containing an odd individual (i.e. a heterospecific). Treatment 3 was the same as treatment 2, but with a predator placed in the adjacent tank. The experiments were designed so each species was used once as focal individuals and once as the odd species. Species pairs were determined randomly: N. azysron was tested with C. viridis (odd species), C. viridis with Caesio teres (odd species), and Caesio teres with N. azysron (odd species). The experiment was conducted inside a closed room of the Sir John Proud Aquarium Facility at the LIRS. Each trial lasted 10 min during which the behavior of the focal fish was recorded with a GoPro Hero 3+ (GoPro, San Mateo USA) positioned above the aquarium. The stock lens of the camera was replaced with a commercially available non-distortive lens (4.14 mm f/3.0 86° HFOV 5MP GP41430; Peau Productions Inc., San Diego, USA). The experimenter remotely controlled the camera with an Ipad® (Apple Inc., California, USA) from behind a wet lab bench, at a distance greater than 3 m to avoid disturbing the fish during the experiments. We tested 10 focal fish per day. Trials for one species in the 4 conditions were completed in 6 days. The identity of the shoals and their position in the test arena were randomly assigned for each trial using a random number generator in Microsoft Excel. Every evening, 5 fish were randomly chosen and transferred between stimulus shoals such that different shoals were used every day. Different odd individuals were used every day. Behavioral experiments were approved by the Queensland Government’s Department of Agriculture and Fisheries Animal Ethics Committee (permit CA 2015-06-868). Mortality was less than 2% and fishes were released on the reef, near their site of capture, at the end of the study. Video analysis We recorded the movements of focal fishes using the behavior logging software ODLog (Macropod Software, Yarraville, VIC, Australia), which logs keystrokes representing events or states. Cropped videos were coded (by DGR), enabling blind analysis by a single observer (FQ) (Clark et al. 2016). A focal fish was considered to associate with a shoal when it was within 3 body-lengths in front of a shoal compartment (c. 14 cm for C. viridis and Caesio teres; c. 11 cm for N. azysron) (Ward et al. 2002; Gómez-Laplaza 2005; Gomez-Laplaza 2009). A fish was not considered in association with a shoal if any part of its body was on the line drawn between the zones in front of the 2 shoals (Figure 2). We recorded the total time that the focal fish spent associating with each shoal and the number of times it moved between shoals (Gomez-Laplaza 2009). Statistical analysis One C. teres and one C. viridis were excluded from treatment 3 because of video recording issues. We compared the absolute time spent by focal fish associating with either shoal in the different conditions using paired t-tests. We used a Wilcoxon signed-rank test when the difference between association times in each zone was non-normally distributed. We compared the percentage of time that fish spent outside the 3 BL association zone and the number of times they switched shoals in the different conditions using Kruskal–Wallis tests. We did not correct for multiple hypothesis testing but interpret our results conservatively (Moran 2003; Nakagawa 2004). Statistical analyses were done in R v3.2.2 (R Core Team 2016). RESULTS Individuals from all three species spent most of their time in the association zone and readily moved between shoals (Table 1, Supplementary Figure S2). Control fishes did not display any preference for a particular shoal of conspecifics, indicating no inherent bias in our experimental setup (all Ps > 0.35, Table 2). N. azysron preferred to shoal with conspecifics over heterospecifics (s = 0, P < 0.001; Figure 3a) but C. viridis and C. teres exhibited no preference for either group (Table 2, Figure 3b and c). None of the 3 species displayed active avoidance or preference for shoals containing an odd individual, regardless of whether a predator was present or not (Table 2, Figure 3). Table 1 Mean (± one standard deviation) association time with a shoal (any of the 2 shoals) and number of times focal fishes switched shoal for the 3 species examined Species Association time (%) No. of switches Neopomacentrus azysron 87.38 ± 11.26 27.93 ±19.61 Caesio teres 87.88 ± 8.74 13.24 ± 8.00 Chromis viridis 93.81 ± 7.55 14.66 ± 13.76 Species Association time (%) No. of switches Neopomacentrus azysron 87.38 ± 11.26 27.93 ±19.61 Caesio teres 87.88 ± 8.74 13.24 ± 8.00 Chromis viridis 93.81 ± 7.55 14.66 ± 13.76 View Large Table 1 Mean (± one standard deviation) association time with a shoal (any of the 2 shoals) and number of times focal fishes switched shoal for the 3 species examined Species Association time (%) No. of switches Neopomacentrus azysron 87.38 ± 11.26 27.93 ±19.61 Caesio teres 87.88 ± 8.74 13.24 ± 8.00 Chromis viridis 93.81 ± 7.55 14.66 ± 13.76 Species Association time (%) No. of switches Neopomacentrus azysron 87.38 ± 11.26 27.93 ±19.61 Caesio teres 87.88 ± 8.74 13.24 ± 8.00 Chromis viridis 93.81 ± 7.55 14.66 ± 13.76 View Large Table 2 Results of the paired t-tests (paired) and Wilcoxon signed-rank tests (wilcox) to compare the association time between 2 shoals Species Treatment test df t/v value P value Neopomacentrus azysron control paired 14 0.648 0.505 treatment 1 wilcox 14 0 <0.001 treatment 2 paired 14 1.653 0.121 treatment 3 paired 14 0.894 0.387 Caesio teres control paired 14 0.939 0.364 treatment 1 paired 14 0.552 0.590 treatment 2 paired 14 −0.462 0.651 treatment 3 paired 13 2.099 0.056 Chromis viridis control wilcox 14 68 0.670 treatment 1 paired 14 −0.941 0.362 treatment 2 paired 13 0.597 0.561 treatment 3 paired 14 −0.349 0.732 Species Treatment test df t/v value P value Neopomacentrus azysron control paired 14 0.648 0.505 treatment 1 wilcox 14 0 <0.001 treatment 2 paired 14 1.653 0.121 treatment 3 paired 14 0.894 0.387 Caesio teres control paired 14 0.939 0.364 treatment 1 paired 14 0.552 0.590 treatment 2 paired 14 −0.462 0.651 treatment 3 paired 13 2.099 0.056 Chromis viridis control wilcox 14 68 0.670 treatment 1 paired 14 −0.941 0.362 treatment 2 paired 13 0.597 0.561 treatment 3 paired 14 −0.349 0.732 Three species were tested in 4 conditions: the control was a choice between 2 shoals of conspecifics; treatment 1 was a choice between a shoal of conspecifics and a shoal of heterospecifics; treatment 2 was a choice between a shoal of conspecifics and a shoal of conspecifics containing an odd individual (i.e. a heterospecific); treatment 3 was the same as treatment 2 but with a predator placed in an adjacent tank. View Large Table 2 Results of the paired t-tests (paired) and Wilcoxon signed-rank tests (wilcox) to compare the association time between 2 shoals Species Treatment test df t/v value P value Neopomacentrus azysron control paired 14 0.648 0.505 treatment 1 wilcox 14 0 <0.001 treatment 2 paired 14 1.653 0.121 treatment 3 paired 14 0.894 0.387 Caesio teres control paired 14 0.939 0.364 treatment 1 paired 14 0.552 0.590 treatment 2 paired 14 −0.462 0.651 treatment 3 paired 13 2.099 0.056 Chromis viridis control wilcox 14 68 0.670 treatment 1 paired 14 −0.941 0.362 treatment 2 paired 13 0.597 0.561 treatment 3 paired 14 −0.349 0.732 Species Treatment test df t/v value P value Neopomacentrus azysron control paired 14 0.648 0.505 treatment 1 wilcox 14 0 <0.001 treatment 2 paired 14 1.653 0.121 treatment 3 paired 14 0.894 0.387 Caesio teres control paired 14 0.939 0.364 treatment 1 paired 14 0.552 0.590 treatment 2 paired 14 −0.462 0.651 treatment 3 paired 13 2.099 0.056 Chromis viridis control wilcox 14 68 0.670 treatment 1 paired 14 −0.941 0.362 treatment 2 paired 13 0.597 0.561 treatment 3 paired 14 −0.349 0.732 Three species were tested in 4 conditions: the control was a choice between 2 shoals of conspecifics; treatment 1 was a choice between a shoal of conspecifics and a shoal of heterospecifics; treatment 2 was a choice between a shoal of conspecifics and a shoal of conspecifics containing an odd individual (i.e. a heterospecific); treatment 3 was the same as treatment 2 but with a predator placed in an adjacent tank. View Large Figure 3 View largeDownload slide Box-plots showing the difference in the association time (Δtassociation; in seconds) of focal fishes with shoals in four conditions: two shoals of conspecifics (control); a shoal of conspecifics and a shoal of heterospecifics (treatment 1; T1); a shoal of conspecifics and a shoal of conspecifics containing a heterospecific (treatment 2; T2); a shoal of conspecifics and a shoal of conspecifics containing a heterospecific in the presence of a predator (treatment 3; T3). Differences are calculated as: Δtassociation = t(left shoal) − t(right shoal) (control) or Δtassociation = t(shoal containing heterospecific individual(s)) − t(shoal of conspecifics) (treatments 1, 2, and 3). A Δtassociation of zero indicates that fish associated equally with both shoals. Empty circles represent individual data points (i.e. focal individuals). Figure 3 View largeDownload slide Box-plots showing the difference in the association time (Δtassociation; in seconds) of focal fishes with shoals in four conditions: two shoals of conspecifics (control); a shoal of conspecifics and a shoal of heterospecifics (treatment 1; T1); a shoal of conspecifics and a shoal of conspecifics containing a heterospecific (treatment 2; T2); a shoal of conspecifics and a shoal of conspecifics containing a heterospecific in the presence of a predator (treatment 3; T3). Differences are calculated as: Δtassociation = t(left shoal) − t(right shoal) (control) or Δtassociation = t(shoal containing heterospecific individual(s)) − t(shoal of conspecifics) (treatments 1, 2, and 3). A Δtassociation of zero indicates that fish associated equally with both shoals. Empty circles represent individual data points (i.e. focal individuals). DISCUSSION The oddity effect in mixed-species shoals We found evidence for the oddity effect in one of the 3 pairs of coral reef fishes tested: N. azysron was the only species that preferred shoaling with conspecifics over heterospecifics (Figure 3a). Two reasons might explain the occurrence of the oddity effect in this species. First, darkish blue-gray N. azysron appear highly conspicuous in a group of pale blue-green C. viridis. Second, in the wild, N. azysron typically escape in the water column when threatened (Hoggett 2016), whereas C. viridis dart for refuge among coral branches (Myers and Lieske 2004), potentially leaving the former isolated and vulnerable. Similar differences in escape strategies among members of mixed-species shoals have previously been documented in field and semi-natural experiments (e.g. Wolf 1985; Allan and Pitcher 1986). The other 2 species tested (C. viridis and C. teres) showed no obvious preference for associating with hetero or conspecifics (Figure 3b and c). Although the oddity effect is predicted to operate strongly in small groups such as those tested in our study (Landeau and Terborgh 1986), its influence on assortative behavior also depends on the type and degree of asymmetry between group members (Krause and Ruxton 2002; Rodgers et al. 2015). For example, though large individuals preferentially associate with similar-sized fish, small-bodied fish sometimes exhibit no preference for groups of different-size individuals (Svensson et al. 2000; Rodgers et al. 2011) or even prefer larger bodied conspecifics (Jones et al. 2010). Such non-assortment might result because: 1) small individuals benefit from joining larger ones if the latter are more profitable to predators (assuming similar handling costs) and hence at greater risk of predation (Peuhkuri 1997; Rodgers et al. 2015); or 2) factors such as aggression and competition override the influence of predation on group choice (Krause and Godin 1994; Jones et al. 2010). Thus, the lack of an oddity effect in 2 of the species pairs tested might occur because these fishes often shoal together in nature and blend relatively well, such that the perceived risk of joining a group of heterospecifics is low. Importantly, we did not test all possible combinations of species pairs in our study and note that different species combinations could have led to different outcomes. Which species comprises the alternative shoal in a 2-choice test appears to have a strong influence on whether fish exhibit a preference for conspecifics or not. For example, threespine sticklebacks (Gasterosteus aculeatus) prefer conspecifics when presented with an alternative shoal of bitterlings (Rhodeus sericeus) but not with an alternative shoal of roach (Rutilus rutilus) (Keenleyside 1955). Similarly, Krause et al. (2000) noted evidence for the oddity effect when G. aculeatus are given the choice between conspecifics and sympatric blackspotted sticklebacks (Gasterosteus wheatlandi) (FitzGerald and Morrissette 1992) but not when given a choice between conspecifics and the European minnow (Phoxinus phoxinus) (Barber et al. 1998). In our study, C. viridis exhibited no preference when given a choice to join a group of conspecifics over a group of C. teres, which are relatively similar in colouration. Had we tested the opposite scenario and presented C. teres with a shoal of C. teres and a shoal of C. viridis, focal individuals might have exhibited choices consistent with the oddity effect. This is because, similar to N. azysron, C. teres occupies upper levels of the water column on the reef and would be left isolated when C. viridis retreats to shelter in response to a threat. Our experiments were not designed to test this hypothesis explicitly, but future studies should explore the interesting possibility that behavioral differences in predator evasion might mediate assortative grouping in addition to heterogeneity in size, coloration, and species identity (see Eshel 1978; Wolf 1985; Allan and Pitcher 1986). Group choice in the presence of an odd individual Early experiments on predator preference and hunting success indicate that predatory fishes selectively target odd-group members (Milinski 1977; Landeau and Terborgh 1986; Theodorakis 1989). Given the vulnerability of odd individuals to predation and the ability of many fishes to identify odd-group members (including themselves), fish might benefit from associating with vulnerable individuals to lower their own predation risk. For example, under high-predation risk, armored brook sticklebacks (Culaea inconstans) prefer to associate with more vulnerable, non-armored fathead minnows (Pimephales promelas) despite this choice resulting in higher competition for food (Mathis and Chivers 2003). Here, we predicted that fish would either 1) join a shoal with an odd individual to reduce their own predation risk, or 2) avoid a shoal with an odd individual because all group members might suffer from increased predation risk. There is evidence that the presence of odd individuals in a shoal increases not only their own vulnerability to predation, but also that of the common phenotype: Landeau and Terborgh (1986) observed that the presence of 1 or 2 odd individuals in a shoal increases predatory attacks on the group by almost 3-fold and the incidence of capture by almost 5-fold. Interestingly, none of the 3 species we tested was attracted or repelled by a shoal containing an odd individual, including the species that exhibited a significant preference for conspecifics, consistent with the oddity effect. This behavior was unaffected by an increase in predation pressure (Figure 3). We note that experimentally simulating increased predation risk sometimes alters fish behavior and group choice (e.g. Krause 1994; Krause and Godin 1994; Mathis and Chivers 2003) but not always (e.g. Pitcher et al. 1986; Theodorakis 1989; Ranta et al. 1992a; Peuhkuri 1997). As such, even when the oddity effect is apparent, our results suggest that any benefit of lower predation risk from associating with an odd fish might be negated by elevated predation risk on the group as a whole. CONCLUSION We tested the oddity effect in three co-occurring species of coral reef fishes, exposing focal individuals to con and heterospecific visual and olfactory stimuli simultaneously (see Ward et al. 2002). Our results are consistent with previous studies suggesting that the oddity effect is not a universal phenomenon in mixed-species shoals (sensu Krause et al. 2000), and that the choice of species pairs appears to be important in eliciting a preference for conspecifics. We hypothesize that behavioral differences in predator evasion might be an important but overlooked trait affecting assortative grouping, in addition to factors such as size, coloration, harassment, and competition. Finally, our results extend the findings of early predation experiments on the oddity effect (Landeau and Terborgh 1986) by suggesting that the potential for reduced individual predation risk from joining a group with an odd member might be offset by an elevated predation risk to the entire group. Since size differences appear to be more consistently associated with the oddity effect than species differences (Krause et al. 2000), future studies should test the generality of this hypothesis with intraspecific group-choice experiments based on size differences—for example, by using a stimulus shoal comprising a large odd individual in a group of small conspecifics. FUNDING This work was supported funded by a Swiss National Science Foundation grant to R.B. (FN 31003A-153067/1) and a Fonds de Recherche du Québec Nature et Technologies postdoctoral fellowship to D.G.R. We thank Patricia Backwell for the idea that led to this study, the staff of the Lizard Island Research Station for logistical support, and members of the Eco-Ethology Lab at the University of Neuchâtel for comments on the manuscript. Data accessibility: The data and analysis script for this study are archived in the repository figshare (Quattrini et al. 2018; https://doi.org/10.6084/m9.figshare.3079942.v1) following best practices (Roche et al. 2015) and were made available to the editors upon initial submission. REFERENCES Allan JR , Pitcher TJ . 1986 . Species segregation during predator evasion in cyprinid fish shoals . Freshwat Biol . 16 : 653 – 659 . Google Scholar CrossRef Search ADS Barber I , Downey L , Braithwaite V . 1998 . Parasitism, oddity and the mechanism of shoal choice . J Fish Biol . 53 : 1365 – 1368 . Google Scholar CrossRef Search ADS Barnard C , Thompson D , Stephens H . 1982 . Time budgets, feeding efficiency and flock dynamics in mixed species flocks of lapwings, golden plovers and gulls . Behaviour . 80 : 44 – 69 . Google Scholar CrossRef Search ADS Bisazza A , Pignatti R , Vallortigara G . 1997 . Detour tests reveal task- and stimulus-specific behavioral lateralization in mosquitofish (Gambusia holbrooki) . Behav Brain Res . 89 : 237 – 242 . Google Scholar CrossRef Search ADS PubMed Bradner J , McRobert S . 2001 . The effect of shoal size on patterns of body colour segregation in mollies . J Fish Biol . 59 : 960 – 967 . Google Scholar CrossRef Search ADS Clark TD , Binning SA , Raby GD , Speers-Roesch B , Sundin J , Jutfelt F , Roche DG . 2016 . Scientific misconduct: the elephant in the lab. A response to parker et al . Trends Ecol Evol . 31 : 899 – 900 . Google Scholar CrossRef Search ADS PubMed Ehrlich PR , Ehrlich AH . 1973 . Coevolution: heterotypic schooling in caribbean reef fishes . Am Nat . 107 : 157 – 160 . Google Scholar CrossRef Search ADS Eshel I . 1978 . On a prey-predator nonzero-sum game and the evolution of gregarious behavior of evasive prey . Am Nat . 112 : 787 – 795 . Google Scholar CrossRef Search ADS FitzGerald G , Morrissette J . 1992 . Kin recognition and choice of shoal mates by threespine sticklebacks . Ethol Ecol Evol . 4 : 273 – 283 . Google Scholar CrossRef Search ADS Foster WA , Treherne JE . 1981 . Evidence for the dilution effect in the selfish herd from fish predation on a marine insect . Nature . 293 : 466 – 467 . Google Scholar CrossRef Search ADS Gómez-Laplaza LM . 2005 . The influence of social status on shoaling preferences in the freshwater angelfish (Pterophyllum scalare) . Behaviour . 142 : 827 – 844 . Google Scholar CrossRef Search ADS Gómez-Laplaza LM . 2009 . Recent social environment affects colour-assortative shoaling in juvenile angelfish (Pterophyllum scalare) . Behav Processes . 82 : 39 – 44 . Google Scholar CrossRef Search ADS PubMed Hamilton WD . 1971 . Geometry for the selfish herd . J Theor Biol . 31 : 295 – 311 . Google Scholar CrossRef Search ADS PubMed Hobson ES . 1963 . Selective feeding by the gafftopsail pompano, Trachinotus rhodopus (Gill), in mixed schools of herring and anchovies in the Gulf of California . Copeia . 1963 : 595 – 596 . Google Scholar CrossRef Search ADS Hoggett A . 2016 . Lizard Island Field Guide-yellowtail demoiselle . Available from: http://lifgaustralianmuseumnetau/Grouphtml?groupId=xfVEM2Tn&hierarchyId=PVWrQCLG (accessed 10 April 2016). Jones KA , Croft DP , Ramnarine IW , Godin JGJ . 2010 . Size-assortative shoaling in the guppy (Poecilia reticulata): the role of active choice . Ethology . 116 : 147 – 154 . Google Scholar CrossRef Search ADS Keenleyside MH . 1955 . Some aspects of the schooling behaviour of fish . Behaviour . 8 : 183 – 247 . Google Scholar CrossRef Search ADS Killen SS , Marras S , Nadler L , Domenici P . 2017 . The role of physiological traits in assortment among and within fish shoals . Philos Trans R Soc B . 372 : 20160233 . Google Scholar CrossRef Search ADS Krause J . 1994 . The influence of food competition and predation risk on size-assortative shoaling in juvenile chub (Leuciscus cephalus) . Ethology . 96 : 105 – 116 . Google Scholar CrossRef Search ADS Krause J , Butlin RK , Peuhkuri N , Pritchard VL . 2000 . The social organization of fish shoals: a test of the predictive power of laboratory experiments for the field . Biol Rev Camb Philos Soc . 75 : 477 – 501 . Google Scholar CrossRef Search ADS PubMed Krause J , Godin JGJ . 1994 . Shoal choice in the banded killifish (Fundulus diaphanus, Teleostei, Cyprinodontidae): effects of predation risk, fish size, species composition and size of shoals . Ethology . 98 : 128 – 136 . Google Scholar CrossRef Search ADS Krause J , Godin JGJ , Brown D . 1996 . Size-assortativeness in multi-species fish shoals . J Fish Biol . 49 : 221 – 225 . Google Scholar CrossRef Search ADS Krause J , Ruxton GD . 2002 . Living in groups . Oxford (UK): Oxford University Press . Landeau L , Terborgh J . 1986 . Oddity and the ‘confusion effect’ in predation . Anim Behav . 34 : 1372 – 1380 . Google Scholar CrossRef Search ADS Mathis A , Chivers DP . 2003 . Overriding the oddity effect in mixed-species aggregations: group choice by armored and nonarmored prey . Behav Ecol . 14 : 334 – 339 . Google Scholar CrossRef Search ADS McRobert SP , Bradner J . 1998 . The influence of body coloration on shoaling preferences in fish . Anim Behav . 56 : 611 – 615 . Google Scholar CrossRef Search ADS PubMed Milinski M . 1977 . Experiments on the selection by predators against spatial oddity of their prey . Ethology . 43 : 311 – 325 . Moran MD . 2003 . Arguments for rejecting the sequential Bonferroni in ecological studies . Oikos . 100 : 403 – 405 . Google Scholar CrossRef Search ADS Mueller HC . 1971 . Oddity and specific searching image more important than conspicuousness in prey selection . Nature . 233 : 345 – 346 . Google Scholar CrossRef Search ADS PubMed Myers R , Lieske E . 2004 . Coral reef guide red sea . London (UK) : Collins Publishers . Nakagawa S . 2004 . A farewell to Bonferroni: the problems of low statistical power and publication bias . Behav Ecol . 15 : 1044 – 1045 . Google Scholar CrossRef Search ADS Overholtzer KL , Motta PJ . 2000 . Effects of mixed-species foraging groups on the feeding and aggression of juvenile parrotfishes . Environ Biol Fishes . 58 : 345 – 354 . Google Scholar CrossRef Search ADS Peuhkuri N . 1997 . Size-assortative shoaling in fish: the effect of oddity on foraging behaviour . Anim Behav . 54 : 271 – 278 . Google Scholar CrossRef Search ADS PubMed Pitcher TJ , Magurran AE , Allan JR . 1986 . Size-segregative behaviour in minnow shoals . J Fish Biol . 29 : 83 – 95 . Google Scholar CrossRef Search ADS Pitcher TJ , Parrish JK . 1993 . Functions of shoaling behaviour in teleosts . In: Pitcher TJ , editor. Behaviour of teleost fishes . 2nd ed . London (UK): Chapman & Hall . p. 363 – 439 . Quattrini F , Bshary R , Roche DG . 2018 . Does the presence of an odd individual affect group choice ? Figshare Fileset . doi: 10.6084/m9.figshare.3079942.v1 R Core Team . 2016 . R: a language and environment for statistical computing . Vienna (Austria) : R Foundation for Statistical Computing . Available from: https://www.R-project.org/ (accessed 23 February 2016). Ranta E , Juvonen SK , Peuhkuri N . 1992a . Further evidence for size-assortative schooling in sticklebacks . J Fish Biol . 41 : 627 – 630 . Google Scholar CrossRef Search ADS Ranta E , Lindström K , Peuhkuri N . 1992b . Size matters when three-spined sticklebacks go to school . Anim Behav . 43 : 160 – 162 . Google Scholar CrossRef Search ADS Ranta E , Peuhkuri N , Laurila A . 1994 . A theoretical exploration of antipredatory and foraging factors promoting phenotype-assorted fish schools . Ecoscience . 1 : 99 – 106 . Google Scholar CrossRef Search ADS Roche DG , Binning SA , Strong LE , Davies JN , Jennions MD . 2013 . Increased behavioural lateralization in parasitized coral reef fish . Behav Ecol Sociobiol . 67 : 1339 – 1344 . Google Scholar CrossRef Search ADS Roche DG , Kruuk LE , Lanfear R , Binning SA . 2015 . Public data archiving in ecology and evolution: how well are we doing ? PLoS Biol . 13 : e1002295 . Google Scholar CrossRef Search ADS PubMed Rodgers GM , Downing B , Morrell LJ . 2015 . Prey body size mediates the predation risk associated with being “odd” . Behav Ecol . 26 : 242 – 246 . Google Scholar CrossRef Search ADS Rodgers GM , Ward JR , Askwith B , Morrell LJ . 2011 . Balancing the dilution and oddity effects: decisions depend on body size . PLoS One . 6 : e14819 . Google Scholar CrossRef Search ADS PubMed Svensson PA , Barber I , Forsgren E . 2000 . Shoaling behaviour of the two-spotted goby . J Fish Biol . 56 : 1477 – 1487 . Google Scholar CrossRef Search ADS Theodorakis CW . 1989 . Size segregation and the effects of oddity on predation risk in minnow schools . Anim Behav . 38 : 496 – 502 . Google Scholar CrossRef Search ADS Tosh CR , Jackson AL , Ruxton GD . 2007 . Individuals from different-looking animal species may group together to confuse shared predators: simulations with artificial neural networks . Proc R Soc Lond B . 274 : 827 – 832 . Google Scholar CrossRef Search ADS Warburton K , Lees N . 1996 . Species discrimination in guppies: learned responses to visual cues . Anim Behav . 52 : 371 – 378 . Google Scholar CrossRef Search ADS Ward AJ , Axford S , Krause J . 2002 . Mixed-species shoaling in fish: the sensory mechanisms and costs of shoal choice . Behav Ecol Sociobiol . 52 : 182 – 187 . Google Scholar CrossRef Search ADS Ward AJW , Currie S . 2013 . Shoaling fish can size-assort by chemical cues alone . Behav Ecol Sociobiol . 67 : 667 – 673 . Google Scholar CrossRef Search ADS Ward AJ , Herbert-Read JE , Sumpter DJ , Krause J . 2011 . Fast and accurate decisions through collective vigilance in fish shoals . Proc Natl Acad Sci USA . 108 : 2312 – 2315 . Google Scholar CrossRef Search ADS PubMed Ward AJ , Krause J . 2001 . Body length assortative shoaling in the European minnow, Phoxinus phoxinus . Anim Behav . 62 : 617 – 621 . Google Scholar CrossRef Search ADS Welty JC . 1934 . Experiments in group behavior of fishes . Physiol Zool . 7 : 85 – 128 . Google Scholar CrossRef Search ADS Wisenden BD , Pollock MS , Tremaine RJ , Webb JM , Wismer ME , Chivers DP . 2003 . Synergistic interactions between chemical alarm cues and the presence of conspecific and heterospecific fish shoals . Behav Ecol Sociobiol . 54 : 485 – 490 . Google Scholar CrossRef Search ADS Wolf NG . 1985 . Odd fish abandon mixed-species groups when threatened . Behav Ecol Sociobiol . 17 : 47 – 52 . Google Scholar CrossRef Search ADS Wrona FJ , Dixon RJ . 1991 . Group size and predation risk: a field analysis of encounter and dilution effects . Am Nat . 137 : 186 – 201 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Behavioral EcologyOxford University Press

Published: May 3, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off