Abstract

Eyespots have long been thought to confer protection against predators, but empirical evidence demonstrating the effectiveness of these markings and their survival value in the wild is limited. Using a mark–recapture experiment, I examined the functional significance of the eyespot on the dorsal fin of a juvenile tropical fish to its survival on coral reefs. None of the juveniles recaptured 1 month after settlement showed evidence of bite marks on the posterior region of the bodies to suggest a deflective function of their eyespot. When I compared the survivors with recruits from the same settlement cohort, I detected no change in the frequency distribution of eyespot size, suggesting no selective pressure operating on this trait. I compared these survivors with conspecifics from the same cohort collected at settlement and then outgrown in the absence of predators under 3 food regimes and 2 levels of intraspecific competition. I found that the eyespots of wild juveniles were larger overall than those of conspecifics maintained in a predator-free environment. The results of this study indicate that larger eyespots per se do not confer a survival advantage in the wild, suggesting that eyespots of this species may not have the long-assumed antipredatory function but play a role in interactions with adult conspecifics. I suggest that juveniles maintain eyespots even when predators or adult conspecifics are absent because they can be afforded at very low costs and may still be beneficial to their bearer under specific ecological conditions.

It is widely recognized that predators have profound effects on species diversity by influencing abundance, composition, and size structure of prey communities (Menge and Sutherland 1976; Sih et al. 1985; Endler 1986; Almany and Webster 2004). Yet, it is still unclear which traits affect the fitness of natural populations by being the target of predators. Because a close encounter with a predator can rapidly become a dead end for its prey, any form of deceptive appearance allowing individuals to avoid predators by preventing detection or recognition in the first place or deflecting attacks to less vital parts of the body once detected is expected to be highly beneficial. Indeed, many animal groups possess cryptic color patterns or conspicuous deflective markings, which have long been assumed to confer effective protection from predators and increase prey survivorship (see Ruxton et al. 2004). If these color patterns and markings have evolved as antipredatory adaptations, then understanding their role and function in nature can substantially improve our appreciation of the selective force through which predators shape the ecology and evolution of many species across habitats (McCollum and Leimberger 1997; Lyytinen et al. 2004).

Protective color patterns and markings are extremely diverse in their form and function (reviewed by Stevens 2007). Whereas most of these are permanent and provide broad-spectrum protection against a wide range of predators, some are facultative and their expression is induced by variation in predation risks (e.g., amphibians, McCollum and Van Buskirk 1996; cephalopods, Hanlon et al. 1999). A growing body of evidence on a variety of morphological traits supports the theory that inducible defenses are generally favored in environments where predators are not permanently present and thus confer a fitness advantage in the presence of predators but are otherwise costly (Lively 1986; Clark and Harvell 1992). Surprisingly, little has been done to determine whether the variation in some color patterns and markings also represents an inducible defense.

Ocelli, often referred to as “eyespots” because of their eye-resembling shape and color pattern, are a widespread feature in animals, most commonly in butterflies, moths, and some fishes. These conspicuous color markings have been attributed with various antipredatory functions, such as deterring hunting predators to initiate an attack as well as startling attacking predators (i.e., intimidation hypothesis) or diverting their attacks toward less vital body parts which may be sacrificed while allowing the prey a chance of escape (i.e., deflective hypothesis; reviewed by Stevens 2005). In particular, such deflective effect has been ascribed to peripheral eyespots located at the posterior end of the body far from the animal's head (e.g., butterflies, Brakefield and Reitsma 1991; fishes, Neudecker 1989; frogs, Van Buskirk et al. 2004). Yet, in spite of decades of research interest in the role and function of eyespots, empirical evidence for the actual antipredator effectiveness of eyespots under natural conditions remains limited (e.g., Vallin et al. 2007), is almost exclusively restricted to some taxa (i.e., insects), and remains rather unconvincing for others, notably fishes (reviewed by Ruxton et al. 2004).

Eyespots have been observed to occur on the tails and dorsal fins of juveniles in a number of fish species (Randall et al. 1997). Similar to other fish species, eyespots of the coral reef fish Pomacentrus amboinensis are located posteriorly and are assumed to misdirect a predator's attack away from vital parts of the body such as the head, by mimicking the real eyes but being located in a less vital region of the body, such as the dorsal fin (misdirection hypothesis, Neudecker 1989). Eyespots in this species may also be expected to deflect a predator's attention away by confusing a predator about its actual distance from a potential prey (reaction distance hypothesis, Meadows 1993). This idea assumes that a larger eyespot relative to the real eye would induce a predator to initiate an attack from a greater distance than it normally would. If P. amboinensis juveniles were able to detect an attacking predator from further away and react to it quickly enough to avoid such a deadly encounter, larger eyespots would be expected to confer an overall higher rate of survival to their bearers. A larger eyespot relative to the real eye may also intimidate a predator into perceiving a P. amboinensis juvenile to be a larger fish than it actually is (i.e., intimidation hypothesis). Nonetheless, the assumption that eyespots in this species have an antipredator function and larger eyespots relative to the size of real eyes better serve as a decoy, thereby increasing juvenile survivorship in the wild, remains untested.

In the presence of predators, the presumed survival advantage that a juvenile would gain by having a larger eyespot relative to its eye should offset any energetic cost incurred to attain and maintain such a defensive trait (e.g., investment in defensive morphology vs. investment in growth). Changes in investment made to maintain defensive traits over time are generally predicted to be associated to different levels of available resources (e.g., Myers and Bazely 1991; Werner and Anholt 1993; Steiner and Pfeiffer 2007). In predatory environments where resources are scarce and competition for them is high, energetic trade-offs are expected to intensify because under these conditions it is harder to grow and defend at the same time and the limited resources available are predominantly allocated to maintenance (Clark and Harvell 1992). When resources are widely available, however, energetic trade-offs are expected to lessen as it is now easier to grow and defend at the same time as allocating energy to essential mechanisms such as maintenance. Whether different sized eyespots are associated to different levels of resource availability (e.g., food) and affected by the presence or absence of conspecific juveniles competing for that resource has never been explored in fishes.

In this study, I aimed to establish whether eyespots in P. amboinensis have an antipredator function by tracking individuals over the first month of life on natural reef habitats. I hypothesized that the benefits of a larger eyespot relative to the eye should manifest in the presence of predators, whereas the associated costs should be most apparent when predators are absent and competition for resources high. Therefore, I tested 1) whether individuals with larger eyespots relative to the size of eyes enjoyed greater chances of survival on the reef and 2) whether the size of eyespots relative to eyes is associated to different levels of food availability and intraspecific competition for food.

MATERIALS AND METHODS

Study site and organisms

The study was conducted in November 2003 on the fringing reef around Lizard Island (14°38′S, 145°28′E) on the northern Great Barrier Reef (Australia). Here, P. amboinensis is abundant and settles on reefs in high numbers during October to February after a pelagic larval life of 15–23 days (Kerrigan 1996). At settlement, this species rapidly (<6 h) metamorphoses into the juvenile form by loosing the tissue transparency typical of the larval pelagic life stage and attains the bright yellow body coloration and the conspicuous black dorsal eyespot distinctive of the juvenile benthic life stage (McCormick et al. 2002, see Figure 1). After settlement, P. amboinensis remains strongly site attached throughout its life (McCormick and Makey 1997; Booth 2002), providing an ideal model for tracking the life history of individual fish in the wild.

Figure 1

Dusky yellow Pomacentrus amboinensis juveniles are frequently observed co-occurring in shallow coral reef habitats with other closely related pomacentrids, such as the bright blue Pomacentrus nagasakiensis. Juveniles of both species are characterized by a black eyespot on the posterior dorsal fin.

Figure 1

Dusky yellow Pomacentrus amboinensis juveniles are frequently observed co-occurring in shallow coral reef habitats with other closely related pomacentrids, such as the bright blue Pomacentrus nagasakiensis. Juveniles of both species are characterized by a black eyespot on the posterior dorsal fin.

Sampling morphology at settlement

I used light traps (for trap design, see Doherty 1987) to collect about 1000 newly metamorphosed P. amboinensis as they approached the reefs surrounding Lizard Island. Three traps were moored over sandy bottom, approximately 100 m apart and 30–50 m from the reef edge. They were deployed 1 m below the surface prior to dusk and cleared of fish just after dawn the following morning. I randomly selected 60 individuals and photographed them parallel to the front glass of a small narrow holding tank (6 × 2 × 4 cm) and against a scale bar. From these images, I measured their standard length (SL, in millimeters), eye area (EYE, in square millimeters), and eyespot area (ESP, in square millimeters) using image analysis software (Optimas 6.5) and then estimated the size of eyespots relative to real eye (ESP:EYE ratio). On the same day, these 60 fish were assigned to 3 experimental feeding treatments.

Experimental feeding treatments

To investigate the effect of resource availability on defensive morphology, I reared the above 60 newly metamorphosed individuals in outdoor 1-l aquaria (n = 1 fish per aquaria) as described in Gagliano and McCormick (2007). Aquaria were held outdoors, ensuring that temperature (29 ± 0.3 °C), salinity (34 ± 0.1 ppt), and light regimes remained as similar as possible to the natural environment. I randomly assigned fish to 3 feeding regimes (ad lib, every second day and every third day) for 30 days. Fish in the ad lib treatment were fed 24- to 36-h-old Artemia sp. nauplii 3 times throughout the day to ensure that they always had food in their tank during daylight hours. Fish in the other 2 treatments were fed once only in the morning of every second or third day, respectively. During the 30-day experimental period, I inspected all aquaria daily and removed any algal growth. I terminated the experiment on day 30, when I sacrificed all fish by cold shock, weighed them (W, g), and rephotographed them against a scale bar. From these images, I measured SL (millimeters), EYE (in square millimeters), and ESP (in square millimeters) and estimated the new ESP:EYE ratio. From the weight–length relationship parameters, I calculated the Fulton's K index (K = weight/length3) as an estimate of fish body condition.

Experimental manipulation of feeding and social conditions

On the same day as the feeding experiment was initiated, I randomly selected another sample of 96 newly metamorphosed P. amboinensis collected by light traps and assigned them to a second experiment. To investigate the extent to which social conditions affect variation in eyespot morphology, while accounting for the possible confounding effect of resource availability, feeding and social conditions were manipulated together in an experiment with 4 treatments: group fed ad lib (G+) × 8, single fed ad lib (S+) × 24, group fed every third day (G−) × 8, and single fed every third day (S−) × 24. Fish in the group treatments were reared in 3-l aquaria (n = 3 fish per aquaria), whereas those in the single treatments were reared in 1-l aquaria (n = 1 fish per aquaria). As in the feeding experiment, fish in the fed ad lib treatments received 24- to 36-h-old Artemia sp. nauplii 3 times every day, whereas fish in the other treatments were fed once only every third day. In the morning of day 8, 15, 22, and 30, I randomly selected 2 aquaria from each group treatment and 6 from each single treatment and measured W, SL, EYE, and ESP as described above.

Sampling morphology and survival of wild juveniles on the reef

During the settlement pulse in November 2003, I searched a 500-m stretch of reef and tagged 900 newly settled individuals in situ as previously described in Gagliano and McCormick (2007). After 30 days postsettlement, I searched and recaptured all surviving tagged individuals from the reef (n = 34). All 34 fish were killed by cold shock and photographed against a scale bar and their SL, EYE, and ESP measured. To determine whether eyespots have a defensive function in the wild, I compared the size of eyespots relative to the size of eyes (EYE:ESP ratio) of a random sample of newly metamorphosed recruits collected at settlement by light traps (n = 60) with the ESP:EYE ratio of juvenile survivors of the November settlement pulse collected from the reef 30 days after tagging. Given previous evidence showing a strong correlation between light trap catches and settlement patterns of P. amboinensis (Milicich et al. 1992; Meekan et al. 1993), newly metamorphosed individuals obtained from traps were considered to be a representative sample of individuals settling on the reef and used to characterize the natural range in ESP:EYE ratio immediately prior to the onset of heavy predation on the reef (e.g., McCormick and Hoey 2004; Almany and Webster 2006).

Analyses

To quantify the precision of the morphological measures, I randomly selected the digital image of 15 individuals, calibrated each image 3 times, and measured SL, EYE, and ESP on each image 3 times at separate occasions. The overall errors associated with obtaining measurements from the digital images were very low (e.g., coefficient of variation [CV] values: SL, 2.12%; EYE, 2.02%; and ESP, 1.55%), indicating that the morphological data were collected and quantified reliably.

I checked all data for normality and homogeneity of variance before performing statistical analyses (Sokal and Rohlf 2001). I used t-tests for dependent samples to compare the mean size difference of real eyes and eyespots. I examined changes in the size of eyes and eyespots over a 1-month period using a repeated-measures analysis of variance (ANOVA), where the effect of ontogeny (i.e., measurements of the same individuals at the beginning and at the end of the experiment) was the within-subjects factor, the effect of feeding treatments was the between-groups factor, and their interaction with each other represented the observed differences in the size of a trait among feeding treatments over time. A Bonferroni correction for repeated measures was applied (α = 0.025). The amount of natural variability in ESP:EYE ratio among individuals was estimated using CVs (%). I used linear regression analyses to examine the relationship between ESP:EYE ratio and fish size. I then tested for the effects of different feeding regimes and social conditions on fish morphology with 1-way and factorial ANOVAs and used analyses of covariance (ANCOVAs) to account for the effect of fish size and physiological condition on the observed variability in ESP:EYE ratio among treatments.

The link between survival and defensive morphology in the wild was assessed using a nonparametric Kolmogorov–Smirnov 2-sample test (α level = 0.05) and comparing the frequency distribution of the ESP:EYE ratio of survivors collected from the reef with that of newly metamorphosed recruits collected using light traps (i.e., my sample of the cohort at settlement from which these survivors originated). All fish used in the present study were sacrificed as part of a large project and in accordance with protocols approved by the James Cook University Animal Ethics Committee.

RESULTS

Eyespots' size at settlement

At settlement, eyespots were significantly larger than real eyes (meanEYE = 0.26 mm2 and meanESP = 11.05 mm2; t-test for dependent samples, t59 = −20.89, P < 0.001). Among newly settled individuals, the natural variability in size of eyespots relative to real eye was as high as 43.31% and was independent of body size (R2 = 0.003, P = 0.67).

Experimental feeding conditions

At the start of the experiment, all feeding groups were similar with respect to all traits measured (1-way ANOVA, P > 0.05 for all traits). However, both real eyes and eyespots of these same individuals predictably increased in size with ontogeny (repeated-measures ANOVA, within-subject factor, P < 0.001 for both traits) and food levels over time (repeated-measures ANOVA, within-subjects × between-groups factor, EYE: F2,57 = 36.04, P < 0.001, ad lib > every second day > every third day; ESP: F2,57 = 14.42, P < 0.001, ad lib > [every second day = every third day]). The ratio between them underwent a mean proportional decrease equal to 13.12% over the 30-day period (repeated-measures ANOVA, within-subject factor, F2,57 = 41.02, P < 0.001). This indicates that real eyes increased in size at a faster rate than eyespots did, although the overall size of eyespots remained significantly larger than real eyes (t-test for dependent samples, t59 = −34.72, P < 0.001). Interestingly, I found that the size of eyespots relative to real eyes did not significantly change among treatments over a 1-month postsettlement period (repeated-measures ANOVA, within-subjects × between-groups factor, F2,57 = 0.32, P = 0.73; Figure 2), but the variability in the ESP:EYE ratio among juveniles dropped to 20.79%.

Figure 2

Size of eyespots relative to real eyes (ESP:EYE ratio) of Pomacentrus amboinensis measured before experimental feeding manipulations and remeasured in the same individuals after a 30-day experimental period. Individual fish were fed ad lib 3 times each day (black bars), fed once only in the morning of every second day (gray bars), or once only in the morning of every third day (white bars). Data are means ± 95% confidence intervals.

Figure 2

Size of eyespots relative to real eyes (ESP:EYE ratio) of Pomacentrus amboinensis measured before experimental feeding manipulations and remeasured in the same individuals after a 30-day experimental period. Individual fish were fed ad lib 3 times each day (black bars), fed once only in the morning of every second day (gray bars), or once only in the morning of every third day (white bars). Data are means ± 95% confidence intervals.

Although individuals from different treatments were significantly different in body size (1-way ANOVA, F2,56 = 47.37, P < 0.001) and physiological condition (1-way ANOVA, F2,56 = 29.26, P < 0.001) at the end of the feeding experiment, the variability in ESP:EYE ratio among the 3 experimental groups remained independent from body size (ANCOVA, F2,56 = 2.07, P = 0.14) or condition (ANCOVA, F2,56 = 0.67, P = 0.52).

Experimental manipulation of feeding and social conditions

Both real eyes and eyespots predictably increased in size over time (factorial ANOVA, EYE: F3,76 = 78.32, P < 0.001; ESP: F3,76 = 13.85, P < 0.001) and similarly across all treatments (factorial ANOVA, EYE: F9,76 = 1.84, P = 0.08; ESP: F9,76 = 1.73, P = 0.10). Although statistically nonsignificant, fish in the group fed every third day (G−treatment), where individuals were predicted to experience the highest degree of competition for food, had the smallest eyes (factorial ANOVA, F1,88 = 3.33, P = 0.07) and eyespots (factorial ANOVA, F1,88 = 3.86, P = 0.053) at the end of the experiment. Nevertheless, the size of the eyespots relative to the real eyes of young P. amboinensis did not change significantly (factorial ANOVA, F1,88 = 0.53, P = 0.47) regardless of feeding conditions (fed ad lib or fed every third day) and social group size (single fish or group of 3 fish). At all sampling times, eyespots remained significantly larger than real eyes (t-test for dependent samples, P < 0.001 for all sampling times). The mean variability among juveniles in the ESP:EYE ratio dropped to 17.32% after the first week and to 14.86% by the end of the fourth week.

Despite significant differences in body size (1-way ANOVA, F3,88 = 9.55, P < 0.001) and physiological condition (1-way ANOVA, F3,88 = 5.78, P < 0.05) among experimental conditions, the variability in ESP:EYE ratio among treatments remained independent from body size (ANCOVA, F3,87 = 0.38, P = 0.77) or condition (ANCOVA, F3,87 = 0.41, P = 0.99).

Eyespots' size and survival of wild juveniles on the reef

None of the wild juvenile P. amboinensis recaptured from reef habitats showed bite marks on the posterior region of their bodies. The mean size of eyespots relative to real eyes of the survivors recaptured after 1 month did not change proportionally since settlement (1-way ANOVA, F1,92 = 3.21, P = 0.08; Figure 3) and remained independent of body size (R2 = 0.01, P = 0.55). However, the natural variability in the ESP:EYE ratio among wild juveniles dropped to 24.40%. Interestingly, I detected no selective pressure acting on the ESP:EYE ratio in the wild (Kolmogorov–Smirnov test, P > 0.10), where 1-month-old juveniles collected from the reef and recruits collected at settlement exhibited similar frequency distributions (recruits ratio range: 1.50–5.41; juvenile ratio range:1.50–4.67; Figure 4).

Figure 3

Mean size of eyespots relative to real eyes (ESP:EYE ratio) of the survivors recaptured after 1 month on the reef (white bar) compared with that of wild recruits at settlement (gray bar) and that of same-age conspecifics that were either fed individually in the laboratory (stippled bar) or maintained in small groups (striped bar) in a predator-free environment. Error bars represent means ± 95% confidence intervals.

Figure 3

Mean size of eyespots relative to real eyes (ESP:EYE ratio) of the survivors recaptured after 1 month on the reef (white bar) compared with that of wild recruits at settlement (gray bar) and that of same-age conspecifics that were either fed individually in the laboratory (stippled bar) or maintained in small groups (striped bar) in a predator-free environment. Error bars represent means ± 95% confidence intervals.

Figure 4

Frequency distribution of EYE:ESP ratio of 1-month-old survivors collected from the reef compared with that of wild recruits collected at settlement (i.e., representing a sample of the cohort at settlement from which the survivors originated).

Figure 4

Frequency distribution of EYE:ESP ratio of 1-month-old survivors collected from the reef compared with that of wild recruits collected at settlement (i.e., representing a sample of the cohort at settlement from which the survivors originated).

When I compared the size of eyespots relative to real eyes of 1-month-old wild juveniles with same-age conspecifics maintained under experimental conditions in the absence of predators, significant differences in ESP:EYE ratio emerged (all feeding groups only [pooled]: 1-way ANOVA, F1,92 = 12.449, P < 0.01; fourth week group conditions [pooled]: 1-way ANOVA, F1,43 = 5.13, P < 0.05; Figure 3). No matter the experimental condition, the mean ESP:EYE ratio was greater overall in wild P. amboinensis experiencing predators on the reef (i.e., they had larger eyespots) than conspecifics kept in the laboratory in the absence of predators.

DISCUSSION

To my knowledge, this is the first experimental study to quantify the extent to which juvenile survivorship of a marine fish is influenced by the size of its eyespot in the wild. Findings from the mark–recapture study combined with laboratory manipulations, where the size of eyespots relative to real eyes of wild juvenile survivors was compared with that of same-age conspecifics maintained under different food and social experimental conditions, indicate that the occurrence of overall larger eyespots relative to real eyes in the wild does not necessarily confer a survival advantage within the predatory environment. I found no evidence that fish with larger eyespots relative to real eyes have a greater probability of surviving during their first month on the reef, when they are the most susceptible to predators (Hoey and McCormick 2004; Gagliano and McCormick 2007). In fact, I detected no selective pressure acting on this trait, and similar to recent findings on insect species (e.g., Lyytinen et al. 2003; Vallin et al. 2007), my results provide no support for the adaptive value of eyespots as an effective antipredator mechanism in this species.

So, if eyespots are not adaptive and may have no antipredatory function in this species, why has evolution not scrapped them yet? One possible explanation could be that these markings are presently decorative leftovers of an antipredator trait that did confer some survival advantage in the recent past but no longer does. This may be a plausible scenario when predators are highly specialized in their choice of prey, which they attack frequently and successfully (Ruxton et al. 2004). In species-rich systems such as coral reefs, predators have been shown to be opportunistic generalists and unable to differentiate among closely related species such as pomacentrids, based on their appearance and colored markings (Almany et al. 2007). In fact, the most common predators of juvenile P. amboinensis are also known to successfully prey on other pomacentrid species (e.g., Pomacentrus nagasakiensis and Pomacentrus chrysurus) commonly co-occurring with P. amboinensis in shallow reef habitats and bearing similar juvenile eyespots on their posterior dorsal fin. Yet, this in itself suggests that predators have become familiar with a prey “type” rather than species and they have habituated to recognize similar visual marking, which they encounter frequently in related species. If correct, this idea implies that predators have been selected to stop being deceived by the markings of juvenile prey and, in this respect, have won the evolutionary arms race between the 2 parties. Clearly, additional work is required to provide support for this interpretation. For example, this could be investigated by looking at historical patterns of eyespot development, function, and selection. A phylogenetic approach could be used to reveal the origin of the eyespots on the posterior dorsal fin of P. amboinensis and related species (i.e., whether eyespots are inherited from a common ancestor) and to establish when and under which conditions a change in the character state (eyespot loss or appearance) occurred. Phylogenetic reconstructions have been successfully used to explain the evolution of coloration patterns in other taxa (e.g., butterflies, Brunton 1998; birds, Badyaev and Hill 2003). Unfortunately, phylogenetic data are not available for P. amboinensis and remains woefully incomplete for the Pomacentrus genus as a whole (data are available for only 19 of 214 species; see Quenouille et al. 2004). Although this approach is currently an inaccessible option for these species, it seems a very promising avenue for future studies.

A more parsimonious possibility to explain the presence of eyespots on the posterior dorsal fin of P. amboinensis is that the selective advantage conferred by eyespots was not detected because it is not uniform among different predatory environments. For example, the defensive value associated with eyespots in insects may vary depending on the predator (Vallin et al. 2007) or increase when a prey encounters naive predators, which are more easily decoyed by eyespots than adults (Lyytinen et al. 2004). In P. amboinensis, the magnitude of predation pressure on newly settled individuals can differ markedly among locations separated by only hundreds of meters, depending on the relative number of different predator species (McCormick and Holmes 2006). Moreover, McCormick and Meekan (2007) recently showed that spatial patchiness in predation pressure is also influenced by the presence of nest-guarding territorial P. amboinensis males. By restricting the predator types that can access the prey resource (i.e., newly recruited P. amboinensis) within defended nest sites, breeders indirectly facilitate the survival of conspecific juveniles. If eyespots conferred any fitness advantage to this species, this may be manifested only under particular conditions and can be lessened or even negated by the characteristics of the community in which individuals settle in.

Of course, it is also possible that fin eyespots of P. amboinensis juveniles are of no actual consequence in a predatory context because their message is not directed to predators. Instead, like subadult plumage in birds, fin eyespots of these juvenile fish may be a form of status signaling (i.e., status-signaling hypothesis; Lyon and Montgomerie 1986) that relays an honest signal of social subordinance to adult territorial males and makes them more willing to accept juveniles within their nest sites. By being allowed to stay within defended nest sites, juvenile P. amboinensis not only experience reduced risks of predation (McCormick and Meekan 2007) but also enjoy increased access to resources found within the nest site without being harassed by adults of their own and/or other species.

A comparison of wild individuals who survived on the reef in the presence of predators to laboratory-derived juveniles who experienced predator-free environments indicates that the mean size of eyespots relative to real eyes was greater overall in wild P. amboinensis than conspecifics kept in any of the experimental conditions. Given that the black pigment of P. amboinensis eyespots is synthesized by melanized organelles contained in melanophores rather than absorbed with food, the possibility that the differences in eyespots' size observed between wild and laboratory-derived juveniles can be directly attributed to differences in the quality of food fish fed on seems unlikely. Interestingly, the observed difference between wild and laboratory-derived juveniles does not appear to result from an increased investment in eyespots in response to predator presence but a decrease when predators are absent. If we attribute an antipredatory value to these eyespots, the present results suggest that P. amboinensis juveniles may be able to weigh up predation risk and adjust the investment made to maintain eyespots. Under this scenario, the observed difference between individuals from wild and predator-free conditions would represent a measure of investment in defense (cf., Tollrian and Harvell 1992). Yet, eyespots may still respond to predation risks while having no actual defensive value (i.e., nonadaptive trait; McPeek 2004), and hence, the difference between individuals from wild and predator-free conditions observed in the present study should be considered a measure of the costs of defense (Steiner 2007). It is, however, also possible that these eyespots have no actual antipredatory function, and the occurrence of a larger eyespot on the dorsal fin of wild P. amboinensis juveniles may be induced by factors other than those accounted for by my experimental setup, such as the social interactions of young P. amboinensis with mature male conspecifics. The observed difference between wild and laboratory-derived juveniles may then be the outcome of an increased investment in eyespots in response to the presence of adult male conspecifics and a decrease in their absence (i.e., in the laboratory). Under this scenario, the observed eyespot plasticity suggests that juveniles of this species may be able to change the investment made to maintain eyespots in response to the social conditions experienced during development. Given the possible benefit of possessing a large eyespot to deflect predators by diverting their attention away from the head or to reduce aggression by mature males by clearly signaling a juvenile status, it still remains unclear what the costs associated to this trait may be.

When I examined the juveniles from the laboratory conditions, I found that their eyes, eyespots, and EYE:ESP ratio did not respond to resource availability (i.e., food) or levels of intraspecific competition for that resource. At the end of the experiment, juveniles that experienced the lowest food levels and the most intense competition for it did exhibit the smallest eyespots, but they also had the smallest eyes. Besides the fact that these results were found to be statistically nonsignificant, they seem more likely to be indicative of poor growth rather than a response of eyespots per se to different social/feeding conditions. Previous authors have suggested that inducible defenses may not always result in measurable costs and pointed out that selection may tend to minimize the cost associated to the maintenance of induced traits, making such cost increasingly harder to detect (Tollrian 1995; Padilla and Adolph 1996). Following this logic, the results presented here suggest that eyespots in P. amboinensis juveniles may be afforded at relatively low or no costs so that they can be easily maintained without affecting energy allocation made to other basic life functions even when their presence may not be needed (e.g., when predators or adult conspecifics are absent).

In conclusion, my results show that the deflective function of eyespots on the posterior dorsal fin of P. amboinensis and possibly cogeneric species is doubtful. These fin markings might be retained by young fish in response to predator presence in reef habitats; nevertheless, the results of this study support the idea that eyespots are unimportant for decoying attacking predators. Although a high proportion of bite marks on the body of surviving wild juveniles would not necessarily demonstrate an antipredatory effect of eyespots because other attributes, such as the ability of young fish to escape predators, may account for it (Edmunds 1974), it would have at least given some support to the idea that eyespots are important in a predation context. However, none of the 34 juvenile survivors recaptured (out of 900) after 1 month since settlement on the reef exhibited bite marks on the region of the posterior dorsal fin to suggest a deflective or, more generally, an antipredatory function of these eyespots. On the other hand, the present findings suggest that these markings may have a completely different function and serve as an honest signal of the sexual immaturity of juveniles to reduce aggression by mature males. Paradoxically, eyespots do fade away from the dorsal fin of P. amboinensis as the animal approaches maturation and disappear completely in the vast majority but not all adult individuals. This opens up the possibility that the function of this signal changes over time by becoming a deceptive symbol of sexual maturity and allowing smaller adult males to go unrecognized as competing fertile individuals. By preventing dominant males from distinguishing these deceiving adults in juveniles' clothing and sexually immature individuals, the presence of an eyespot on the posterior dorsal fin may allow smaller males that are less likely to compete for females and/or defend a nest site successfully, to avoid fights with dominant males and increase access to food resources and potential mates (e.g., acting as sneakers). Although ontogenetic changes in coloring and pattern have been observed in many other vertebrate taxa (e.g., Cooper 1998; Kynard et al. 2002; Hawlena et al. 2006), the phenomenon, especially as it relates to fish behavioral ecology, remains poorly understood. Because coral reefs contain such a rich assortment of species, where opportunities for predator–prey interactions are high and social systems are extremely diverse, fish have the potential to help us better understand the functional role of color markings and the evolutionary significance of color ontogenetic changes. Given the paucity of empirical data available on this vertebrate group, more studies addressing these issues are urgently needed.

FUNDING

Layne Beachly Foundation; Nancy Vernon Rankine Fund.

I thank G. Hill, V. Messmer, and S. Smith for their assistance in the field and M. Depczynski, J.A.Y. Moore, L. Van Herwerden, W. Cresswell, and 2 anonymous reviewers for their thoughtful comments on an earlier version of the manuscript. The study was conducted under appropriate permits from the Great Barrier Reef Marine Park Authority. This paper is an output of the AIMS@JCU joint venture.

References

Almany
GR
Peacock
LF
Syms
C
McCormick
MI
Jones
GP
Predators target rare prey in coral reef fish assemblages
Oecologia
 , 
2007
, vol. 
152
 (pg. 
751
-
761
)
Almany
GR
Webster
MS
Odd species out as predators reduce diversity of coral reef fishes
Ecology
 , 
2004
, vol. 
85
 (pg. 
2933
-
2937
)
Almany
GR
Webster
MS
The predation gauntlet: early post-settlement mortality in reef fishes
Coral Reefs
 , 
2006
, vol. 
25
 (pg. 
19
-
22
)
Badyaev
AV
Hill
GH
Avian sexual dichromatism in relation to phylogeny and ecology
Annu Rev Ecol Evol Syst
 , 
2003
, vol. 
34
 (pg. 
27
-
49
)
Booth
DJ
Distribution changes after settlement in six species of damselfish (Pomacentridae) in One Tree Island lagoon, Great Barrier Reef
Mar Ecol Prog Ser
 , 
2002
, vol. 
226
 (pg. 
157
-
164
)
Brakefield
M
Reitsma
N
Phenotypic plasticity, seasonal climate and the population biology of Bicyclus butterflies
Ecol Entomol
 , 
1991
, vol. 
16
 (pg. 
291
-
303
)
Brunton
CFA
The evolution of ultraviolet patterns in European Colias butterflies (Lepidoptera, Pieridae): a phylogeny using mitochondrial DNA
Heredity
 , 
1998
, vol. 
80
 (pg. 
611
-
616
)
Clark
CW
Harvell
CD
Inducible defenses and the allocation of resources: a minimal model
Am Nat
 , 
1992
, vol. 
139
 (pg. 
521
-
539
)
Cooper
WE
Conditions favoring anticipatory and reactive displays deflecting predatory attack
Behav Ecol
 , 
1998
, vol. 
9
 (pg. 
598
-
604
)
Doherty
PJ
Light traps: selective but useful devices for quantifying the distributions and abundances of larval fishes
Bull Mar Sci
 , 
1987
, vol. 
41
 (pg. 
423
-
431
)
Edmunds
M
Significance of beak marks on butterfly wings
Oikos
 , 
1974
, vol. 
25
 (pg. 
117
-
118
)
Endler
JA
Natural selection in the wild
1986
Princeton (NJ)
Princeton University Press
Gagliano
M
McCormick
MI
Compensating in the wild: is flexible growth the key to early juvenile survival?
Oikos
 , 
2007
, vol. 
116
 (pg. 
111
-
120
)
Hanlon
RT
Forsythe
JW
Joneschild
DE
Crypsis, conspicuousness, mimicry and polyphenism as antipredator defences of foraging octopuses on Indo-Pacific coral reefs, with a method of quantifying crypsis from video tapes
Biol J Linn Soc
 , 
1999
, vol. 
66
 (pg. 
1
-
22
)
Hawlena
D
Boochnik
R
Abramsky
Z
Bouskila
A
Blue tail and striped body: why do lizards change their infant costume when growing up?
Behav Ecol
 , 
2006
, vol. 
17
 (pg. 
889
-
896
)
Hoey
AS
McCormick
MI
Selective predation for low body condition at the larval-juvenile transition of a coral reef fish
Oecologia
 , 
2004
, vol. 
139
 (pg. 
23
-
29
)
Kerrigan
BA
Temporal patterns in size and condition at settlement in two tropical reef fishes (Pomacentridae: Pomacentrus amboinensis and P. nagasakiensis)
Mar Ecol Prog Ser
 , 
1996
, vol. 
135
 (pg. 
27
-
41
)
Kynard
B
Henyey
E
Horgan
M
Ontogenetic behavior, migration, and social behavior of pallid sturgeon, Scaphirhynchus albus, and shovelnose sturgeon, S. platorynchus, with notes on the adaptive significance of body color
Environ Biol Fishes
 , 
2002
, vol. 
63
 (pg. 
389
-
403
)
Lively
CM
Canalization versus developmental conversion in a spatially variable environment
Am Nat
 , 
1986
, vol. 
128
 (pg. 
561
-
572
)
Lyon
BE
Montgomerie
RD
Delayed plumage maturation in passerine birds: reliable signaling by subordinate males?
Evolution
 , 
1986
, vol. 
40
 (pg. 
605
-
615
)
Lyytinen
A
Brakefield
PM
Lindström
L
Mappes
J
Does predation maintain eyespot plasticity in Bicyclus anynana?
Proc R Soc Lond B Biol Sci
 , 
2004
, vol. 
271
 (pg. 
279
-
283
)
Lyytinen
A
Brakefield
PM
Mappes
J
Significance of butterfly eyespots as an anti-predator device in ground-based and aerial attacks
Oikos
 , 
2003
, vol. 
100
 (pg. 
373
-
379
)
McCollum
SA
Leimberger
JD
Predator-induced morphological changes in an amphibian: predation by dragonflies affects tadpole shape and color
Oecologia
 , 
1997
, vol. 
109
 (pg. 
615
-
621
)
McCollum
SA
Van Buskirk
J
Costs and benefits of a predator-induced polyphenism in the gray treefrog Hyla chrysoscelis
Evolution
 , 
1996
, vol. 
50
 (pg. 
583
-
593
)
McCormick
MI
Hoey
AS
Larval growth history determines juvenile growth and survival in a tropical marine fish
Oikos
 , 
2004
, vol. 
106
 (pg. 
225
-
242
)
McCormick
MI
Holmes
TH
Prey experience of predation influences mortality rates at settlement in a coral reef fish, Pomacentrus amboinensis
J Fish Biol
 , 
2006
, vol. 
68
 (pg. 
969
-
974
)
McCormick
MI
Makey
L
Post-settlement transition in coral reef fishes: overlooked complexity in niche shifts
Mar Ecol Prog Ser
 , 
1997
, vol. 
153
 (pg. 
247
-
257
)
McCormick
MI
Makey
L
Dufour
V
Comparative study of metamorphosis in tropical reef fishes
Mar Biol
 , 
2002
, vol. 
141
 (pg. 
841
-
853
)
McCormick
MI
Meekan
MG
Social facilitation of selective mortality
Ecology
 , 
2007
, vol. 
88
 (pg. 
1562
-
1570
)
McPeek
MA
The growth-predation risk trade-off: so what is the mechanism?
Am Nat
 , 
2004
, vol. 
163
 (pg. 
E88
-
E111
)
Meadows
DW
Morphological variation in eyespots of the foureye butterflyfish (Chaetodon capistratus): implications for eyespot function
Copeia
 , 
1993
, vol. 
1993
 (pg. 
235
-
240
)
Meekan
MG
Milicich
MJ
Doherty
PJ
Larval production drives temporal patterns of larval supply and recruitment of a coral reef damselfish
Mar Ecol Prog Ser
 , 
1993
, vol. 
93
 (pg. 
217
-
225
)
Menge
BA
Sutherland
JP
Species diversity gradients: synthesis of the role of predation, competition, and temporal heterogeneity
Am Nat
 , 
1976
, vol. 
110
 (pg. 
351
-
369
)
Milicich
MJ
Meekan
MG
Doherty
PJ
Larval supply—a good predictor of recruitment of 3 species of reef fish (Pomacentridae)
Mar Ecol Prog Ser
 , 
1992
, vol. 
86
 (pg. 
153
-
166
)
Myers
JH
Bazely
D
Tallamy
DW
Raupp
MJ
Thorns, spines, prickles, and hairs: are they stimulated by herbivory and do they deter herbivores?
Phytochemical induction by herbivores
 , 
1991
New York
Wiley
(pg. 
p. 325
-
344
)
Neudecker
S
Eye camouflage and false eyespots: chaetodontid responses to predators
Environ Biol Fishes
 , 
1989
, vol. 
25
 (pg. 
143
-
157
)
Padilla
DK
Adolph
SC
Plastic inducible morphologies are not always adaptive: the importance of time delays in a stochastic environment
Evol Ecol
 , 
1996
, vol. 
10
 (pg. 
105
-
117
)
Quenouille
B
Bermingham
E
Planes
S
Molecular systematics of the damselfishes (Teleostei: Pomacentridae): Bayesian phylogenetic analyses of mitochondrial and nuclear DNA sequences
Mol Phylogenet Evol
 , 
2004
, vol. 
31
 (pg. 
66
-
88
)
Randall
JE
Allen
GR
Steene
RC
Fishes of the Great Barrier Reef and Coral Sea
1997
Bathurst (Australia)
Crawford House Publishing Pty Ltd.
Ruxton
GD
Sherratt
TN
Speed
MP
Avoiding attacks: the evolutionary ecology of crypsis, warning signals and mimicry
2004
Oxford
Oxford University Press.
Sih
A
Crowley
PH
McPeek
MA
Petranka
JW
Strohmeier
K
Predation, competition and prey communities: a review of field experiments
Annu Rev Ecol Syst
 , 
1985
, vol. 
16
 (pg. 
269
-
311
)
Sokal
RR
Rohlf
FJ
Biometry
2001
3rd ed.
New York
WH Freeman
Steiner
UK
Investment in defense and cost of predator-induced defense along a resource gradient
Oecologia
 , 
2007
, vol. 
152
 (pg. 
201
-
210
)
Steiner
UK
Pfeiffer
T
Optimizing time and resource allocation trade-offs for investment into morphological and behavioral defense
Am Nat.
 , 
2007
, vol. 
169
 (pg. 
118
-
129
)
Stevens
M
The role of eyespots as anti-predator mechanisms, principally demonstrated in Lepidoptera
Biol Rev
 , 
2005
, vol. 
80
 (pg. 
573
-
588
)
Stevens
M
Predator perception and the interrelation between different forms of protective coloration
Proc R Soc Lond B Biol Sci
 , 
2007
, vol. 
274
 (pg. 
1457
-
1464
)
Tollrian
R
Predator-induced morphological defenses: costs, life history shifts, and maternal effects in Daphnia pulex
Ecology
 , 
1995
, vol. 
76
 (pg. 
1691
-
1705
)
Tollrian
R
Harvell
CD
The ecology and evolution of inducible defenses
1992
Princeton (NJ)
Princeton University Press
Vallin
A
Jakobsson
S
Wiklund
C
“An eye for an eye?”—on the generality of the intimidating quality of eyespots in a butterfly and a hawkmoth
Behav Ecol Sociobiol
 , 
2007
, vol. 
61
 (pg. 
1419
-
1424
)
Van Buskirk
J
Aschwanden
J
Buckelmuller
I
Reolon
S
Ruttiman
S
Bold tail coloration protects tadpoles from dragonfly strikes
Copeia
 , 
2004
, vol. 
3
 (pg. 
599
-
602
)
Werner
EE
Anholt
BR
Ecological consequences of the trade-off between growth and mortality rates mediated by foraging activity
Am Nat
 , 
1993
, vol. 
142
 (pg. 
242
-
272
)