- Top of page
- Conflict of Interest
Animals seldom exist within a nutritionally homogeneous environment, and as a result of variable nutritional composition and prey accessibility, they can experience a range of dietary options (Rapport 1980). Optimal foraging theory predicts that an individual should prefer prey species of high nutritional value relative to the energy spent to locate, capture, and consume the prey (Charnov 1976; Pyke et al. 1977). Differences in the nutrient composition of prey can play a key role in determining species-specific preferences (Jensen et al. 2012). However, the nutritional value of a given prey species may vary in response to differences in the condition or reproductive status of individuals, making optimum prey choice difficult (Fitzgibbon 1990; Gende et al. 2001; Lane et al. 2011). The relative accessibility, or vulnerability, of different prey species may also be important (Harder 1983; Hoogland et al. 2006; Plath et al. 2011). The presence of antipredator defenses or morphological features that constrain feeding can increase the time required to locate, manipulate, and consume food (Werner and Hall 1974; Temeles et al. 2009), reducing their value. Relative nutritional value may also vary within an individual, with consumers selectively targeting specific parts that provide the greatest nutritional benefit (Andrew and Jones 1990; Gende et al. 2001; Pekar et al. 2010; Pitman and Durban 2012) or the least protected parts of a prey organism.
On coral reefs, many of the associated fishes are dependent on live corals, as food, for shelter, or during recruitment (Munday et al. 2008). Coral-feeding fishes are among the most specialized species found on coral reefs, selectively consuming corals from particular genera or species (Berumen et al. 2005; Pratchett 2007; Cole et al. 2008, 2010; Rotjan and Lewis 2008; Brooker et al. 2013). The underlying basis of this selectivity is not well understood but could relate to a variety of factors such as biochemical composition, morphology, or antipredator defenses. It has often been assumed that selectivity relates to variation in the nutritional value between corals (Pisapia et al. 2012), and recent studies have shown that consuming a preferred coral can have positive effects on corallivorous fishes, improving relative growth rates (Berumen and Pratchett 2008), body condition (Berumen et al. 2005; Brooker et al. 2013), and reproductive output (Brooker et al. 2013). However, the few studies that have attempted to relate the biochemical profiles of coral tissue, in particular the levels of energetic macronutrients, to corallivore preferences have failed to find strong correlations with fitness-related benefits (Tricas 1989; Keesing 1990; Rotjan and Lewis 2005; Pisapia et al. 2012). Furthermore, patterns of coral-feeding within different coral species have received little attention, and it is not known if coral-feeding fishes target specific parts of the coral colony, either due to nutritional variation of differences in prey accessibility.
Scleractinian corals are generally composed of colonies of individual polyps, all extending from an aragonite exoskeleton. The basic anatomy of a coral polyp is relatively simple, consisting of a gastrointestinal chamber enclosed by a tentacle-ringed mouth. Each polyp produces an individual exoskeletal cup, the corallite, that provides protection for the polyp (Klaus et al. 2007). Polyps are connected by gastrovascular canals that run through the thin layer of interpolyp tissue, the coenosarc. Exoskeletal structure and polyp morphology vary extensively both between- and within-coral taxa (Klaus et al. 2007; Todd 2008), and this variation could affect how efficiently coral tissue can be consumed. For example, by selectively foraging on the coral Pocillopora meandrina, a species with clustered polyps, the butterflyfish, Chaetodon multicinctus increased its calorific intake per bite relative to when foraging on other corals (Tricas 1989). If corallivores attempt to maximize their efficiency when foraging, then preferences for specific corals may therefore reflect their morphological traits. To date, studies of corallivory and corallivore foraging preferences have generally considered each coral species to be an independent prey type (Cole et al. 2008) and have not tested whether corallivores use these corals uniformly or are influenced by factors, such as biochemical or morphological variation, that may occur within a single coral (but see Rotjan and Lewis 2009). Investigating prey selection at this finer scale may help define the processes driving prey selection in corallivorous fishes.
The objective of this study was to investigate, for the first time, the relative roles of nutrition and polyp accessibility in determining within-colony feeding selectivity by the corallivorous filefish, Oxymonacanthus longirostris (Bloch & Schneider, 1801; Fig. 1). This filefish is an obligate corallivore that feeds almost exclusively on corals from the genus Acropora (Kokita and Nakazono 2001). On the southern Great Barrier Reef (GBR), it primarily feeds on Acropora nobilis (Dana, 1846), which is an abundant branching coral in that region (Veron 2000). However, it also exhibits a strong dietary preference for Acropora millepora (Ehrenberg, 1834) and other less abundant coral species (Brooker et al. 2013). Patterns of feeding within these coral species are unknown. Here, we specifically set out to (1) confirm that O. longirostris primarily feeds on coral polyps; (2) determine whether or not O. longirostris shows a preference for different parts of A. nobilis coral colonies in the field and whether this is related to polyp density or corallite structure; (3) compare feeding patterns to determine whether food accessibility determines foraging location and whether fish are able to modify feeding patterns in response to food accessibility; and finally, (4) experimentally test whether nonuniform patterns of feeding on preferred coral species (A. millepora and Acropora tenuis [Dana, 1846]) reflect structural differences between polyps that may affect foraging efficiency.
- Top of page
- Conflict of Interest
Our field studies and laboratory experiments demonstrate that O. longirostris does not feed uniformly from coral colonies, but is selecting feeding positions with greater polyp accessibility, rather than those that are more nutritious. In the field, O. longirostris fed nonuniformly on the branching coral, A. nobilis, a species that forms the bulk of its diet (Brooker et al. 2013). Fish mostly fed centrally on each coral branch, avoiding areas near the growing tips and bases where branches intercept. Foraging observations also confirmed that O. longirostris targets individual polyps. In a pairwise choice experiment, where two factors, Acropora species and the point of origin of fragments (top or bottom sections of branches), fish selected fragments with comparatively larger, or numerous polyps. Fish appear to modify their foraging to select the most efficient prey available. When food accessibility was standardized along a branch, fish feed mostly on the central sections of the branch, irrespective of the level of accessibility. However, when food accessibility was manipulated so that it varied along the branch, fish consistently fed on the section of branch with the shallowest corallites regardless of its location. Together, these results suggest that patterns of within-coral selectivity by O. longirostris may reflect active choices made to increase foraging efficiency.
The actual tissue consumed by presumed corallivores is often not known (Cole et al. 2008). Our aquarium observations show that O. longirostris is predominantly a coral polyp feeder, selectively targeting individual polyps while avoiding the coenosarc. It is likely that this selectivity reflects the relative benefit of coral polyps as a food resource. Each coral polyp consists of a fleshy body cavity extending to the basal plate of the corallite cup, enclosed by the mouth and a ring of tentacles. In contrast, the coenosarc is a relatively thin layer of tissue that covers the underlying skeleton between these polyps. Therefore, selectively targeting polyps should allow a greater volume of tissue to be removed per bite, offsetting any increase in search times. Other corallivores including many butterflyfishes are also assumed to preferentially consume coral polyps (Alwany et al. 2003; Cole et al. 2009). While many species, including O. longirostris, have jaw and mouth structures that appear adapted for removing polyps there is limited direct evidence for this, with this assumption often based on gut content analysis that may fail to distinguish between polyps and general tissue (e.g., Hiatt and Strasburg 1960; Sano et al. 1984; Harmelin-Vivien 1989). As O. longirostris targets polyps, variation in polyp morphology, defensive structures, or the biochemical composition of polyps that increases or decreases the amount of energy consumed could have a direct influence on prey preferences both within and between coral species. In addition, it is possible that fish target polyps that maximize the effectiveness of their specialized trophic morphology.
In the field, O. longirostris exhibited highly nonuniform patterns of foraging on the branching coral A. nobilis. Foraging theory predicts individuals should target prey that maximizes energetic return (Pyke et al. 1977). Variation in the tissue composition or surface structure within a coral colony may alter the relative prey value by increasing or reducing the efficiency with which it can be consumed or assimilated. However, no significant variation was found in the morphometric variables of the A. nobilis coral branches examined, suggesting that within-colony selectivity is not driven by structural differences of the corallites, at least for this coral species.
There is evidence that the biochemistry of coral tissue can vary within a colony due to metabolic processes. For instance, the concentration of lipid, which is often indicated as being particularly important for corallivorous fishes (Tricas 1989; Rotjan and Lewis 2009), can vary within individual Acropora branches. Fang et al. (1989) found that polyps near the growing tip of the branching species, Acropora formosa, had lower lipid concentrations than polyps further down the branch, suggesting a biochemical gradient occurs as metabolites are transported up toward the growth point. While this suggests foraging near the growing tip may be less nutritionally beneficial, fish also avoided foraging near the base of branches. Foraging near the base may be less efficient due to the morphological constraints of locating suitable polyps in narrow areas where branches intersect and may require the fish to orientate itself at a suboptimal angle when searching or foraging. Midbranch, fish would have the greatest range of unrestricted motion. It is possible that predation risk may also influence feeding position; feeding near branch tips may increase potential exposure to predators, while feeding near branch bases may restrict movement and escape potential.
Behavioral experiments using simulated A. nobilis branches indicated that O. longirostris can distinguish between potential prey based on small morphological differences and, when preys are nutritionally similar, will modify their foraging patterns to preferentially select prey that are the most accessible and will presumably require the minimum effort to acquire. When simulated coral branches had identically sized artificial corallites along the branch, making food accessibility equal between segments, fish consistently feed on the central segment. This replicates the foraging patterns observed on A. nobilis, a species that morphometric measurements indicated has similar polyp morphology from the base to the tip of the branch. This suggests that, when there is limited structural variation, O. longirostris may have an innate drive to feed centrally along the branches of arborecent corals. The underlying basis for this behavior is not known but it may relate to nutritional variation between polyps along a branch if this is consistent between branches (Fang et al. 1989) or morphological constraints that affect foraging efficiency. However, when artificial corallite extension was manipulated this central foraging pattern was overridden, with fish preferentially foraging on the segment with the shallowest artificial corallite size regardless of its location. This consistent modification of foraging selectivity implies that the shallow artificial corallites were the most attractive to fish, either due to food being closer to the surface of the corallite thereby reducing the effort needed to extract it or increasing the amount that could be removed per bite, or food being more visible and so reducing the effort needed to search between bites. This result indicates that foraging decision-making by O. longirostris is flexible with fish able to recognize and respond to small differences in prey characteristics and able to modify their foraging behavior when presented with a novel prey to maximize foraging efficiency.
In the pairwise choice experiment using live coral fragments of two preferred Acropora species (A. millepora and A. tenuis), fish varied their prey preferences depending on the combination of coral species and point of origin of fragments (top or bottom sections of branches) presented. The preference patterns observed appear to reflect the morphological differences between coral fragments, specifically those that relate to polyp size and density. For instance, no significant difference was found along A. millepora branches with regard to any of the morphological variables recorded, and no foraging preference was exhibited by O. longirostris. However, fish exhibited a preference for the lower parts of A. tenuis branches where thecal extension was significantly less, and polyps were therefore less protected. Fish also exhibited a general preference for the bottom sections of A. tenuis branches over either section of A. millepora. While corallite density was slightly higher on A. millepora, A tenuis was found to have larger corallites. This may increase the relative amount of tissue that can be removed per bite, increasing overall foraging efficiency (Tricas 1989). No preference was observed between coral species when fish were provided with larger sections of coral composed of several whole branches. As A. millepora is known to be a preferred prey for O. longirostris (Brooker et al. 2013), it is therefore possible that overall both species represent equally valuable prey for these fishes. However, A. millepora may still be preferentially selected in the wild as fish chose the upper sections of A. millepora over those of A. tenuis, and lower sections of A. tenuis branches would remain difficult to access within fully intact colonies. The relationship between variation in corallite structure and prey preferences suggests that small-scale morphological differences between and within corals can affect the foraging decisions of O. longirostris.
Foraging selectivity is exhibited in many corallivorous species (Cole et al. 2008) with the consumption of preferred coral having beneficial effects on a variety of fitness-related parameters (Berumen et al. 2005; Berumen and Pratchett 2008; Brooker et al. 2013). It is generally assumed that these preferences relate to the nutritional content of coral tissue (Berumen et al. 2011; Pisapia et al. 2012). Despite this, attempts to relate preferences for specific corals to the relative levels of the major energetic macronutrients, such as lipids, protein, and carbohydrates, have failed to find strong correlations (Tricas 1989; Keesing 1990). However, these studies have generally considered the biochemical profile of each sampled colony as a single replicate. When within-colony differences were assessed, namely the total reproductive effort of polyps, Rotjan and Lewis (2009) found parrotfish consumed areas of Montastraea colonies with high numbers of gametes, ostensibly due to their higher protein and lipid levels. If the nutritional value of coral tissue consistently varies within a colony, and corallivores only target specific parts, then relevant differences in nutritional quality between coral species may have failed to be recognized due to a sampling methodology that does not account for these within-colony foraging patterns. Future work should therefore consider the biochemical variation within corals when attempting to determine a nutritional basis for foraging preferences.
Our results provide strong support for the hypothesis that coral morphology can influence corallivore foraging preferences. Morphology has previously been indicated in the preferences of the butterflyfish, C. multicinctus, where fish exhibited a strong preference for the massive Porites lobata over the branching Porites compressa (Tricas 1989), implying that the relatively flat foraging surface of P. lobata was the key driver of the preference. Many corallivorous fishes, including O. longirostris, preferentially target morphologically similar Acropora corals, generally digitate species with short branches and a relatively open corallite structure (Cole et al. 2008; Brooker et al. 2013). These corals may allow fish to ingest a relatively large amount of tissue per bite while requiring limited reorientation between bites. It is therefore possible that for ecologically similar corallivores, such as many butterflyfishes, coral morphology may also play a key role in determining dietary preferences. While it is likely that a variety of interacting factors influence the foraging preferences of these species, further work that determines the relative importance of nutritional quality versus accessibility may help to decipher why corallivores prefer certain corals.
In conclusion, our study shows that this corallivorous fish is a highly selective polyp feeder, with within-colony feeding selectivity probably driven by a combination of both innate preferences and responses to small-sale differences in polyp morphology that may affect foraging efficiency. Acropora corals appear to be highly variable in their value as prey and this can affect condition and fitness of individuals (Berumen and Pratchett 2008; Brooker et al. 2013). As obligate corallivores must achieve a nutritional balance from within a relatively narrow range of potential prey, precise behavioral mechanisms that increase foraging efficiency may help these species to maximize their performance.