FISH
Fisheries Coral-Reef Fish Where unexploited by humans, coral-reef Rshes typ- 0004 ically exhibit high standing stocks, the maximum M. A. Hixon, Department of Zoology, Oregon State being about 240 t km\2 (about 24 t C km\2). High University, Corvallis, OR 97331-2914, USA standing crops reSect the high primary productivity 3 2 1 Copyright ^ 2001 Academic Press of coral reefs, often exceeding 10 gC m\ y\ , much of which is consumed directly or indirectly by Rshes. doi:10.1006/rwos.2001.0015 Correspondingly, reported Rshery yields have reach- ed 44 t km\2 y\1, with an estimated global potential Diversity, Distribution, and of 6M t y\1. These Rsheries provide food, bait, and Conservation live Rsh for the aquarium and restaurant trades. However, the estimated maximum sustainable yield 0001 Coral-reef Rshes comprise the most speciose assem- from shallow areas of actively growing coral reefs is blages of vertebrates on the Earth. The variety of 2 1 around 20}30 t km\ y\ , so many reefs are clearly shapes, sizes, colors, behavior, and ecology exhib- overexploited. Indeed, overRshing of coral reefs oc- ited by reef Rshes is amazing. Adult body sizes range curs worldwide, due primarily to unregulated multi- from gobies (Gobiidae) less than 1 cm in length to species and multigear exploitation in developing tiger sharks (Carcharhinidae) reportedly over 9 m nations. Few and inadequate stock assessments and long. It has been estimated that about 30% of the other quantitative Rshery analyses, susceptibility of some 15 000 described species of marine Rshes in- Rsh at spawning aggregations (see below), and de- habit coral reefs worldwide, and hundreds of species structive Rshing practices (including the use of dyna- can coexist on the same reef. Taxonomically, reef mite, cyanide, and bleach) are contributing factors. Rshes are dominated by about 30 families, mostly In the PaciRc, some 200}300 reef-Rsh species are the perciform chaetodontoids (butterSyRsh and taken by Rsheries, about 20 of which comprise some angelRsh families), labroids (damselRsh, wrasse, and 75% of the catch by weight. As Rshing intensiRes in parrotRsh families), gobioids (gobies), and acan- a given locality, large Rshes, especially piscivores thuroids (surgeonRshes). (see below), are typically depleted Rrst, followed by 0002 The latitudinal distribution of reef Rshes follows less preferred, smaller, and more productive plank- that of reef-building corals, which are usually lim- tivores and benthivores. (Fishing is naturally inhib- ited to shallow tropical waters bounded by the 203C ited in some regions by ciguatera Rsh poisoning, isotherms (roughly between the latitudes of 303N caused by dinoSagellate toxins concentrated in the and S). The longitudinal center of diversity is the tissues of some reef Rsh.) The indirect effects of Indo-Australasian archipelago of the Indo-PaciRc re- overRshing include the demise of piscivores perhaps gion. Local patterns of diversity are correlated with enhancing local populations of prey species, and the those of corals, which provide shelter and harbor demise of urchin-eating species (such as trigger- prey. There is a high degree of endemism in reef Rshes, Balistidae) and various herbivorous Rshes Rshes, especially on more isolated reefs, and many providing predatory and competitive release, respec- species (about 9%) have highly restricted geographi- tively, of sea urchins, which then overgraze and cal ranges. bioerode reefs. 0003 The major human activities that threaten reef Rshes include overRshing, especially by destructive R shing practices, and habitat destruction, including Morphology both local effects near human population centers and the ongoing global decline of reefs due to coral A typical perciform reef Rsh (virtually an oxymoron) 0005 bleaching. Worldwide, about 31% of coral-reef is laterally compressed, with a closed swimbladder Rshes are now considered critically endangered and and Rns positioned in a way that facilitates highly 24% threatened. The major solution for local con- maneuverable slow-speed swimming. Compared to servation is &no-take' marine protected areas, which more generalized relatives, reef Rsh have a greater have proven effective in replenishing depleted popu- proportion of musculature devoted to both locomo- lations. tion and feeding. Their jaws and pharyngeal appar- atus are complex and typically well developed for
RWOS 015 CHANAKSHI NP 2 FISH / Coral-Reef Fish
suction feeding of smaller invertebrate prey, with blennies (Blenniidae) mimic cleaner wrasse and tremendous variation reSecting a wide variety of thereby parasitize host Rsh. AnemoneRshes (Po- diets. For example, most butterSyRsh (Chaetodon- macentridae, especially the genus Amphiprion) live tidae) have forceps-like jaws that extract individual in a mutualistic association with several genera of polyps from corals, many damselRsh (Pomacen- large anemones. By circumventing discharge of the tridae, e.g., genus Chromis) have highly protrusible cnidarian’s nematocysts, the Rsh gain protection jaws that facilitate pipette-like suction feeding of from predators by hiding in the stinging tentacles of zooplankton, and parrotRshes (Scaridae) have fused the anemone. In turn, the Rsh defend their home beak-like jaw teeth and molar-like pharyngeal teeth from butterSyRshes and other predators that attack enabling some species to excavate algae from dead anemones. However, some host anemones survive reef surfaces. (This excavation and subsequent def- well without anemoneRsh, in which case the rela- ecation of coral sand can bioerode up to 9 kg of tionship is commensal rather than mutualistic. Fi- calcium carbonate per square meter annually.) Tet- nally, some gobies cohabit the burrows of digging raodontiform reef Rshes typically swim relatively shrimp. The shrimp provides shared shelter and the slowly with their dorsal and anal Rns, and conse- goby alerts the shrimp to the presence of predators. quently are morphologically well defended from predation by large dorsal}ventral spines (trigger- Territoriality Rshes, Balistidae), toxins (puffers, Tetraodontidae), The most overt form of competition involves terri- 0009 or quill-like scales (porucupineRshes, Diodontidae). toriality or defense of all or part of an individual’s The latter two families have fused dentition which is home range. Many reef Rshes behave aggressively well adapted for consuming hard-shelled invert- toward members of both their own and other spe- ebrates. cies, but the most obviously territorial species are 0006 Diurnal reef Rshes are primarily visual predators. benthic-feeding damselRshes (Pomacentridae, e.g., Visual acuity is high and retinal structure indicates genus Stegastes). By pugnaciously defending areas color vision. Coloration is highly variable, ranging about a meter square from herbivorous Rshes, dam- from cryptic to dazzling. Bright &poster' colors are selRsh prevent overgrazing and can thus maintain hypothesized to serve as visual signals in aggression, dense patches of seaweeds. These algal mats serve as courtship, and other social interactions. Sexually a food source for the damselRsh as well as habitat dimorphic coloration is associated with haremic so- for small juvenile Rsh of various species that manage cial systems (see below). Nocturnal reef Rshes are to avoid eviction. At a local spatial scale, the algal either visually oriented, having relatively large eyes mats can both smother corals as well as maintain (e.g., squirrelRshes, Holocentridae), or rely on olfac- high species diversity of seaweeds. By forming dense tion (e.g., moray eels, Muraenidae). schools, nonterritorial herbivores (parrotRshes and surgeonRshes) can successfully invade damselRsh Behavior territories.
0007 Overt behavioral interactions between coral-reef Piscivory and Defense Rshes include mutualism (when both species bene- R Predation is a major factor affecting the behavior 0010 t), interference competition (often manifested as R territoriality), and predator}prey relationships. and ecology of reef shes. There are three major modes of piscivory. Open-water pursuers, such as reef sharks (Carcharhinidae) and jacks (Caran- Mutualism gidae), simply overtake their prey with bursts of 0008 Three of the best-documented cases of mutualism speed. Bottom-oriented stalkers, such as grouper occur in reef Rshes. &Cleaning symbiosis' occurs (Serranidae) and trumpetRshes (Aulostomidae), when small microcarnivorous Rsh consume ecto- slowly approach their prey before a sudden attack. parasites or necrotic tissue off larger host Rsh, Bottom-sitting ambushers, such as lizardRshes which often allow cleaners to feed within their (Synodontidae) and anglerRshes (Antennariidae), mouths and gill cavities. The major cleaners are sit-and-wait cryptically for prey to approach them. various gobies (Gobiidae) and wrasses (Labridae). The vision of piscivores is often suited for crepuscu- Some of the cleaner wrasses are specialists that lar twilight, when the vision of their prey is least maintain Rxed cleaning stations regularly visited by acute (being adapted for either diurnal or nocturnal hosts, which assume solicitous postures. The inter- foraging). Hence, many prey species are inactive action is not always mutualistic in that cleaners during dawn and dusk, resulting in crepuscular occasionally bite their hosts, and some saber-tooth &quiet periods' when both diurnal and nocturnal
RWOS 015 CHANAKSHI NP FISH / Coral-Reef Fish 3
species shelter in the reef framework. (ParrotRshes sal (bottom-oriented) juvenile and adult stage. Most may further secrete mucous cocoons around them- bony reef Rshes broadcast spawn, releasing gametes selves at night, and small wrasses may bury in the directly into the water column where they are swept sand.) Otherwise, prey defensive behavior when to the open ocean. Smaller species spawn at their foraging or resting typically involves remaining war- home reefs and some larger species, such as some ily near structural shelter and shoaling either within grouper (Serranidae) and snapper (Lutjanidae), mi- or among species. Associated with day}night shifts grate to traditional sites and form massive spawning in activity are daily migrations between safe resting aggregations. Gametes are released during a paired areas and relatively exposed feeding areas. Cari- or group &spawning ascent' followed by rapid return bbean grunts (Haemulidae) spend the day schooling to the seaSoor. Exceptions to broadcast spawning inactively on reefs, and after dusk migrate to nearby include demersal spawners that brood eggs until seagrass beds and feed. Reproducing reef Rshes may they hatch, either externally (e.g., egg masses de- avoid predation by spawning (in some combination) fended by damselRshes) or internally (e.g., mouth- offshore, in midwater, or at night. Spawning during brooding cardinalRshes, Apogonidae), and a few ebbing spring tides that carry eggs offshore or guar- ovoviviparous or viviparous species that give birth ding broods of demersal eggs further defends to well-developed juveniles (including reef sharks propagules from reef-based predators. Subsequent and rays). Annual fecundity of broadcast spawners settlement of larvae back to the reef which occurs ranges from about 10 000 to over a million eggs per mostly at night, is also an apparent antipredatory female. Spawning is weakly seasonal compared to adaptation. temperate species, typically peaking during summer months but not strongly related to any particular Reproduction environmental variable. Lunar and semilunar spawning cycles are common. These are presumably Social Systems and Sex Reversal adaptations to transport larvae offshore away from reef-based predation, to maximize the number of 0011 The best-studied examples of highly structured so- settlement-stage larvae returning during favorable R cial systems in reef shes are the harems of wrasses conditions that vary on lunar cycles, or to beneRt R R and parrot shes. Typically, these sh are born as spawning adults in some way. females that defend individual territories or occupy Little is known about the behavior and ecology of 0013 a shared home range. A larger male defends a group reef-Rsh larvae. Duration of the pelagic larval stage of females from other males, thereby sequestering ranges from about 9 to well over 100 days, aver- matings. When the male dies, the dominant (typi- aging about a month. Larval prey include a variety cally largest) female changes sex (protogyny) and of small zooplankters. Comparisons of fecundity at becomes the new harem master. At high population spawning to subsequent larval settlement back to R sizes, some sh may be born as males, develop huge the reef suggest that larval mortality is both ex- R testes, resemble females, in ltrate harems, and sneak tremely high and extremely variable, apparently due spawnings with the resident females. Spatially iso- mostly to predation. Patterns of endemism, settle- R lated at their home anemones, anemone shes have ment to isolated islands, and limited data tracking social systems in which the largest individual is larvae directly suggest that there is considerable lar- female, the second largest is male, and the remain- val retention at the scale of entire islands. Later- R ing sh are immature (i.e., monogamy). Upon the stage larvae are active swimmers and may control death of the female, the male changes sex (pro- their dispersal by selecting currents among depths. R tandry) and the behaviorally dominant juvenile sh The overall reproductive strategy is apparently to matures into a male. Simultaneous hermaphroditism disperse the larvae offshore from reef-based pred- occurs among a few sparids and serranine sea ators, but then to retain them close enough to shore R basses. These sh have elaborate courtship behav- for subsequent settlement. iors during which individuals switch male and Settlement, the transition from pelagic larva to 0014 female roles between successive pair spawnings. life on the reef (or nearby nursery habitat), occurs at Regardless of the broad variety of mating systems a total length of about 8 mm to about 200 mm. R found in reef shes, each individual behaves in Larger larvae are either morphologically distinct a way that maximizes lifetime reproductive success. (e.g., the acronurus of surgeonRshes) or essentially pelagic juveniles (e.g., squirrelRshes and porcu- Life Cycle pineRshes). Choice of settlement habitat is apparent 0012 The typical bony reef Rsh has a bipartite life cycle: in some species, and both seagrass beds and man- a pelagic egg and larval stage followed by a demer- groves can serve as nursery habitats. Some wrasses
RWOS 015 CHANAKSHI NP 4 FISH / Coral-Reef Fish
larvae bury in the sand for several days before der being benthic invertebrate eaters or detritivores. emerging as new juveniles. There is typically weak The benthivores can be further subdivided based on metamorphosis during settlement involving the prey taxa (e.g., corallivores) or other categories growth of scales and onset of pigmentation. Esti- (e.g., consumers of hard-shelled invertebrates). mates of settlement are generally called &recruitment' Grunts that migrate from reefs at night and feed in and are based on counts of the smallest juveniles surrounding seagrass beds subsequently return nutri- that can be found some time after settlement. Once ents to the reef as feces. There is also considerable settled, most reef Rsh are thought to live less than consumption of Rsh feces by other Rsh on the reef. a decade, although some small damselRsh live at Fishes thus contribute substantially to nutrient trap- least 15 years. ping (via planktivory and nocturnal migration) and recycling (via coprophagy and detritivory) on coral reefs. Within each feeding guild, there is typically Ecology resource partitioning: each species consumes a par- 0015 Coral-reef Rshes are superb model systems for ticular subset of the available prey or forages in studying population dynamics and community a distinct microhabitat. Communities are also struc- structure of demersal marine Rshes because they are tured temporally, with a diurnal assemblage being eminently observable and experimentally manipul- replaced by a nocturnal assemblage (the resting as- able in situ. semblage sheltering in the reef framework). The diurnal assemblage is dominated by perciform and Population Dynamics tetraodontiform Rshes, whereas the nocturnal as- semblage is dominated by beryciform Rshes (evolu- 0016 Because reefs are patchy at all spatial scales and reef R R tionary relicts apparently relegated to the night by sh are largely sedentary, coral-reef shes form more recently evolved Rshes). metapopulations: groups of local populations linked by larval dispersal. Many local populations are de- mographically open, such that reproductive output Maintenance of Species Diversity drifts away and is thus unrelated to subsequent Four major hypotheses have been proposed to ex- 0018 larval settlement originating from elsewhere. Ulti- plain how many species of ecologically similar mately, the degree of openness depends on the R R coral-reef shes can coexist locally. There are data spatial scale examined. For example, anemone sh that both corroborate and falsify each hypothesis in populations are completely open at the scale of each various systems, suggesting that no universal gener- anemone, may be partially closed at the scale of an alization is possible. The Rrst two hypotheses are oceanic island, and mostly closed at the scale of an based on the assumption that local populations are archipelago. It is clear that variability in population not only saturated with settlement-stage larvae, but size is driven by variation in recruitment due to also regularly reach sizes where resources become larval mortality (and perhaps spawning success). limiting. First, the &competition hypothesis,' bor- Recruitment varies considerably at virtually every rowed from terrestrial vertebrate ecology, suggests spatial and temporal scale examined. The mecha- that coexistence is maintained despite ongoing inter- nisms naturally regulating populations are less clear, speciRc competition by Rne-scale resource partition- but are probably twofold. First, given that density- ing (or niche diversiRcation) among species. Second, dependent growth is common and that there is the &lottery hypothesis,' derived to explain coexist- a general exponential relationship between body R R ence among similar territorial damsel shes that did size and egg production in sh, density-dependent not appear to partition resources, is based on the fecundity is likely. Second, early postsettlement assumptions that, in the long run, competing species mortality is often density-dependent, and has been are approximately equal in larval supply, settlement demonstrated experimentally to be caused by pred- rates, habitat and other resources requirements, and ation. competitive ability. Thus, settling larvae are likened to lottery tickets, and it becomes unpredictable Community Structure which species will replace which following the ran- 0017 Due to high local species diversity, reef-Rsh commu- dom appearance of open space due to the death of nities are complex. There are about Rve major feed- a territory holder or the creation of new habitat. ing guilds, each containing dozens of species locally The relatively restrictive assumptions of this hypoth- (with approximate percentage of total Rsh biomass): esis can be relaxed if one considers the &storage zooplanktivores (up to 70%), herbivores (up to effect,' which is based on the multi-year life span of 25%), and piscivores (up to 55%), with the remain- reef Rshes and the fact that settlement varies
RWOS 015 CHANAKSHI NP FISH / Coral-Reef Fish 5
through time. Even though a species is at times an diurnal and nocturnal assemblages. Communities inferior competitor, as long as adults can persist are variously structured by recruitment, competi- until the next substantial settlement event, that spe- tion, and predation in a way that maintains high cies can persist in the community indeRnitely. The local species diversity. third hypothesis, &recruitment limitation,' assumes that larval supply is so low that populations seldom See also reach levels where competition for limiting re- sources occurs, so that postsettlement mortality is Climate: Corals and Climate Change. Coastal Re- 0020 density independent and coexistence among species gimes: Coral Reefs. Diversity, Marine Species. Fish: is guaranteed. Finally, the &prediction hypothesis' Feeding/Foraging (Behaviour, Food Chains, Feeding predicts that early postsettlement predation, rather Mechanisms, Density Dependence); Predation/Mortality; than limited larval supply, keeps populations from Reproduction. Fisheries: Fisheries for Coral Reef/Tropi- reaching levels where competition occurs, thereby cal Species. ensuring coexistence. Further Reading Summary Birkeland C (ed.) (1997) Life and Death of Coral Reefs. New York: Chapman and Hall. 0019 About 4500 Rsh species inhabit coral reefs globally, R BoK hlke JE and Chaplin CCG (1993) Fishes of yet their sheries are overexploited, their habitat is the Bahamas and Adjacent Tropical Waters. 2nd edn. threatened, and there are indications that species are Austin: University of Texas Press. endangered. Coral-reef Rshes are so diverse that Caley MJ (ed.) (1998) Recruitment and population there are many exceptions to virtually every general- dynamics of coral-reef Rshes. Australian Journal of ization that can be made about them. The typical Ecology 23(3). species is a distinctively colored, highly compressed Helfman GS (ed.) (1978) Patterns of community structure perciform that readily maneuvers and picks small in Rshes. Environmental Biology of Fish 3(1). invertebrates from complex reef habitats. Regarding Longhurst AR and Pauly D (1987) Ecology of Tropical behavior, mutualism, territoriality, antipredatory Oceans. San Diego: Academic Press. mechanisms, and complex social systems (sometimes Montgomery WL (1990) Zoogeography, ecology and behavior of coral reef Rshes. In: Dubinskiz (ed.) Coral involving sex reversal) are common. Reproduction is Reefs, pp. 329}364. Amsterdam: Elsevier. typically via broadcast spawning, with pelagic larval Polunin NVC and Roberts CM (1996) Reef Fisheries. duration averaging about a month, and life span less London: Chapman and Hall. than a decade. Population dynamics are apparently Randall JE, Allen GR and Steene RC (1997) Fishes of the driven by Suctuations in larval mortality, and popu- Great Barrier Reef and Coral Sea. 2nd edn.Honolulu: lations are probably regulated in the absence of University of Hawaii Press. Rshing by density-dependent fecundity and early Sale PF (ed.) (1991) The Ecology of Fishes on Coral postsettlement mortality via predation. Community Reefs. San Diego: Academic Press. structure is complex, involving numerous feeding Sale PF (ed.) (2001) The Ecology of Fishes on Coral guilds, and includes day}night transitions between Reefs. 2nd edn. San Diego: Academic Press. Thresher RE (1984) Reproduction in Reef Fishes. Neptune City: T.F.H. Publishers.
RWOS 015 CHANAKSHI NP