Ecosystems (2004) 7: 358-367 DOI: 10.1007/S10021-004-0184-X ECOSYSTEMS! ) 2004 Springer-Verlag High Complexity Food Webs in Low- diversity Eastern Pacific Reef-Coral Communities Peter W. Glynn Division of Marine Biology and Fisheries, Rosenstiel School of Marine and Atmospheric Science, 4600 Rickenbacker Causeway, Miami, Florida 33149, USA ABSTRACT Community-wide feeding interrelationships in a low-diversity coral reef off the Pacific coast of Pan- am? (Uva Island reef) demonstrate complex path- ways involving herbivore, strong corallivore, and carnivore interactions. Four trophic levels with 31 interguild links are identified in a generalized food web, and documented feeding interrelationships with 287+ species links are portrayed in a coral- corallivore subweb. The importance of trophic groups changes greatly with time, from unknown causes over annual to decadal-scale periods, and during very strong El Nino-Southern Oscillation events such that intermittent intense herbivory by echinoids (Diadema) and corallivory by gastropod moUusks, the crown-of-thorns sea star Acanthaster, hermit crabs, and fishes result in high levels of coral mortality and bioerosion of reef substratum. Intrar- egional differences in species composition and abundances affecting food-web interactions are briefly described for nonupwelling (Uva Island) and upwelling areas (Pearl Islands) in Panam?. Seasonal upwelling in the Pearl Islands results in high plank- ton productivity, which likely augments production in invertebrates, fishes, marine mammals, and sea- birds, but these pathways still remain largely un- quantified. The corallivore Acanthaster is absent from upwelling centers in Panam? and from up- welling and nonupwelling areas in the southern and central Gal?pagos Islands, and the highly de- structive, facultative corallivore Eucidaris galapagen- sis occurs only in the latter offshore islands and at Cocos Island. Relatively recent declines in the abundances of manta rays, sharks, and spiny lob- sters are correlated with, but not necessarily caus- ally linked to, increasing fishing activities in the late 1970s to early 1980s. The extent to which the com- plex yet highly unstable Uva Island food web is representative of other eastern Pacific coral reef ecosystems remains to be investigated. Key words: food web; coral reef; eastern Pacific; Panam?. INTRODUCTION With only about 8-10 zooxanthellate (reef-build- ing) coral species present on any given reef, and three genera {Pocillopora, Porites, and Pavona) con- tributing dominantly to reef accretion in the eastern Pacific, the array and complexity of known trophic interactions are unexpectedly high in this low-di- versity coral reef region. High-diversity coral reef Received 10 May 2002; accepted 30 December 2002; published online 27 April 2004. Corresponding author: e-mail: pglynn@rsmas.mianii.edu ecosystems in the Indo-West Pacific and Greater Caribbean biogeographic regions support from two to more than ten times the number of reef-building coral and fish species present on eastern Pacific coral reefs (Paulay 1997). Numerous direct and in- direct feeding effects involving polychaetous worms, crustaceans, moUusks, echinoderms, and fishes have been identified during the last 30 years of coral community studies ranging from the Gulf of California to Ecuador (Barham and others 1973; Gilchrist 1985; Glynn 1973, 1974, 1976, 1977b, 1982a, 1982b, 1983a, 1983b, 1984, 1985a; Glynn and others 1979, 1982; Glynn and Colgan 1988; 358 Eastern Pacific Coral Reef Food Webs 359 '?V:^ Gulf of Califarnia 70' Mexicci Gulf of Uva ?hiriqui ^ '^ Is. reef PACIFIC OCEAN Galapagos Islands Gulf of Panama 110? 100? 90? 80? 70? Figure 1 Eastern Pacific coral reef areas and location of Uva Island coral reef study site. Guzman 1988a, 1988b; Guzman and L?pez 1991; Guzman and Robertson 1989; Reyes Bonilla and Calder?n Aguilera 1999; Wellington 1982a). How- ever, no comprehensive analysis of food-web rela- tionships has yet been published for any coral reef community in the eastern Pacific. I construct the first ecosystem-level food web of a coral reef in Panam? that my colleagues and I have continuously studied since 1970. This coral reef is located in the Gulf of Chiriqui (Figure 1), a nonup- welling environment supporting numerous small patch and fringing pocilloporid reefs. The reef is built predominantly by two frame-building coral species in the genus Pocillopora (P. damicornis and P. elegans), with secondary contributions from species in the genera Parit?s, Gardineroseris, Pavona, Psammo- cora, and Millepora (a hydrocoral). Unlike coral reefs in other regions, where crustose coralline algae are often abundant and contribute importantly as bind- ing agents in framework construction, calcifying algae are generally unimportant to reef develop- ment in the eastern Pacific. The Uva Island patch reef covers about 2.5 ha and consists of a reef flat, reef slope, and seaward reef base at a maximum depth of approximately 6 m relative to mean low water (MLW) (Glynn and Mat? 1997; Eakin 2001). Core drilling shows the reef framework to be 8-12 m thick, with a maximum age of about 5000 years (Glynn and Macintyre 1977). Strong seasonal pat- terns at Uva Island greatly influence environmental conditions between the wet and dry seasons (Table Elevated sea temperatures associated with two exceptionally strong El Nino-Southern Oscillation (ENSO) events in 1982-83 and 1997-98 resulted in extensive reef-coral mortalities over most of the equatorial eastern Pacific (Glynn 1990, 2002; Glynn and Colgan 1992, Glynn and others 2001). ENSO- induced coral mortality in the Gal?pagos was 97% and 26% following 1982-83 and 1997-98, respec- tively. The Uva reef suffered about 75% overall coral mortality following the 1982-83 ENSO and about 13% from the 1997-98 ENSO. These distur- bances need to be considered because they signifi- cantly altered the relative abundances of corals, the numerous species that prey on them, and the im- portant links in the food-web design. It is also im- portant to note that the magnitude of fluxes varies greatly over time, based on the presence of various species during a 30-year period. Several of these species have demonstrated marked variations in abundance, intermittently contributing signifi- cantly to food-web interactions or, when rare, scarcely having any effect. COMMUNITY-WIDE FOOD WEB Four trophic levels with 31 interguild links (includ- ing at least five intraguild pathways) have been identified in the Uva Island coral community (Fig- ure 2). All nonreferenced, generalized, and species- specific pathways presented here and below for macrobiota and fishes are based on personal con- sumer-prey observations and gut-content analyses performed on field populations. Detritus, constitut- ing particulate organic matter (for example, decay- ing organisms and feces), receives contributions from all reef biota and is consumed by decomposers and by suspension and deposit feeders. Decompos- ers such as microbes and species of meiofauna and microfauna are present in organic-rich sediment patches at the reef base, often encircling coral col- onies and forming deposits in the internal structure of reefs where they accumulate in the highly po- rous basal reef framework. The organic matter con- tent of sediments is also enriched by particulate fallout from river runoff and plankton advected onto the reef. Studies of meiofauna species under- taken on a Costa Rican reef found foraminifers, copepods, nematodes, and gastropods to predomi- nate (Guzman and others 1987), and this is ex- pected for the Uva Island reef. This rich (largely microscopic) community demonstrates complex feeding interactions involving meiofauna that prey on microbes, and microfauna that consume meio- 360 P. W. Glynn Table 1 Physical Environmental Conditions Influencing Offshore Coral Reefs in the Gulf of Chiriqui, Panam? Factor Wet Season Dry Season 2,598 Cloud Cover" median % Rainfall mm*', mean annual Salinity, surface" Median (range) Tides^ semidiurnal, unequal mean spring range = 3.29 m Light Penetration'' depth (m) of 10% surface radiation Sea Temperature (?C)''^ mean annual SST a~ 28?C Nutrient Concentration'' PO4-P (|xg-at 1"') at 40 m depth 23.2-41.7 423, maximum (Sep) 28.0 (27.0-32.0) low water exposures near midnight low salinity stress from rainfall 10, minimum (Sep) 0.5''-1.0'= (Jul-Sep) 30-31, maximum (14 wk) 0.3, minimum (Oct, Nov) "Glynn 1977b ''Kwiecinski and Chial 1983 ^U.S. Department of Commerce, calculated from tidal predictions for Balboa, Panama ^Glynn unpublished observations 'Dana 1975 ^Renner 1963 ^Glynn and others 2001 6.8-9.2 16, minimum (Feb) 34.0 (28.0-36.0) low water exposures at midday intense solar exposure, heating, desiccation 37, maximum (Jan) 25 minimum (1 wk) 1.1, maximum (Mar) fauna; hence, the redirected arrow in trophic cate- gory 4 (Figure 2). Reef waters in the Gulf of Chiriqui are often filled with phytoplankton and Zooplankton, depending upon nutricline shoaling and nutrients advected to reefs via wet-season river discharge (L. D'Croz un- published data). The sporadic abundance of plank- ton populations and ill-defined current movements around the Uva reef make this potential food source difficult to track and quantify. Phytoplankton and Zooplankton are consumed by suspension feeders, such as sponges, sea anemones, polychaete worms, lithophage bivalves, vermetid gastropods, barna- cles, porcelanid crabs, and brittle stars, and by zoo- plankton carnivores. These animals in turn contrib- ute toward the organic matter content of sediments. Consumers of Zooplankton include corals and a variety of fishes (Paranthias colonus, Kyphosus el- egans, and damselfishes, mostly species of Chromis), and manta rays {Manta birostris). The relative con- tributions of Zooplankton and photosynthates from algal endosymbionts toward the caloric intake of zooxanthellate corals are unknown in the eastern Pacific. However, Wellington (1982b), in a field study in the Pearl Islands (Panam?), demonstrated that maximum skeletal growth in two massive coral species {Pavona clavus and Pavona gigantea) was de- pendent on the availability of Zooplankton (> 95 |jLm net mesh size), whereas the branching coral Pocillopora damicornis grew independent of zoo- plankton supply. Turf and crustose coralline algae are most abun- dant on reef fiats and in reef base habitats, whereas green frondose algae occur most commonly in the deeper reef base and talus slope zones. Reef slopes in forereef and backreef zones are dominated by nearly continuous stands of live Pocillopora spp. (Ta- ble 2). The occurrence of macroalgae is intermittent and may depend on nutrient concentrations arising from pycnocline shoaling and rivers and runoff as well as fiuctuations in herbivore populations. When abundant, macroalgae such as Caulerpa spp. can overgrow and kill low-lying corals such as Psammo- cora stellata and Porites panamensis (Glynn and Mat? 1997). Endolithic algae have been observed in coral skeletons but are unstudied, and little is known about their distribution and abundance in eastern tropical Pacific reefs. Reef fiat turf algae are heavily grazed by large schools of scarid and acanthurid fishes that move into this zone at high water. Zooxanthellate corals harbor both dinoflagellate endosymbionts and crustacean symbionts, the latter moving among colony branches or occupying shel- ters within coral skeletons. Symbiodinium clades C Eastern Pacific Coral Reef Food Webs 361 TROPHIC LEVELS 4 Top-Level Carnivores ETTl C Small-Medium Carnivores l??t Corail i vores 13 Zooplankton Decomposers Benthic Algae ~i. ^ Zooplankti vores Suspension Feeders ??T Deposit Feeders Phytoplankton Zooxanthellate Corals Symbiotic Metazoans Symbiotic Algae Figure 2 Conceptual coral reel food web at Uva Island, nonupwelling Gulf of Chiriqui, Panam?. Trophic level assignments follow Opitz (1996), which are from a modeling study of a Caribbean coral reef. The following categories are recognized: 1 benthic algae?algal turf, endolithic algae, crustose coralline algae, and macroalgae; 2 phytoplankton?femtoplankton to mesoplankton; 3 detritus?particulate organic matter; 4 symbiotic algae?Symbiodinium clades A, C, and D; 5 decompos- ers?microorganisms, meiofauna and microfauna; 6 suspension feeders?sponges, polychaete worms, lithophage bivalves, and barnacles; 7 deposit feeders?sea cucumbers, shrimp, and spiny lobster; 8 symbiotic metazoans?crabs, shrimp; 9 Zooplankton?cnidarians, crustaceans, and various larval stages; 10 zooplanktivores?cnidarians, damselfishes, and manta rays; 11 zooxanthellate corals?scleractinian corals and hydrocorals; 12 herbivores?gastropods, echinoids, damselfishes, angelfishes, parrotfishes, kyphosids, acanthurids, and turtles; 13 corallivores?hermit crabs, gastropod moUusks, sea stars, and pufferflsh; 14 invertivores?polychaete worms, octopus, hawkfish, jacks (Caranx speciosus), triggerflsh, wrasses, and serranids; 15 small to medium carnivores?polychaete worms, crustaceans (Hymenocera, spiny lobster), puffers, balistids, eels, snappers, wrasses, groupers, jacks, and sharks; 16 top-level carnivores?sharks, barracuda, wahoo, snappers, grou- pers, and mackerels. and D are present in scleractinian corals and Sym- biodinium clade A in Millepora hydrocorals (Glynn and otfiers 2001). At least 55 species of decapod crustaceans have been found on Pocillopora damicor- nis colonies alone in the Gulf of Panam? (Abele and Patton 1976). Some of these directly graze coral polyps (Glynn 1983b), feed on coral mucus, remove food from polyps, or feed on fat bodies produced by the coral host (Stimson 1990). Coral tissues (or tissues plus skeleton) are con- sumed by numerous corallivores, noted in more detail below. The diets of many corallivore species are not restricted; they also consume benthic algae, suspension feeders, and invertivores (here treated as noncoral consumers of various invertebrate taxa) that prey on crustaceans living symbiotically with corals. Resident invertebrate and fish carnivores feed on herbivores, planktivores, corallivores, and invertivores. Panulirus gracilis (spiny lobster), here grouped with carnivores, largely feeds on the flesh of moribund and decaying organisms. Sea turtles, notably the olive Ridley (Lepidochelys oliv?cea) and the loggerhead (Caretta caretta), destroy live pocillo- porid frameworks while searching for and feeding on sponges attached to the coral's dead subsurface skeletons. Finally, the largest transient carnivores, such as sharks, wahoo, mackerels, and tunas, prey on resident carnivores and consumers at lower tro- phic levels as well. Whitetip reef sharks, Triaenodon obesus, have also been observed to cause extensive breakage of pocilloporid frameworks while pursu- ing their prey sheltered in the reef (Jim?nez 1996- 97). Bird and mammal consumers have not been observed. Even when low water exposes the reef flats and causes massive mortality of invertebrates and fishes, no birds appear to feed on them. 362 P. W. Glynn Table 2 Uva Reef Benthic Cover and Abundances of the More Common Vagile Invertebrates and Fishes. Reef zone: RB, reef base; FR. fore reef; RF, reef flat; BR, bade reef Organism Source Benthic Cover (%) live Podllopora spp. 0.23 (RB) 56.5 ' (FR) 7.6 (RF) 53.8 (BR) Eakin 1996 filamentous algae 6.5 (RB) 3.6 (FR) 0.0 (RF) 14.0 (BR) sampling 1988-1994 crustose coralline and 89.4 (RB) 38.4 . (FR) 91.6 ( RF) 32.0 (BR) fleshy algae Invertebrate Abundances Non-pocilloporid corals 8.4 (RB) Glynn 1974 relative % abundance" Diadema mexicanum 58 ? 13 (RB) 21 ? 8 (FR) < 1.0 (RF) < 1.0 (BR) Eakin 2001*' ind m^^ < 1.0 (RB) 6 ? 1 (FR) < 1.0 (RF) < 1.0 (BR) Glynn unpublished data^ Acanthaster planci 12.5 ? 2.3 Fong and Glynn 1998"* ind ha^', entire reef 1.0 (0-2) Glynn unpublished data" Trapezia spp/ 1.5- ?8.5 (FR) 3- ?9.5 (RF) Glynn 1976 Alpheus lottinf 2 (FR) 2 (RF) 1 median ind colony^ ' Hymenocera picta^ 1-118 Glynn 1977a ind ha^', entire reef Pherecardia striata^ 170 ? 95 RF Glynn 1984 ind m^^ median ? c.l. 0 RB Fish Abundances Glynn unpublished data mean ? 95% ci. ha"' n = 27 transects sampled during 2000-2002 Balistidae 106.0 ? 15.8 RB, FR, and RF zones flsh sizes > 15 cm TL Scaridae 100.0 ? 37.3 Arothron spp. 66.2 ? 15.4 Lutjanidae 39.8 ? 37.3 Acanthuridae 34.2 ? 20.1 Chaetodontidae + 30.0 ? 8.5 Pomacanthidae Labridae 20.8 ? 8.3 Serranidae 9.3 ? 5.1 Diodontidae 2.3 ? 2.3 ^calculated relative to live Podllopora spp. cover ''sampling period, 1988-1994 ^sampling period, 2000-2002 ^maximum population density 1987, mean ? SE, sampling period 1983-1984 ^mean and range, sampling period 2000-2002 ^Podllopora spp. sampled in 1974 ^range, sampled in 1976 '?sampled in 1980 CORAL-CORALLIVORE SUBWEB The coral-corallivore subweb depicts 287+ known interspecific links among coral prey, invertebrate and fish corallivores, invertivores, and top-level fish predators (Figure 3). This subweb includes only the chief reef-building corals {Podllopora spp.. Pavona spp., Gardineroseris planulata, and Porites lobata) and taxa that often contribute significantly to coral community cover {Psammocora stellata and Millepora intricata). An undetermined but probably minor amount of feeding, by means of extracoelenteric digestion, oc- curs interspecifically among corals growing in close proximity. This kind of interaction is indicated be- tween Podllopora spp. and Pavona spp., where Well- ington (1980) demonstrated that a feeding response initially dominated by Podllopora damicornis was later reversed through the development of sweeper tentacles by Pavona gigantea. The tissues of all scleractinian coral species are consumed by at least two corallivore species, but Eastern Pacific Coral Reef Food Webs 363 Hemigaleidae, Carcharhinidae, Sphyraenidae, Scombridae, Serranidae ^ Muraenidae, Lutjanidae, Labridae, Serranidae y 23 Figure 3 Documented feeding interrelationsiiips among zooxantiiellate scleractinian corals and hydrocorals, and coralli- vores and their predators at Uva Island, nonupwelling Gulf of Chiriqui, Panam?. Strong pathways are indicated by the thick arrows, denoting frequent prey consumption. Species identities: 1, Gardineroseris planulata; 2, Pavona spp. (Pavona davus, Pavona gigantea, and Pavona variant); 3, Pocillopora spp. (Pocillopora damicomis and Podllopora elegans); 4, Porites lobata; 5, Millepora intricata; 6, Psammocora stellata; 7, Acanthaster plana; 8, Quoyula madreporarum; 9, Phestilla sp.; 10, Jenneria pustulata; 11, Pentaceraster cumingi; 12, Arothron spp. {A. meleagris and A. hispidus); 13, Stegastes acapulcoensis; 14, Chaetodontidae and Pomacanthidae (Johnrandallia nigrirostris, Chaetodon humeralis, Holacanthus passer, and Pomacanthus zonipectus); 15, hermit crabs (Trizopagurus magnificus, Aniculus elegans, and Caldnus obscurus); 16, Monacanthidae (Aluterus scriptus and Cantherhinus dumerilii); 17, Scaridae (Scarus perrico and S. rubroviolaceus); 18, Hymenocera pida; 19, Pherecardia striata; 20, Balistidae (Sufflamen verres and Pseudobalistes naufragium); 21, Odopus spp.; 22, Diodontidae (Diodon holocanthus and Diodon hystrix); 23, Muraenidae (Gymnothorax castaneus and Gyninothorax dovii), Lutjanidae (Lutjanus viridis, Lutjanus argentiventris, and Lutjanus novemfasdatus), Labridae (Bodianus diplotaenia, Novaculichthys taeniourus, and Thalassonia grammaticum), and Serranidae (Epinephelus panamensis and Epinephelus labriformis); and 24, Hemigaleidae (Triaenodon obesus), Carcharhinidae (Carcharhi- nus leucas), Sphyraenidae (Sphyraena ensis), Scombridae (Acanthocybium solandri), and Serranidae (Myderoperca xenarcha). Pocillopora spp., which are the principal framework builders on Panamanian reefs, are preyed on by at least nine groups of consumers, some with multiple species. Pufferfishes {Arothron), parrotfishes (Scari- dae), and filefishes (Monacanthidae) bite off colony branch-tips, and hermit crabs (Trizopagurus, Anicu- lus, and Caldnus) scrape skeletal surfaces, generat- ing coarse to fine-grained calcareous sediments (Glynn and others 1972). Removing only tissues and leaving the skeleton intact are the gastropods Jenneria pustulata and Quoyula madreporarum, but- terflyfishes and angelfishes, the damselfish Stegastes acapulcoensis, and Acanthaster planci, the crown-of- thorns sea star. Although damselfish nip and ingest coral tissues when establishing or enlarging algal mats, the contribution of this food source toward the fish's dietary needs is unclear (Wellington 1982a). Acanthaster is a corallivore generalist, prey- ing on nearly all coral taxa but consuming espe- cially large amounts of Pocillopora and Gardineroseris (Glynn 1974). Acanthaster feeds more on small or broken branches of Pocillopora, which receive less protection by crustacean guards than large intact colonies (Glynn 1983a). Both adult and juvenile butterflyfishes and angelfishes nip and ingest Pocil- lopora polyps. They have not been seen feeding on other coral genera. Jenneria and Acanthaster can kill relatively large (approximately 30 cm in diameter) whole colonies of Pocillopora. Like the gastropod Drupella in the western Pacific, Jenneria sometimes 364 P. W. Glynn overwhelms Pocillopora with feeding aggregations exceeding 50-100 individuals on single 30-cm-di- ameter colonies. Arothron spp. also bite off the branch tips of Mille- pora intricata and Psammocora spp., and remove cen- timeter-sized fragments from protuberances and ridges on the massive colonies of Porites lobata and Pavona spp. Balistids, notably Pseudobalistes naufra- ?ium, also bite 2- to 5-cm-diameter sections from Porites and Pavona corals, but in search of lithoph- agine bivalves that they extract and consume. Some of these fragments survive this breakage and pro- duce new colonies, thus contributing to the asexual propagation of these corals (Guzman 1988a). Phes- tilla sp., a small and secretive aeolid nudibranch, feeds on Porites lobata at night. The sea star Pentac- eraster cumingi occasionally feeds on Psammocora stel- lata in reef slope and base zones with its everted stomach also digesting sponges and other metazo- ans attached to the coral colony's basal branches. No predators have been observed feeding on Phes- tilla sp. or P. cumingi. Especially strong trophic pathways in this subweb, involving pocilloporid corals and Gardin- eroseris planulata, are supported by several studies. The consumption of Pocillopora spp. tissues by Jen- neria pustulata, Acanthaster planci, and Arothron me- leagris is notably high compared to that by hermit crabs, other gastropods and fishes, and other mem- bers of the corallivore guild [quantified by Glynn and others (1972) and Glynn (1973, 1976, 1977b, 1984, 1985a)]. From 1983 to 1985, Jenneria con- sumed about twice the amount of coral tissues as did Acanthaster and the latter about seven times that oi Arothron (Glynn 1985a), but these values varied widely through time depending on the fluctuating abundances of corallivore populations. On the Uva Tee?,Gardineroseris is preferentially preyed on by Acanthaster (Fong and Glynn 2000). In turn, field and mesocosm studies (Glynn 1977b, 1984) indi- cate that Acanthaster abundance on the Uva reef is controlled by the harlequin shrimp Hymenocera picta and the polychaete worm Pherecardia striata. Shrimps and worms often feed on moribund Acan- thaster and may, therefore, be scavengers rather than predators. Finally, several pathways from cor- allivores or corallivore consumers lead to higher- level resident carnivores such as eels and snappers, and ultimately to top-level transient carnivores such as groupers, wahoos, and sharks. TEMPORAL AND SPATIAL VARIATIONS As already noted, eastern Pacific coral reefs have been subject to significant declines in live coral cover over the past 30 years, during periods of strong ENSO disturbances (Glynn and Colley 2001; Reyes Bonilla and others 2002). Corallivores and other consumers have also demonstrated significant fiuctuations in abundances over this period. For example, Acanthaster planci ranged from 36 ind (14 ind ha"^) on the Uva reef in 1972 (Glynn 1973) to 0 in 2001 (Fong and Glynn 1998; P. W. Glynn unpublished data). Also, Jenneria pustulata popula- tions have ranged from 20,000 to 60,000 (8000- 24,000 ind ha"^) on the Uva reef in the 1970s (Glynn and others 1972; Oramas 1979) to 400,000 to 700,000 (160,000-280,000 ind ha"^) in 1982 (Glynn 1985a). Tens of thousands of Jenneria were observed consuming whole large (approximately 30 cm in diameter) pocilloporid colonies in 1982. These abundances declined sharply following 1983, a likely result of elevated ENSO temperatures and coral mortality, and remained low from 1985 on- ward (Achurra Cardenas and Valdes Arauz 1980; Glynn 1985a; P. W. Glynn unpublished data). Dia- dema mexicanum mean population densities in- creased dramatically after 1983, from less than 5 ind m^^ in reef base and forereef zones to 50 and 20 ind voT^, respectively (Eakin 2001). These high densities began to decline in the mid 1990s, ap- proaching pre-1983 levels by 2000, with the disap- pearance of shelter space due primarily to the ero- sive feeding activities of these sea urchins. In contrast to the aforementioned large changes in invertebrate numbers, mean Arothron meleagris abundances of 100-200 ind (40-80 ind ha"') on the Uva reef remained relatively constant for over 30 years (Glynn 1985a; Guzman and Robertson 1989; P. W. Glynn unpublished data). Although A. meleagris feeds predominantly on Pocillopora dami- cornis at the Uva reef, it concentrates on Porites lobata at Ca?o Island, Costa Rica. The diets and feeding preferences of the puffer at different sites are not related in a consistent way to the relative abundances of coral prey or their fluctuations in abundance following major mortality events (Guz- man and Robertson 1989). Notable, but poorly quantified, declines in abun- dances also have been observed in other consumer populations. In the 1970s, during each yearly re- search cruise to the Uva reef, 5-10 plankton-feed- ing Manta birostris could be seen gliding along the reef front during a 60-min dive. Often several (up to 13 ind) whitetip reef sharks and occasionally bull sharks, Carcharhinus leucas, were seen on the Uva reef in the 1970s. These species have been observed only rarely in the 1980s and 1990s, following shark-fishing activities off Uva Island in the late 1970s. The cryptic spiny lobster, Panulirus gracilis, Eastern Pacific Coral Reef Food Webs 365 was abundant under massive corals and hidden in the pocilloporid frameworks in the early to mid- 1970s, but virtually disappeared by the late 1970s and early 1980s, following large-scale extraction by commercial divers. Variations in food webs occur spatially within the eastern Pacific, although most differences have not been well quantified. In Panam?, for example, some significant differences are evident between coral reefs located in the nonupwelling Gulf of Chiriqui (Uva Island reef) and the seasonally up- welling environment of the Gulf of Panam?. The productivity of primary producers (phytoplankton and benthic algae) and the production of their con- sumers (Zooplankton, suspension feeders, and her- bivores) are expected to increase during the up- welling season, and remain relatively high for some undetermined time thereafter. Diadema mexicanum attains a significantly larger test diameter in the up welling Gulf of Panam? (= 18.7 ? 0.8 mm, n = 82, Uva reef, 16 March 2002) and therefore prob- ably demonstrates a higher per-capita rate of con- sumption in the former area. Also, Acanthaster is absent from the Gulf of Panam?, and Jenneria is apparently less abundant in the upwelling gulf than at Uva Island. Known only in the Gal?pagos Islands and at Cocos Island (Lessios and others 1999), the abundant and large grazing echinoid Eucidaris ga- lapagensis has a strong effect on algal-covered car- bonate surfaces. This sea urchin has bioeroded vir- tually all reef frameworks in the Gal?pagos Islands affected by the 1982-83 ENS O disturbance (Macin- tyre and others 1992; Glynn 1994; Reaka-Kudla and others 1997). SUMMARY COMMENTS A notable feature of the Uva Island trophic web is the large number of coral-corallivore pathways and the interannual variability of corallivory. At least eight invertebrate and 11 fish species consume cor- als on the Uva reef. This is probably attributable to the numerous consumer species that are adapted to feed on pocilloporid corals, which are notably abundant and among the main constructors of coral reefs in this region. Although a detailed subweb portraying invertebrate corallivore pathways is un- available for other coral reef ecosystems, some per- spective can be gained by comparing the fish coral- livores summarized by Jones and others (1991) for reef areas ranging from the central to the western Pacific. The highest numbers of fish corallivores listed were for Hawaii and Okinawa (nine species), followed by the Marshall Islands (six species), and outer-slope habitats on the Great Barrier Reef (five species). Thus, it is evident that the species-poor zooxanthellate coral fauna of eastern Pacific reefs is subject to intense pr?dation by fish corallivores, possibly proportionately more so than in species- rich communities in the central and west Pacific. Marked fluctuations in the population densities of Acanthaster and Jenneria result in highly variable rates of pocilloporid and agariciid coral mortalities (Glynn 1985a; Pong and Glynn 2001). Coral mor- tality resulting from ENS O warming also alters the relative abundances of coral prey and their spatial relationships relative to corallivore feeding activities (Glynn 1985b). Surviving coral populations may be subject to increased corallivory during and follow- ing such disturbances. Damselfish lawn expansion and accelerated bioerosion (largely by Diadema) of corals killed during ENSO disturbances cause rapid erosion and collapse of reef frameworks. This attempt to portray a coral reef food web in the eastern Pacific underlines the complexity of feeding relationships in a low-diversity coral com- munity, which will surely reveal additional intrica- cies with continuing studies. This exercise reveals that relatively little is known of the food-web links at the highest and lowest trophic levels, due in large measure to the mobility and sporadic feeding be- haviors of large predators and the dynamic nature of plankton and suspended/dissolved organic mat- ter supplies. Additional gaps in our knowledge re- late to species consumption rates and changes in diet that occur during development. These research areas offer challenging opportunities that will re- quire innovative approaches to map all critical tro- phic pathways, quantify energy and material flows, and determine the extent that consumers regulate coral community structure. ACKNOWLEDGEMENTS Studies contributing to this article were made pos- sible by support from the Smithsonian Tropical Re- search Institute, the National Geographic Society, and the National Science Foundation (Biological Oceanography Program). 1 thank the following Panamanian authorities for permission to work in their territorial waters: Recursos Marinos, INRENARE (Instituto Nacional de Recursos Natura- les Renovables) and the Parque Nacional de Coiba. For logistical support and permission to carry out research in the Gal?pagos Islands, 1 am grateful to the staff of the Charles Darwin Research Station and the Gal?pagos National Park Service. Com- ments by P. Fong, H. M. Guzman, N. Idrisi, D. Lirman, T. R. McClanahan, and H. Reyes Bonilla have helped to clarify several topics addressed in this article. A. L. Hazra and D. Lirman kindly helped 366 P. W. Glynn to produce the food-web graphics. Finally, I thank T. R. McClanahan for the invitation to address coral reef trophic relationships in the eastern Pacific and for his helpful critique of this contribution. REFERENCES Abele LG, Patton WK. 1976. The size of coral heads and the community biology of associated decapod crustaceans. J Bio- geogr 3:35-47. Achurra Cardenas JB, Valdes Arauz E. 1980. "Estudio sobre los efectos predatorios de los moluscos CoralUophila abbreviata yJenneria pustulata de ambos litorales del Istmo de Panam?." Thesis, Escuela de Biolog?a, Universidad de Panam?. 66 pp Barham EG, Gowdy RW, Wolfson FH. 1973. Acanthaster (Echi- nodermata, Asteroidea) in the Gulf of California. Fish Bull (Seattle) 71:927-42. Dana TF. 1975. Development of contemporary eastern Pacific coral reefs. Mar Biol 33:355-74. Eakin CM. 1996. Where have all the carbonates gone? A model comparison of calcium carbonate budgets before and after the 1982-1983 El Ni?o at Uva Island in the eastern Pacific. Coral Reefs 15:109-19. Eakin CM. 2001. A tale of two ENSO events: carbonate budgets and the infiuence of two warming disturbances and interven- ing variability, Uva Island, Panam?. Bull Mar Sei 69:171-86. Fong P, Glynn PW. 1998. A dynamic size-structured population model: does disturbance control size structure of a population of the massive coral Gardineroseris planulata in the eastern Pacific? Mar Biol 130:663-74. Fong P, Glynn PW. 2000. A regional model to predict coral population dynamics in response to El Nino-Southern Oscil- lation. Ecol Appl 10:842-54. Fong P, Glynn PW. 2001. Population abundance and size-struc- ture of an eastern tropical Pacific reef coral after the 1997-98 ENSO: a simulation model predicts field measures. Bull Mar Sei 69:187-202. Gilchrist SL. (1985) "Hermit crab corallivore activity." In: Pro- ceedings of the Fifth International Coral Reef Congress, Tahiti. Editors: M, Viviem , Harmelin and B. Salvat. Moorea, French Polynesia, Publisher: Antenne Museum-EPHE. Vol 5. p 2II-4. Glynn PW. 1973. Acanthaster. effect on coral reef growth in Panam?. Science 180:504-6. Glynn PW. 1974. The impact of Acanthaster on corals and coral reefs in the eastern Pacific. Environ Conserv 1:295-304. Glynn PW. 1976. Some physical and biological determinants of coral community structure in the eastern Pacific. Ecol Monogr 40:431-56. Glynn PW. 1977a. Coral growth in upwelling and nonupwelling areas off the Pacific coast of Panam?. J Mar Res 35:567-85. Glynn PW. (1977b) Interactions between Acanthaster and Hym- enocera in the field and laboratory. In: Proceedings of the Third International Coral Reef Symposium, Miami. Editor: D.H. Tay- lor City: Miami, Florida, Rosenthal School of Marine and Atmospheric Science, University of Miami. Vol I. p 209-15 Glynn PW. (1982a) Acanthaster population regulation by a shrimp and a worm. In: Proceedings of the Fourth Interna- tional Coral Reef Symposium, Manila. Editor: E.D. Gomez Quezon City, Phillippines: Marine Sciences Center, University of the Phillippines. Vol 2. p 607-12 Glynn PW. 1982b. Coral communities and their modifications relative to past and prospective Central American seaways. AdvMar Biol 19:91-132. Glynn PW. 1983a. Crustacean symbionts and the defense of corals: coevolution on the reef? In: Nitecki MH, Ed. Coevolu- tion. Chicago: University of Chicago Press, p 111-78. Glynn PW. 1983b. Increased survivorship in corals harboring crustacean symbionts. Mar Biol Lett 4:105-11. Glynn PW. 1984. An amphinomid worm predator of the crown- of-thorns sea star and general pr?dation on asteroids in east- ern and western Pacific coral reefs. Bull Mar Sei 35:54-71. Glynn PW. (1985a) "Corallivore population sizes and feeding effects following El Ni?o (1982-83)-associated coral mortality in Panam?." In: Proceedings of the Fifth International Coral Reef Congress, Tahiti. Editors: C. Gabria and B. Salvat. Moorea, French Polynesia: Antenne Museum-EPHE. Vol 4. p 183-8 Glynn PW. 1985b. El Nino-associated disturbance to coral reefs and post disturbance mortality by Acanthaster planci. Mar Ecol Prog Ser 26:295-300. Glynn PW. 1990. Coral mortality and disturbances to coral reefs in the tropical eastern Pacific. In: Glynn PW, Ed. Global eco- logical consequences of the 1982-83 El Nino-Southern Oscil- lation. Amsterdam: Elsevier, p 55-126. Glynn PW. 1994. State of coral reefs in the Gal?pagos Islands: natural vs. anthropogenic impacts. Mar PoUut Bull 29:131-40. Glynn PW. 2002. "An overview of the effects of the I997-I998 El Nino-Southern Oscillation on eastern Pacific coral commu- nities and coral reefs: an overview." In: Proceedings of the Ninth International Coral Reef Symposium, Bali. Editor: M, Kasim Moosa. Bali: Ministry of Environment Indonesian In- stitute of Science. Vol. 2 p II69-II74 Glynn PW, Colgan MW. 1988. "Defense of corals and enhance- ment of coral diversity by territorial damselfishes." In: Pro- ceedings of the Sixth International Coral Reef Symposium, Australia. Editor: J.H. Choat Townsville, Australia: 6th Inter- national Coral Reef Symposium, Executive Committee. Vol 2. p 157-63 Glynn PW, Colgan MW. 1992. Sporadic disturbances in fluctu- ating coral reef environments: El Ni?o and coral reef develop- ment in the eastern Pacific. Am Zool 32:707-18. Glynn PW, Colley SB. 2001. A collection of studies on the effects of the 1997-98 El Nino-Southern Oscillation event on corals and coral reefs in the eastern tropical Pacific. Bull Mar Sei 69:1-288. Glynn PW, Macintyre IG. (1977) "Growth rate and age of coral reefs on the Pacific coast of Panam?." In: Proceedings of the Third International Coral Reef Symposium, Miami. Editor: D.H. Taylor City: Miami, Florida: Rosenthal School of Marine and Atmospheric Science, University of Miami. Vol 2. p 251-9 Glynn PW, Mat? JL. (1997) "Field guide to the Pacific coral reefs of Panam?. In: Proceedings of the Eighth International Coral Reef Symposium, Panam?." Editors: H.A. Lessios and I.G. Macintyre City: Balboa republic of Ponamz Publisher: Smith- sonian Tropical Research Institute Year: 1997 Vol I. p 145-66 Glynn PW, Mat? JL, Baker AC, Calder?n MO. 2001. Coral bleaching and mortality in Panam? and Ecuador during the 1997-1998 El Nino-Southern Oscillation event: spatial/tem- poral patterns and comparisons with the 1982-1983 event. Bull Mar Sei 69:79-109. Glynn PW, Prahl H, von Guhl F. 1982. Coral reefs of Gorgona Island, Colombia, with special reference to corallivores and their infiuence on community structure and reef develop- ment. An Inst Invest Mar Punta Betin 12:185-214. Eastern Pacific Coral Reef Food Webs 367 Glynn PW, Stewart RH, McCosker JE. 1972. Pacific coral reefs of Panama: structure, distribution and predators. Geol Rund- scfiau 61:483-519. Glynn PW, Wellington GM, Birkeland C. 1979. Coral reef growtfi in tfie Gal?pagos: limitation by sea urchins. Science 203:47-9. Guzman HM. 1988a. Distribuci?n y abundancia de organismos coral?voros en los arrecifes coralinos de la Isla del Ca?o, Costa Rica. Rev Biol Trop 36(2A):191-207. Guzman HM. 1988b. Feeding activity of tfie corallivorous gas- tropod Quoyula monodonta (Blainville). Rev Biol Trop 36(2A): 209-12. Guzman HM, L?pez JD. 1991. Diet of the corallivorous pufferfish Arothron meleagris (Pisces: Tetraodontidae) at Gorgona Island, Colombia. Rev Biol Trop 39:203-6. Guzman HM, Obando VL, Cort?s J. 1987. Meiofauna associated with a Pacific coral reef in Costa Rica. Coral Reefs 6:107-12. Guzman HM, Robertson DR. 1989. Population and feeding re- sponses of the corallivorous pufferfish Arothron meleagris to coral mortality in the eastern Pacific. Mar Ecol Prog Ser 55: 121-31. Jim?nez C. 1996-97. Coral colony fragmentation by whitetip reef sharks at Coiba Island National Park, Panam?. Rev Biol Trop 44/45:698-700. Jones GP, Ferrell DJ, Sale PP. 1991. Fish pr?dation and its impact on the invertebrates of coral reefs and adjacent sediments. In: Sale PF, Ed. The ecology of fishes on coral reefs. San Diego: Academic, p 156-79. Kwiecinski B, Chial ZB. 1983. Algunos aspectos de la oceanogra- f?a del Golfo de Chiriqu?, su comparaci?n con el Golfo de Panam?. Rev Biol Trop 31:323-5. Lessios HA, Kessing BD, Robertson DR, Paulay G. 1999. Phylo- geography of the pantropical sea urchin Eucidaris in relation to land barriers and ocean currents. Evolution 53:806-17. Macintyre, IG, Glynn, PW, Cortes, J (1992) "Holocene reef his- tory in the eastern pacific: Mainland Costa rica Cano Island, and Gal?pogos Islands." In: Proceedings of Seventh Interna- tional Coral Reef Symposium, Guam Opitz S. (1996) "Trophic interactions in Caribbean coral reefs." Technical Report 43. Makati City (Philippines): International Center for Living Aquatic Resources Management. 341 p Dramas N, FA. (1979) "Estudio ecol?gico del Jenneria pustulata (MoUusca, Gastropoda)." Thesis, Escuela de Biolog?, Univer- sidad de Panam?. 100 p Paulay G. 1997. Diversity and distribution of reef organisms. In: Birkeland C, Ed. Life and death of coral reefs. New York: Chapman and Hall, p 298-353. Reaka-Kudla ML, Feingold JS, Glynn PW. 1997. Experimental studies of rapid bioerosion of coral reefs in the Gal?pagos Islands. Coral Reefs 15:101-7. Renner JA. 1963. Sea surface temperature monthly average and anomaly charts for the eastern tropical Pacific Ocean, 1947- 58. Spec Sei Rep US Fish Wildl Serv (Fish) 442:1-57. Reyes Bonilla H, Calder?n-Aguilera LE. 1999. Population den- sity, distribution and consumption rates of three corallivores at Cabo Pulmo reef. Gulf of California, Mexico. P.S.Z.N. Mar Ecol 20:347-57. Reyes Bonilla H, Carriquiry JD, Leyte-Morales GE, Cupul-Ma- ga?a AL. 2002. Effects of the El Nino-Southern Oscillation and the anti-El Ni?o event (1997-1999) on coral reefs of the western coast of M?xico. Coral Reefs 21. 368-372 Stimson J. 1990. Stimulation of fat body production in the coral Pocillopora by mutualistic crabs of the genus Trapezia. Mar Biol 106:211-8. US Department of Commerce. 1952-96, Tide tables. West Coast of North and South America. Coast and Geodetic Survey (ESSA). Washington (DC): US Government Printing Office. Wellington GM. 1980. Reversal of digestive interactions between Pacific reef corals: mediation by sweeper tentacles. Oecologia (Berl) 47:340-3. Wellington GM. 1982a. Depth zonation of corals in the Gulf of Panam?: control and facilitation by resident reef fishes. Ecol Monogr 52:223-41. Wellington GM. 1982b. An experimental analysis of the effects of light and Zooplankton on coral zonation. Oecologia (Berl) 52:311-20.