122 



COLIN 



organisms may prey directly on calcifying organisms and in 

 the proces% often damage the skeleton. 



A few species of fishes vigorously attack coral skele- 

 tons, biting off and ingesting the tips of branched species. 

 Randall (1974) observed the pufferfish, Arothoron nigro- 

 punctatus, feeding heavily (85 to 100% of gut contents) on 

 corals, particularly Acropora and Montipora. Hiatt and 

 Strasburg (1960) found corals in the guts of nine plectog- 

 nath fishes (two triggerfishes, three filefishes, three puffers, 

 and one sharpnose puffer). Most, but not all individuals of 

 any species, had ingested branched coral tips in various 

 amounts. Although none of these fishes are obligate coral 

 predators, many contain coral tips in such quantity that 

 these must constitute a regular part of their diet at 

 Enewetak (Randall, 1974). 



Large portions of coral skeleton will, on occasion, have 

 the ends of the branches removed, often with piles of coral 

 fragments left in the depression. This is seen in Pontes rus 

 at Enewetak (Fig. 21) and in other species. It is assumed 

 this phenomenon results from the activities of fishes which 

 feed on coral branches, but the feeding by some of these 

 fishes is seldom observed. 



Other coral-feeding fishes tend to eat only the polyps, 

 leaving the skeleton essentially intact. In such cases, the 

 polyp normally regenerates. A number of butterflyfishes 

 (Chaetodontidae) and damselfishes of the genus 

 Plectrogli^phidodon feed on corals in this matter (Motta, 

 1980; Randall, 1974; Reese, 1973, 1975, 1977) and are 

 discussed elsewhere. Randall (1974) notes also that the 

 blenny Ecsenius bicolor at Enewetak has been observed 

 feeding on Acropora. 



Some herbivorous fishes occasionally scrape at the sur- 

 face of living corals doing more damage than the chaeto- 

 dontids. Scarids produced a characteristic scrape mark on 

 corals with an elongate furrow, often with a slight ridge 

 along its midline where the two sides of the beak fuse. 

 Hiatt and Strasburg (1960) found some species of scarids 

 at Enewetak had fed on corals. Randall (1974) has 

 reviewed the question of parrot fish grazing on live corals 

 and discusses an apparent disparity between published 

 data on coral feeding by scarids at Heron Island, Great 

 Barrier Reef, and Hiatt and Strasburg's (1960) information. 

 He found no obvious reason for the differences observed 

 but suggested that local coral cover may influence how 

 much coral is ingested by parrot fishes. Although some 

 scarids do graze live corals, the impact of this behavior is 

 probably minor compared to the effect on sediment pro- 

 duction and deposition. 



Randall (1974), Ogden (1977), and others have docu- 

 mented the role of scarids in sediment production. The 

 rasping of rock or coral for its algal film is the first step. 

 This material is then ground to a fine consistency by the 

 pharyngeal mill of the parrot fish, passed through the gut, 

 and eventually expelled. The rain of sedimentary material 

 shed when parrot fishes defecate is impressive, and the 

 amount of sediment produced from hard substrates by this 

 mechanism is enormous. 



Also important in sediment production are fishes which 

 reduce the hard parts of invertebrates (mollusc shells, echi- 

 noid tests and spines, crustaceans, etc.) to bits. Randall 

 (1974) reports that plectognaths with their fused or but- 

 tressed teeth, lethrinids with molariform teeth, labrids with 

 pharyngeal teeth, and dasyatid and myliobatid rays with 

 plate-like jaws are well adapted for this purpose. 



Massive corals at Enewetak are attacked by a number 

 of biological agents. Although seldom visible, these agents 

 weaken the skeleton to the point that physical factors can 

 break the colony loose or cause it to crumble. Highsmith 

 (1981a) reports that clinoid sponges accounted for 70 to 

 80% of skeletal damage in various massive corals from 

 Enewetak. They did not burrow deeply into the skeleton, 

 only a few millimeters, but extended interconnected 

 chambers laterally beneath dead surfaces of the coral 

 colony. Highsmith (1981a) reported that 65 to 95% of the 

 boring was within the "dead area" of skeletons. In a mas- 

 sive coral this "dead area" includes the area around the 

 basal attachment and dead spots on the colony surface. 

 Similarly, these dead areas are heavily eroded by grazing 

 organisms. When exposed to light or scraped (as when 

 overlying skeletal material is removed), clinoid sponges 

 engage in rapid burrowing (Ruetzler, 1975) Heavy grazing 

 pressure, combined with this response, may produce rapid 

 erosion rates at basal attachments. 



Highsmith (1981a) points out that skeletal weakening 

 at the base, combined with storm-induced water motion, 

 may not be sufficient to dislodge most massive colonies. 

 However, coral rubble on the bottom can be put into 

 motion by storm waves and, to a point, may be the most 

 significant force in breaking heads loose. Eventually 

 though, "as massive corals grow, they become more sus- 

 ceptible to breakage by storm currents and less susceptible 

 to breakage by suspended rubble or to biocrosion detach- 

 ment." 



The alpheid shrimps occurring in deep grooves on 

 Goniastrea retiformis apparently form the grooves, not by 

 boring or erosion, but by preventing growth of coral in 

 that area while the remainder of the colony continues to 

 increase in size (Fig. 22). These grooves, though, provide 

 dead areas which penetrate deeply into the G. retiforrr}is 

 head and are penetrated by boring organisms (Highsmith, 

 1981a). 



Highsmith (1981b) suggested that bioerosional damage 

 to corals is positively correlated with increasing skeletal 

 density. Five species of Enewetak corals {Ouloph\^lha 

 crispa. Fauia pallida, Goniastrea retiformis. Pavona clauus, 

 and Pontes lutea) had a positive correlation between 

 bioerosion and density. This correlation did not correspond 

 to differences in growth rates. The slowest growing 

 species, F. pallida, was the least bored. 



Among molluscs, the boring bivalve Lithophaga curta 

 preferentially colonized the coral Montipora berrvi 

 (Highsmith, 1980). Boring bivalves in general have thin, 

 weak shells and, if exposed, are easily eaten by fish preda- 

 tors. Highsmith (1981a), for example, reported that the 



