extensively fed upon by butterf lyfishes and balistid triggerfishes when 

 exposed. Using a similar experimental design, Wellington (1982) found that 

 Pocillopora colonies exposed to predation by mainly tetraodontid pufferfishes off 

 the Pacific coast of Panama suffered strongly reduced growth rates relative to 

 protected controls. These studies suggest that Pocillopora colonies may be 

 generally rare where fish coral livores are abundant. 



Perhaps more important to coral distribution and abundance than direct 

 predation by fishes are the indirect effects of certain damselfishes (Pomacentri- 

 dae). These fish defend permanent individual territories about 1 m in diameter, 

 resulting in locally decreased grazing by other fishes, and thus increased algal 

 growth as distinct mats used as food sources and nesting sites (see next 

 section). Field observations have shown that damselfish kill coral either 

 directly by removing polyps, thus providing open substrate for their algal mats 

 (Kaufman 1977, Wellington 1982), or indirectly by the reduced grazing within 

 territories allowing algae to smother corals (Potts 1977). In either case, 

 bioerosion of corals may be accelerated within territories, apparently because 

 the algal mats provide refuges and/or food sources for settling boring organisms, 

 which subsequently reduce the ability of the coral to withstand wave shock (Risk 

 and Sammarco 1982). Thus, the algal mats within damselfish territories are gen- 

 erally detrimental to coral colonies (see also Lobel 1980). However, small 

 corals located within the periphery of a territory but still outside the algal 

 mat can be protected from coral livores by the resident damselfish (Wellington 

 1982). 



How damselfish territoriality affects the recruitment of coral larvae is 

 unclear. Previous work has shown that algal mats inhibit settlement by corals 

 (e.g., Birkeland 1977). However, Sammarco and Carlton (1981) reported enhanced 

 recruitment of mostly Acropora on substrates placed inside damselfish territories 

 on the Great Barrier Reef for four months relative to caged substrates or those 

 exposed to grazers outside territories. Unfortunately, the ultimate fates of 

 these recruits were not followed. Using a similar experimental design in Hawaii, 

 Hixon and Brostoff (in prep.) found the recruitment of Pocillopora over the 

 course of a year on caged surfaces to be greater than that on exposed surfaces 

 either inside or outside damselfish territories. However, there was indication 

 of low survival of these recruits due to smothering by algae. Moreover, in a 

 laboratory microcosm experiment at the same site, Brock (1979) found an inverse 

 relationship between the density of grazing scarid parrotfishes and coral 

 recruitment on exposed tank walls. Bak and Engle (1979) have also attributed the 

 high mortality rates of juvenile corals to grazing by parrotfishes, which are 

 typically excluded from damselfish territories. It appears that on open 

 surfaces, coral recruits may initially experience enhanced survival where they 

 are protected from extensive fish grazing, either inside cages or within 

 damselfish territories, but will eventually be overgrown by algae. Those larvae 

 settling where exposed to fish grazing may initially suffer high mortality, being 

 consumed along with the fishes' other prey, but some individuals will eventually 

 reach an invulnerable size class where they are both immune to incidental 

 predation and freed from competition with algae (Birkeland 1977). 



^lery few data have been reported on how fishes alter competitive interactions 

 between corals. Sammarco and Williams (1982) have hypothesized that the algal 

 mats of damselfish territories may act as refuges for some rarer corals, thus 

 increasing the local evenness of relative abundances among coral species. By far 

 the most extensive published study of the effects of fishes on coral community 

 structure is that of Wellington (1982), who worked on a fringing reef off the 

 Pacific coast of Panama. In this system, branching Pocillopora spp. dominated 

 shallow areas (0-6 m depth), while the massive Pavona gigantea dominated deeper 

 areas (6-10 m depth). Using a clever series of experiments and observations, 



80 



