Specificity of larval settlement of the Caribbean Orange Icing Sponge,Mycale laevis

The Caribbean sponge Mycale laevis is often found growing in close proximity to living scleractinian corals. This commonly observed sponge–coral association has been considered a mutualism, with the coral providing substratum for the sponge, and the sponge protecting the coral skeleton from boring organisms. We examined the specificity of sponge recruitment to live corals, expecting a positive and specific settlement response if a mutualism exists. Benthic surveys conducted off Key Largo, Florida, and Bocas del Toro, Panama, revealed that individuals of M. laevis grew on substrata that included dead coral and other species of sponges. Selectivity analysis indicated that at three of the four survey sites, M. laevis was not randomly distributed, but associated with live corals more frequently than expected from proportional coral cover. However, settlement assays demonstrated that larvae of M. laevis did not preferentially respond to the presence of live coral. We have previously demonstrated that adults of M. laevis are chemically undefended and readily eaten by spongivorous fishes unless protected by adjacent substrata such as live corals. In overfished areas, where spongivore density is low, the sponge is not selectively distributed near corals. Initial results of settlement experiments with different substrata suggested that larvae of M. laevis responded positively to the presence of the chemically defended sponge Amphimedon compressa, perhaps indicating an associational defense. Further experiments revealed that larvae were reacting to artificially high concentrations of exudates from cut surfaces of Am. compressa; settlement was not enhanced in response to healed pieces of Am. compressa. In addition, the larvae of M. laevis did not selectively respond to live coral or to chemically defended heterospecifics. These results indicate that the commonly observed proximity of M. laevis to live corals is not driven by larval settlement behavior, but instead by post‐settlement mortality due to predation.