Foraging by the herbivorous parrotfish Sparisoma radians

The foraging behavior of the bucktooth parrotfish Sparisoma radians was studied in seagrass beds off St. Croix, US Virgin Islands and in laboratory preference tests. Thalassia testudinum was the dominant item in the field diet with the epiphytized distal portion of the blades most favored. Other seagrasses, Syringodium filiforme and Halodule wrightii, were taken in relation to their abundance. Several algae were also eaten, particularly Halimeda spp. and Penicillus spp. Although abundances of the algae varied, the fish maintained a steady mixture of plant species in their diet. Laboratory feeding tests for various plants presented in pairs showed clear preferences in nearly all cases. The preference hierarchy was (1) T. testudinum with epiphytes, (2) H. wrightii, (3) T. testudinum without epiphytes, (4) S. filiforme, and the algae, (5) Dictyota divaricata, (6) Enteromorpha flexuosa, (7) Caulerpa mexicana, (8) Halimeda incrassata and (9) Penicillis pyriformis. Preference did not vary significantly with satiation. The catch per unit effort (kilocalories absorbed per bite) was calculated for each plant, using (a) absorption values calculated for S. radians fed the different plants, (b) the calorific values for tested plants, and (c) an estimate of the amount of material taken per bite. The ranking of catch per unit effort closely paralleled the preference hierarchy with the exception of C. mexicana which has a toxin. Fish fed diets of single plants, mixed plant diet, and starved controls showed differential survival which paralleled the preference hierarchy, the most preferred plants leading to longest survival. Comparison of laboratory results with field feeding behavior shows that inclusion of plants in the diet is not related directly to preference rank, availability or survival value, but that the fish deliberately eat a variety of plants presumably which maintains balanced diet. This result indicates that models of optimal foraging for herbivores should include nutrient constraints and avoidance of toxins in order to predict accurately the behavior of an animal in the field.