Food and Shelter as Determinants of Food Choice by an Herbivorous Marine Amphipod

Because food and habitat are closely linked for small herbivores that live on plants, food choice in the field may be constrained by the need to choose plants that provide safe living sites. We investigated the importance of food value and refuge value in determining the plant utilization patterns of the herbivorous marine amphipod Ampithoe longimana. When offered a choice of five common seaweeds, this amphipod fed most readily on Dictyota and Hypnea and less readily on Sargassum, Chondria, and Calonitophyllum. Rates of feeding on the different seaweeds were unrelated to seaweed gross morphology, toughness, nitrogen, or protein content. When cultured on each of these seaweeds in the laboratory, amphipod survivorship was high on Dictyota (82%), intermediate (35 and 18%, respectively) on Sargassum and Hypnea, and low (0%) on the other seaweeds. Survivorship on the different diets was strongly correlated (r = 0.930) with algal protein content; however, neither protein content nor amphipod performance of the different diets was significantly related to feeding rates on those diets. Additionally, amphipods from the three seaweed species that produced some survivors did not differ in growth rate, fecundity, egg size, or age at first ovulation. Variance in survivorship, and related measures, among sibling groups of amphipods suggested that this amphipod population possessed heritable variation for performance on the different seaweed species. In the field, abundance of A. longimana on the different species of algae was more clearly related to the preference of omnivorous fishes for these algae than to feeding rates of the amphipods when given those algae in the laboratory.A. longimana was more abundant on Dictyota and Sargassum (both unpalatable to omnivorous fishes), than on Hypnea, Chondria, and Calonitophyllum (all of which are palatable to fishes). During the season when omnivorous fishes were abundant, density of A. longimana increased on Dictyota, which is chemically defended from fishes, but decreased or remained unchanged on the seaweeds that are more palatable to fishes. Competition with other amphipods as a group did not appear to explain the distribution of A. longimana among seaweeds, since there were no negative correlations between A. longimana abundance and total amphipod abundance in any month. The lack of any consistent relationship between host—plant use in the field and either feeding preference or diet value, as measured by survivorship and reproduction, suggests that host—plant use by A. longimana may be strongly constrained by requirements for shelter from predation.