Date Awarded


Document Type


Degree Name

Doctor of Philosophy (Ph.D.)


Virginia Institute of Marine Science


Laboratory experiments determined the effects of two levels of habitat complexity upon pipefish (Syngnathus fuscus) foraging for amphipods. Habitats were composed of equal densities of either narrow (low complexity) or wide (high complexity) leafed artificial seagrass. The response to habitat, as measured by rate of encounter with amphipods, probability of attack after encounter, probability of success after attack, and overall rate of amphipod consumption, was determined for combinations of two fish size classes and three amphipod size classes. Small fish did not experience visually inhibitive effects in either habitat, while large fish had their visual fields impinged upon in the wide leaf habitat and encountered fewer amphipods. There was a general trend for encounter rate to increase with amphipod size. Large fish attack probability was positively related to amphipod size in the narrow leaf habitat, but negatively related to amphipod size in the wide leaf habitat. Small fish attack probability was negatively related to amphipod size in both habitats. Success was negatively related to a ratio of prey size to fish size, and showed no overall effect of habitat. Pipefish have flexible behaviors, allowing them to minimize unsuccessful attacks. Due to their position in the structure of vegetation, amphipods have a distribution of vulnerabilities; a criterion by which pipefish select prey. Size-selective predation on gammarid amphipods by pipefish was examined utilizing simulation modeling and laboratory experimentation. Three computer simulation models were developed: (1) a mechanistic model based on empirically derived size-dependent mechanisms of pipefish-amphipod interaction, (2) an optimal diet breadth model in which the rate of energy intake is maximized, and (3) an optimal diet breadth model where switching from energy maximization to time minimization occurs as consumption becomes limited by gastric processing (i.e. satiation). None of these models successfully accounted for the observed pattern of prey size selection. Pipefish concentrated their feeding upon smaller, energetically more profitable amphipods, in excess of what was predicted by either the mechanistic or optimal diet breadth models. This pattern of selection was evident through out 4 hour feeding bouts, indicating that diet breadth compression did not occur.



© The Author