This study empirically examines predictions concerning boundary permeability as a function of patch contrast (i.e., the magnitude of difference in measures across patch interfaces), light level, and faunal mobility and density, as well as relative rates of emigration and immigration. I assessed permeability using invertebrate transfer in a macroalgal-seagrass-sand landscape, with particular emphasis on the caridean shrimp Thor floridanus. Proportional emigration was estimated by staining algal fauna in situ using neutral red (a live immersion stain) and then later collecting all animals in the algal clumps and in the surrounding sand or seagrass. Permeability was in part a function of boundary contrast in this milieu; macroalgal patches were more of a closed system in sand than in seagrass. Contrary to predictions, proportional emigration across both the algal-seagrass and algal-sand boundaries was inversely related to faunal density. Another unexpected result was that the rate of change in emigration with increasing density was unaffected by differential contrast among habitat types. The observed enhanced permeability for more motile fauna was consistent with predictions. A related finding was that permeability was greater at night than during the day, probably as a function of increased nocturnal vagility of benthic fauna. Emigration was not in equilibrium with immigration at the temporal scale of this study (hours) for the majority of species tested. Emigration-immigration disequilibria should be most likely when the temporal scale is small, and in such cases caution should be exercised when extrapolating from one transfer parameter to the other.