Understanding the dynamics of open marine populations is difficult. Ecological processes may vary with the spatial structure of the habitat, and this variation may subsequently affect demographic rates. In a series of observational and experimental studies in the Bahamas, I examined the roles of emigration, mortality, and predation in the local population dynamics of juvenile coral-reef fishes. First, I documented mortality and emigration rates in populations of bluehead and yellowhead wrasse. Assuming that all losses were due solely to mortality would have significantly underestimated survivorship for both species on patch reefs, and for yellowheads on continuous reefs. Mortality differed between species, but emigration did not differ between species or reef types. Mortality of blueheads was density-dependent with respect to both conspecific density and total wrasse density on continuous reefs. In contrast, mortality of yellowheads varied inversely with the density of blueheads on patch reefs. Emigration rates varied inversely with distance to the nearest reef inhabited by conspecifics. In subsequent experiments, I manipulated densities of yellowhead wrasse and beaugregory damselfish, and determined that the relationship between density and mortality varied with reef spatial structure. On natural reefs, mortality rates of the wrasse were highly variable among reefs. On artificial reefs, mortality rates of both species were density-dependent on spatially isolated reefs, yet high and density-independent on aggregated reefs. Heterogeneity in the spatial structure of natural reefs likely caused variation in predation risk that resulted in high variability in mortality rates compared to artificial reefs. A final experiment demonstrated that a single resident predator caused substantial mortality of the damselfish, regardless of reef spacing. Patterns suggested that resident predators caused density-dependent mortality in their prey through a type 3 functional response on all reefs, but on aggregated reefs this density dependence was overwhelmed by high, density-independent mortality caused by transient predators. These results (1) suggest post-settlement movement should be better documented in reef-fish experiments, (2) demonstrate that the role of early post-settlement processes, such as predation, can be modified by the spatial structure of the habitat, and (3) have ramifications for the implementation of marine reserves. / Graduation date: 2004