The threespot damselfish, Stegastes planifrons, maintains individual territories that are clustered on coral patch reefs. My objective was to understand the effects of territory clustering on behavior and fitness. Fish with territories in the center of a cluster had (relative to edge fish): higher mating success (number of eggs), higher aggressive chase rates with conspecifics, lower chase rates to heterospecifics, lower overall chase rates, lower grazing rates by intruders, and smaller territories. Feeding rate, survivorship, and age at maturity did not vary with territory position. Therefore, central fish appeared to have higher fitness, which was probably related to the lower energetic costs of territory defense there. Center and edge territories differed in habitat complexity, and the density of potential algal competitors, egg predators, and various food and invertebrate species. These microhabitat features could provide different quality shelter, nest or feeding sites and thus might explain the positional differences in fitness. An experiment in which I changed the position of treatment fish from the center to the edge of a cluster, without altering microhabitat, showed that position per Se, and not microhabitat variation, caused the center-edge differences. Vacated space in the center of a cluster was fought over more vigorously and reoccuppied sooner than similar space on the edge. Settlement to one of two depopulated clusters was preferentially to the cluster center. These data indicated that threespots compete for the more desirable central positions. Therefore, these populations can be considered simultaneously recruitment limited (in terms of local population size) and resource limited (in terms of local reproductive output and perhaps global population size). Aggressive chases with conspecifics were lower on the cluster edge than at any distance toward the center, while chases to heterospecifics had the opposite pattern. The results of chases with conspecifics did not fit the predictions of the model by Stamps et al. (1987) . This discrepancy may be a result of habituation between territorial neighbors. / Graduation date: 1995