Patterns of settlement of larvae and population dynamics of juveniles are poorly known for coral reef fishes. During 1987 to 1989, I studied these phenomena in the domino damselfish (Dascyllus albisella), a species endemic to the Hawaiian Islands. Larvae settle onto branching coral heads as new recruits (10-15 mm in length), usually with conspecific groups, and remain on the coral heads through juvenile life until maturity (70 mm total length). By conducting experiments on natural patch reefs and on an artificially distributed grid of coral heads, I found that most larvae settle at night, and that they settle preferentially on corals supporting large conspecific groups compared to small groups or empty corals. Within a group, juveniles form a linear dominance hierarchy based on fish size; aggressive interactions are mainly directed by larger fish towards smaller fish. Tagging studies demonstrated that growth was retarded in larger groups and for fish of low social status, but that survival, especially of new recruits, was enhanced in larger groups. Therefore, I identified both a growth cost and a survival benefit to group living. I derived a measure of net benefit of group living by combining size-specific growth and survival data into an estimate of the probability of reaching mature size. This estimate increased with group size in 1988 but not in 1987. I developed a simulation model which used my field data on settlement rate, settlement preferences, and juvenile growth and survival to predict demography of juvenile groups. The model successfully predicted seasonal fluctuations in mean group size, and estimated the number of fish maturing in 1987 and 1988, as a function of settlement rate and preferences and of juvenile growth and survival. Numbers maturing were directly related to settlement rate in both years, except at high rates in 1987, suggesting that primary recruitment limitation of adult numbers could be occurring. Settlement preferences also influenced numbers maturing. At all settlement rates, numbers maturing differed between years, suggesting that secondary recruitment limitation of adult numbers may also occur. / Graduation date: 1991