The role of a marine protected area in enhancing local fisheries, through the emigration or spillover of exploitable fishes, was studied in a coral reef park (Mombasa Marine Park, Kenya) and fishery over a seven-year period during a time when the park's border changed and pull seines were eliminated. We measured catches before and after the park's establishment and during the management changes and compared these catches with the unmanaged side of the park. Additionally, we placed baited traps on both sides of the park over a full tidal cycle which allowed us to measure the spillover from the park compared to the deeper, rougher, and less fished reef edge. The total wet mass of catches per trap, the mean size of the trapped fish, and the number of fish species caught per trap declined as a function of the distance away from the park edge on both the southern and northern sides. However, this relationship was truncated on the unmanaged side which also had smaller catches, smaller fish, and fewer species than the managed side. Trap fishers on the managed side adapted to the spillover by increasing the traps per fisher, which had the effect of reducing the catch per trap. Tides and reef morphology also appeared to interact and influenced catches, but we found no relationships between catches and benthic substratum cover, which was usually dominated by seagrass and sand. Spillover from the deeper reef edge was evident for the managed but not the unmanaged side of the park, but may be due to differences in reef morphology interacting with tidal patterns rather than management. On the managed side, the park significantly increased the catch per fisher and catch per area by >50%, but even after the park's size was reduced, the total catch was reduced by ∼30%. The reduced park was still ∼50% of the total area. Consequently, the catch per area increase was insufficient to compensate for the lost area over this early period of the park's establishment. Spillover was greatest for the dominant fisheries species. These were moderately vagile species in the rabbitfish (Siganidae; herbivores), emperors (Lethrinidae; carnivores), and surgeonfish (Acanthuridae; herbivores) families, which had instantaneous emigration rates from the park to the reserve fishing ground of ∼0.5. Our field survey, combined with previous modeling studies, based on adult emigration rates from marine reserves, suggests that tropical fisheries dominated by rabbitfish, emperors, and surgeonfish should be enhanced by closed areas of ∼10–15% of the total area. The optimal protected area may increase if larval export is important, but the predicted response should not be measurable for >10 years, beyond the length of our study, as breeding stock develop inside protected areas.