of the stock is exposed to fishing, and protection is lower. This “dispersal imbalance” (i.e., more fish leave the reserve than enter) will intensify if fishing effort concentrates near the reserve boundary (Walters, 2000). Simulation models often indicate that a large fraction (20–50%) of productive fishing areas may have to be designated as reserves to provide the desired level of insurance (Roberts, 2000).
When the mobility of adults is high, as in many pelagic and migratory fish species, reserves have often been discounted as an effective management tool. Whereas many coral reef fish have small territories as adults and may disperse during their planktonic larval stage, numerous fish species migrate hundreds or even thousands of kilometers annually (Harden Jones, 1968). Many high-value fish species, including cod (Gadus morhua) and herring (Clupea harengus) migrate long distances and so would obtain only intermittent protection from reserves as they pass through them. However, reserves on spawning grounds or in nursery areas for such species can offer protection and may be a management option. Even for highly migratory species such as swordfish (Xiphias gladius) or tunas, MPAs that protect nursery areas or vulnerable population bottlenecks may be effective as management tools.
Although it is generally assumed that all individuals of a migratory species migrate, recent evidence suggests that many individual fish cover relatively short distances. For example, most of 11,000 galjoen (Dichitius capensis) tagged in a marine reserve in South Africa remained within a few kilometers of their tagging sites, while about 18% dispersed tens to hundreds of kilometers (Attwood and Bennett, 1994). Hence, reserves may protect the less mobile individuals of these migratory species.
Fishing also affects habitat and non-commercial species. However, most studies of reserve performance evaluate only a narrow range of taxa, focusing on fish assemblages. Studies of East African coral reefs demonstrate that fishing has broader impacts on the ecosystem, including changes in biogenic habitat. In this case, the fishery removed keystone predators of sea urchins, leading to urchin population explosions. As urchin grazing intensified, there was increased bioerosion of coral with reductions in coral cover (McClanahan and Shafir, 1990). In addition, sea urchins out-competed herbivorous fish for algae, leading to declines in the fishery for these herbivorous species (McClanahan et al., 1994).
In Jamaica, the ultimate cause of reef degradation following a mass mortality of sea urchins and two hurricanes was attributed to overfishing of herbivorous fish. After the removal of herbivorous fish, sea urchins were the primary