Marine ecosystems can be exposed to natural and anthropogenic disturbances that can lead to ecological failures. Marine reserves have been lately suggested to protect marine populations and communities that have been affected by habitat destruction and harvest. This research evaluates the potential role of two marine reserves established in Oregon in 1967 (Whale Cove) and 1993 (Gregory Point). The red sea urchin (Strongylocentrotus franciscanus) was selected as indicator of population recovery since it is the only species that is commercially harvested. Changes in density, biomass, average size, size structure, growth and mortality rates were evaluated through time to assess population recovery. These parameters were also compared between reserves and adjacent exploited areas to evaluate the effect of exploitation. Results from Whale Cove (old reserve) indicate that the population in this area is fully recovered. On the contrary, the population in Gregory Point (new reserve) showed signs of recovery after six years of being protected. The importance of red urchins as source populations to provide larvae to adjacent areas was explored by the analysis of drifter's trajectories. Both reserves might be connected in a network where larvae produced in Whale Cove will provide recruits to Gregory Point and adjacent exploited areas, as well as populations in northern California. Gregory Point releases larvae that become recruits for Whale Cove only when spawning takes place in winter, otherwise larvae travel to central California. No clear trends were found in growth and mortality rates between reserves and non-reserves; differences were more related with food availability, competitors, and age specific mortality. We applied qualitative simulations to characterize and differentiate the community network inside reserves and exploited areas. Results suggest that communities from a particular site can be represented by a set of alternative models with consistent species interactions. Differences in predator-prey interactions as well as non-predatory relationships (interference competition, mutualism, amensalism) were found among sites. Each set of models represents a hypothesis of community organization that agreed with natural history information. Alternative models suggest that kelp forest communities are dynamic and can shift from one network configuration to another providing a buffer against a variable environment. / Graduation date: 2002