The case for marine reserves is strengthening, and both deterministic and stochastic calculations show that fisheries management using reserves may achieve harvests comparable with management without reserves. Thus, depending upon the metric used, reserves need not disadvantage harvest. Reserves provide a buffer that increases the chances of sustainability of the stock, and thus the fishery. In this paper, I develop methods (deterministic and stochastic) that allow one to determine how much habitat needs to be set aside as reserve, once societal decisions concerning the goals of reserves are made. The answer to the question: “how much habitat needs to be allocated to reserves” is not a simple single number. Rather, it is a procedure that can be employed once biological, operational and social information are provided. The methods also apply to reserves used to aid stock recovery.