Two extreme types of fisheries are set up for analysis. In type I natural mortality and recruitment are negligible while fishing is in progress; in type II natural mortality, recruitment and fishing mortality all occur throughout the year at a uniform rate. With either type, the catch per unit effort is proportional to the average population on handwhile fishing is in progress. This statistic is less useful than it may appear, because, in cases of type I, recruitment occurs between the fishing seasons, and in cases of type II, the slow process of removal (or replacement) of accumulated stock following a change in fishing effort makes it difficult to determine the significance of any change in catch per unit effort.A more useful statistic is the rate of exploitation (a fish's annual expectation of death by capture). A theoretical relationship between rate of exploitation and gear in use is here developed for type I fisheries, and is shown to apply fairly well to many cases of type II—so that if rate of exploitation for one year be determined, that for other years can be calculated. For certain special cases of type I fisheries, several methods are proposed by which rate of exploitation might be estimated from ordinary statistics of catch and effort. For other cases of type I, and all cases of type II, an application of Petersen's tagging technique, or data from other outside sources, appear almost essential to determination of the rate of exploitation.