Rational management of fisheries in order to optimize harvests requires a constant input of data on the state of the Fish stocks, and in particular, of their mortality rates. Such data are usually obtained by laborious gathering of catch and effort statistics, accompanied by separate investigations designed to produce estimates of the biological and fishery parameters which govern the productivity of the stocks. In most tropical countries, the diversity and complexity of multi-species, multi-gear ertissnal fisheries makes conventional stock assessment very difficult to execute and often also prohibitively expensive. However, over the past decade, the development of length-frequency based stock assessment methodologies has rendered many of the traditional data acquisition systems outdated or redundant. In particular, detailed estimates of catch and effort are no longer essential to the fishery manager, even though they remain highly desirable. In essence, all parameters necessary for a basic assessment of the state of a fish stock can now be derived from routine collection of length-frequency data.