Extinction is conceivable for some species, including aquarium-fish stocks of high value and limited geographical range, but a biological basis exists for management of such small fishes. Immediate data needs are much greater for management of stocks of larger species which most reef fishing targets. The strengths and limitations of underwater visual census are now better recognized, and there is improved capacity for fish ageing, especially through otolith analysis, as a basis for growth and mortality assessments. Information technology has grown rapidly in scope, and modelling approaches offer a better foundation now for incorporating bio logical data into analytical approaches to sustainability. Reef fishery dynamics may widely be simpler than implied by their ‘multispecies‘ con dition, and in spite of their assumptions, surplus production models have proved useful in stock assessment. Empirical and exploratory approaches to sustainability, however, are more desirable than ever, as uncertainty about long-term ecosystem effects of fishing increases. Work of an experi mental nature urgently needs also to be directed at rehabilitation techni ques, especially marine fishery reserves. Successful management is, however, unlikely without local community involvement. Understandingthe scope of such participation demands a greater social science input to reef fishery studies; here as elsewhere there are encouraging trends in recent studies.