By-catch was surveyed in the commercial prawn trawl fleets of the Clarence River system, the largest estuarine prawn fishery in New South Wales, Australia. The catch was censussed in all tows during replicate fishing trips in each month in the 1989-90, 1990-91 and 1991-92 prawn trawl seasons in the two areas of this fishery: the Clarence River and Lake Woolooweyah. Significant species-specific variabilities in abundances were detected at a variety of spatial and temporal scales. Despite comparatively small ratios of by-catch to catch of prawns (annual estimates in each area of between 0.13:1 and 0.45:1), hundreds of thousands of small recreationally and commercially important finfish were caught and discarded by the fleet. The results are discussed in terms of (i) their contrast to common generalisations about by-catch in prawn trawl fisheries, (ii) the accuracy and precision of the methods used to quantify and compare by-catches, (iii) the consequences for conflicts between prawn trawl fisheries and other commercial and recreational fisheries that target by-catch species, and (iv) the various strategies available to fisheries managers to resolve potential conflicts.