Slender baardmanUmbrina robinsoni are an important component of recreational shore-angler and spearfisher catches along the eastern seaboard of South Africa. Stocks of U. robinsoni at three sites -False Bay, Stil Bay and the KwaZulu-Natal (KZN) coast -were modelled using a per-recruit approach. Total (Z) and fishing (F) mortality rates were estimated by catch-curve analyses using measures of individual size (length or weight) recorded by researchers, divers (log books) or during spearfishing competitions. Based on estimates of F during the period 2001-2003, spawner biomass per-recruit ratios were estimated to be either at or below the 25% threshold (False Bay SB/R = 21% SB/R F=0 , Stil Bay SB/R = 25% SB/R F=0 , and KZN SB/R = 21% SB/R F=0 ), suggesting that rates of F were too high. Reductions in F necessary to achieve target fishing mortality levels (F SB40 ) at the current minimum size limit (l c 40 cm total length) were 51% for Stil Bay and the KZN coast and 57% for False Bay. Based on the bag frequencies from 927 diver outings in KZN (1989KZN ( -2003, a reduction in bag limit from the current five to two fish is predicted to reduce F in this region by approximately 25%. Increasing the l c to 50 cm is predicted to increase SB/R ratios to 36% SB/R F=0 in False Bay, 43% SB/R F=0 in KZN and 52% SB/R F=0 in Stil Bay, at current levels of F. Owing to the philopatric nature of U. robinsoni and the consequent existence of temporary refugia, catch curves are likely to underestimate fishing mortality. The reductions in F estimated to attain the target reference points are therefore probably conservative.