Mussels play an important supplementary role in the diet of local communities on the Transkei coast in the Eastern Cape province of South Africa. The exploitation of mussels date back to about 1350 years ago, but in the last 3 decades, exploitation of the brown mussel Perna perna has become unsustainable with mussels collected as small as 30-40mm when they are only just sexually mature. Dye and Dyantyi (2002) developed a technique to rehabilitate areas denuded of adult mussels. The government sponsored Mussel Rehabilitation Project (MRP) to use this technique but only some sites have been successfully rehabilitated, reaching c. 80 % cover within a year whilst others only reach about 5%. At an unexploited site (Riet River), I tested the effects of mussel size and wave strength on the effectiveness of the rehabilitation technique, hypothesizing that different size classes may respond differently due to differences in their energy allocation (growth vs reproduction), while wave action determines food supply. Small (1-2cm) and large mussels (3-4cm) were deployed for rehabilitation at 2 exposed and 2 sheltered sites, separated by 100s m. A similar study was repeated in Coffee Bay where shores are exploited. Six sites were selected, 3 sites that had been successfully rehabilitated and 3 that were unsuccessful according to the MRP. Again, two size classes were used but these differed from the first experiment. Mussels of 3-4cm size were now rated as small and 5-6cm as large. Two methods were used to re-attach mussels, the original and the same method with the addition of mesh bags during mussel deployment. Treatments were examined on three occasions at approximately one month intervals. At Riet River, the sites chosen did not show differences in wave strength (measured using dynamometers) or water flux (measured using erosion of cement balls) so that water motion was excluded from the analyses. Small mussels grew faster and had weaker attachment than large mussels. There was no difference in condition index between small and large mussels, or in the numbers of recruits settling among the byssus threads of deployed mussels of the two size classes. In Coffee Bay, there was no relationship between rehabilitation success and maximum wave force, and no difference in bulk water flux among sites. Small mussels deployed using mesh bags survived better than non-meshed or large mussels of either treatment. There was no difference in condition index (CI) between mesh and no-mesh, or between small and large mussels. As in the case of Riet River, small mussels grew faster than large mussels, but large mussels attached stronger than small mussels, with no effect of mesh. Although the factors that improve reseeding of mussels can be identified (use of mesh, use of small mussels, choice of sites with high recruitment rates), successful long-term rehabilitation requires appropriate subsequent management of re-seeded sites.