An annular flume was used to measure the effect of increasing current velocity on mussel (Mytilus edulis) feeding rate and the stability of mussel beds sampled from the mouth of the Exe estuary (SW England). It was found that, in contrast to earlier flume studies, the feeding rates of mussels from open coast sites were unaffected by current velocities up to 0.8 m s -1 . Algal cell depletion in the water column above mussels was a function of current velocity, increasing with declining currents below 0.05 m s -1 . The erodability/stability of the mussel bed, measured in terms of critical erosion velocity, sediment mass eroded and mean erosion rate, was found to be a function of the nature of the substrate and the density of the mussels. Erosion of mussel beds on sandy substrate showed a non-linear relationship with mussel bed density. In comparison with the sand (0% mussel cover), sediment resuspension was about five and four times higher for 25% and 50% cover, respectively. This was due to the increased turbulence and scouring around the clumps of mussels in low-density parts of the bed, and this resulted in some mussels detaching from the bed. At ~100% mussel cover, the sandy bed was more protected by the dense surface layer of mussels, and none became detached during erosion due to the high number of byssal attachments between individuals. The sediment resuspension from the 100% mussel cover was about three times lower than the 0% cover. Erosion of the bed with 50% cover resulted in burial of a large proportion of the mussels, with a 6 cm increase in sediment level. However, the mussels returned to the surface and recovered in 1-2 days, due to a combination of migration upwards and substrate settlement. Channels on the edge of the main Exmouth mussel bed were characterised by a more stable substrate comprising pebbles and sand with varying mussel densities. At these sites, where mussels experience high current velocities on spring tides (up to 0.9 m s -1 ), there was no difference between the erodability of pebble/sand substrate with 0% and 100% mussel cover. The sediment erosion was also lower than the 100% mussel cover on the sandy substrate, particularly at currents >0.4 m s -1 . Sampling of different parts of the mussel bed at Exmouth showed mussels at low densities were made up of smaller clumps with a lower mass ratio of mussels to attached substrate (pebbles/sand), thus providing a greater degree of anchorage.