Intertidal mussel beds were severely damaged by scouring ice floes during the winter of 1995/96. Aerial surveys before and after the winter showed that more clusters of mussel beds vanished in a region with a higher areal share of tidal flats and a lower salinity, suggesting that the amount of ice present determined the magnitude of the disturbance on beds of Mytilus edulis. Nehls and Thiel (1993) observed a strikingly similar spatial pattern of disturbances caused by severe storms in the Wadden Sea. Areas on mussel beds mechanically undisturbed by ice showed no reduced abundance and biomass of mussels, indicating that temperature alone was of little importance as a lethal factor. Conversely, Cerastoderma edule was strongly affected by low temperature. On average 80% died during the winter with extinctions >93% in the high tidal zone. At the lowest tidal level, surviving cockles were larger than those killed by the frost. A reinvestigation of sampled sites in autumn revealed that substantial further mortality occurred during spring and summer which may constitute a time-lag effect of the preceding winter. There was no increased mortality in juvenile and adult Mya arenaria during the winter 1995/96, confirming that this clam is a hard-winter-species just like Macoma balthica.