This study deals with the influence of benthos abundance when assessing the potential impact of a small wind park on wintering sea ducks. Using the Before-After-Control-Impact design, it was suggested in a recent study (Guillemette et al. 1998) that the wind park provoked a decline in the abundance and a change in the distribution of common eiders Somateria mollissima and common scoters Melanitta nigra. However, the observed decline in sea duck abundance occurred concomitantly with a decline of benthic food supplies. We measured concomitant food and common eider abundance for a fourth year at Tunø Knob to test the hypothesis that, if food abundance increases again, we should also observe an increase in duck abundance. The methods used in this study are similar to those applied in the aforementioned study. The results showed that the average number of common eiders increased considerably in 1997-98 (3,361 individuals) compared to 1996-97 (458), even surpassing the level observed during the baseline year in 1994-95 (1,821). A significant increase in the abundance of common scoters occurred in 1997-98 compared to 1995-96 and 1996-97, but not in relation to the baseline year. The abundance of food for sea ducks also increased in 1997-98 where the frequency of occurrence of most potential prey reached the level observed during the baseline year. The density of blue mussels, a preferred prey species, was 1,113 individuals m-2 in 1997-98 compared to 11,111 individuals m-2 during the baseline year and only 100 individuals m-2 in 1996-97. Computations of the amount of food supply eaten by wintering common eiders suggest that, during the baseline year, prey were superabundant. This may explain why we observed a great abundance of common eiders in 1997-98 despite the fact that mussel abundance did not reach the level observed during the baseline year. Finally, the distribution of common eiders in 1997-98 on the study site as a whole was very similar to the distribution observed during the baseline year. A similar observation was made around the wind park. These results support the hypothesis that the decline of sea ducks observed during the two After years was not caused by the wind park. We conclude that without measuring the abundance and the distribution of food supply, it will remain difficult to make any reliable impact assessment of an offshore wind park on sea ducks.