We examined the influence of food availability and population density on the morphometry and shell length body mass relationship of Mytilus edulis. Mussels were reared in the laboratory for 22 mo at 8 different density levels in 11 chambers supplied with natural seston at 2 different concentrations. This allowed us to assess separately the effects of food availability and mussel density. The shell length/width and shell height/width ratios were affected by food, density and time. Mussels tended to be narrower (flatter) at high density and at low food level. Therefore, narrow shells could result from reduced food concentration in high density situations without implying physical interference. Shell mass was also influenced by both food and density levels, but to a lesser extent than tissue dry mass. In contrast with soft tissue mass, shell mass increased significantly for all food and density levels between October 1995 and October 1996, The elevation of the shell length-body mass-population density relationship was lower at low food availability, The slope of the tridimensional relationship, however, increased with decreasing food level, indicating apparent asymmetric competition for all food-density treatments pooled together, This pattern, however, is misleading because mussels obviously cannot interact among chambers. Since the slopes of length mass relationships are used in predicting self-thinning exponents in space-regulated situations, it follows that self-thinning exponents in mussels should be sensitive to background food level, thus limiting the use of self-thinning relationships for resolving factors regulating growth.