We combined extensive water sampling with monthly growth measurements of juvenile sea scallops held in cages 0 to 200 cm above the bottom to (1) construct predictive empirical models of shell and soft-tissue growth based on oceanographic variables, and (2) determine whether scallops on or near the bottom can derive a food supplement from resuspended sediment when seasonal phytoplankton production is low. Variation in growth was strongly dependent on depth, but this relationship was not consistent over time or tissue type. In late fall, when phytoplankton biomass was generally low (-1 pg chl I-'), the adductor muscle of scallops on the bottom lost mass (-1.5 mg dry wt d.'), but for scallops held only 20 cm higher in the water column, growth was 2.5 mg d.'. During the winter, softtissue growth on the bottom was significantly lower than that of scallops held above the sediment surface. At this time, there was no variation in shell growth with respect to depth. At the end of the study, soft-tissue weight (excluding muscle tissue) of scallops on the bottom was -40% less than that of scallops growing 250 cm above bottom. Rather than providing an energetic benefit, results suggest that high concentrations of seston near the bottom inhibit growth. Empirical regression models of scallop growth using data from water sampling every 2 wk accounted for up to 68% of growth variation, with temperature and seston quality being the most important predictor variables. Marginal improvements to the model uslng data collected hourly with In situ probes suggest that estimates of food supply should be corrected, i.e. reduced, when high flows or high seston concentrations limit filtration rates. In addition, results indicate that attention to the magnitude and variation of predictor variables without consideration of their seasonal coherence may be a primary factor limiting the ability to construct truly predictive models of bivalve growth.