One of the main elements in developing an optimal release strategy for an enhancement e¡ort is to evaluate and select release sites that will support growth and survival of newly released, cultured ¢sh. Three potential release sites (New Castle (NC), Broad Cove (BC) and Oyster River (OR)) in the Great Bay Estuary, NH, USAwere evaluated for pilot-scale releases of winter £ounder (Pseudopleuronectes americanus Walbaum). Cultured juvenile £ounder were placed in cages at each of the three sites. Sites were evaluated based on growth and survival of the ¢sh in relation to water temperature, prey availability and sediment composition. Fish grew faster in the two upper estuarine sites BC (0.54 mm day À1 ) and OR (0.56 mm day À1 ) than at the site at the mouth of the estuary (NC 5 0.37 mm day À1 ). Fish survival (44^53%) and water temperature (17.8^19.7 1C) did not vary between sites. Benthic samples showed that prey was available to, and eaten by, the £ounder. Sediment composition was the main di¡erence between the three sites, with one site (NC) characterized by gravel whereas the other two sites were sandy. These results corroborate other studies showing the importance of sediment quality for the distribution of £at¢sh populations. From these results, we can con¢dently eliminate NC as a potential release area and recommend that sandy sites are better for stocking cultured juvenile winter £ounder than gravely sites.