Soft-shell clams ( Mya arenaria L.) are a commercially important intertidal bivalve in Maine, USA where they are managed as a common property resource. Ten million 8–12 mm clam juveniles are reared annually at the Beals Island Regional Shellfish Hatchery and are transplanted onto publicly-owned mudflats during the fall or early winter. The success of this strategy is unpredictable because of sporadic occurrence of winter ice formation and subsequent scouring of the intertidal zone. These events, in addition to other severe winter storms, can result in complete mortality of the transplanted seed. Transplanting hatchery-reared seed in the spring, therefore, is preferable, but until now there has been no economically effective technique to hold millions of soft-shell clam seed over the winter. A field experiment designed to test the interactive effect of clam size and density on survival was conducted during the winter of 1991–1992 at a sheltered and exposed site near Beals, Maine, USA. Three sizes of hatchery-reared clam seed (Small, X ̄ = 4.3 mm shell length [SL]; Medium, 8.2 mm SL; Large, 11.5 mm SL) were used. Clam density was approximate and depended on clam size (Small, 18 750 ind, 37 500 ind, and 56 250 ind; Medium, 11 000 ind and 22 000 ind; Large, 4500 ind and 9000 ind). Clams of each size/density combination were added separately to nylon window screen, zippered bags (aperture = 1.8 mm). Bags containing clams were housed in wooden-framed subunits (0.45 m × 0.45 m × 0.08 m) covered with a 12 mm extruded mesh netting. Subunits were arranged vertically in groups of 4–6 and were submerged 1 m above the bottom in shallow water (3–8 m) from November 1991 to April 1992. Mean survival of medium and large size clams at both sites was 97.7%. Significant density effects were detected for both size groups at the sheltered site, although the difference between treatment levels was less than 2%. Mean survival of small clams pooled across sites was 67.8%. At the sheltered site, clams in the uppermost level of the overwintering units had consistently lower survival rates than those nearer the bottom. Density effects for small clams were detected only at the sheltered site where animals held at the highest density had significantly higher survival rates than animals at either of the two lower densities (75.3% vs. 60.3%). Clam loss over the winter may be reduced by suspending seed in seawater instead of transplanting it to intertidal flats.