ABSTRACT: The influence of Zostera marina L. seed-density on germination and initial seedling success was investigated using seed-addition field experiments at 2 scales in the Chesapeake Bay region in 1999 and 2000. We first tested whethergermination rates and initial seedling establishment were affected by initial seed-densities of 2.5, 25, 250, and 1250 seeds m^-2 within 4 m² plots. We then tested whether plot size affects germination rates, following the hypothesisthat rates of seed predation might be different in large and small plots. We broadcast seeds at a single density (500 seeds m^-2) but at a much larger plot size (100 m², or 25 times the size of the small plots). In the springfollowing seed broadcast, seedlings were present in most 4 m² plots (seedling densities of 0.6 to 15.4% of the number of seeds released in 1999, and 3.3 to 23.3% of those released in 2000) and in all 100 m² plots (4.3% to 13.9%).Seed-density effects were not significant in 1999 or 2000, while site effects were significant in both years. The percentages of seedlings in the larger plots were similar to those in the smaller plots. These results suggest that there were nodensity-dependent effects on germination and initial seedling establishment, and that within the size range of plots examined in this study, such processes are not likely to be scale-dependent. The significant differences among the sites may be related tomicro-topographic complexities of the bottom caused by both biotic and abiotic factors that allow seeds to be retained close to where they settle. Our data, combined with previously published data on seed dispersal and patch dynamics, stress theimportance of conserving existing beds, regardless of bed size and shoot density, since these are sources of seeds that may establish new patches. The data may also help in developing strategies for the restoration of denuded sites using seeds instead ofadult plants.