Nesting Common Terns (Sterna hirundo) were studied at a number of barrier beaches and small islands of tidal salt marsh in New Jersey and the Eastern Shore of Maryland-Virginina from 1980 through 1982. Data were collected on clutch sizes, nest spacing, and nesting success. The principal null hypothesis tested was that no difference in reproductive success exists between beach and marsh habitats. Nests were monitored from egg-laying in mid-May until mid-July when young fledged. Clutch sizes varied among colonies and across years but no systematic effect of year, habitat, or colony size on mean clutch size per colony was detected. Analyses of nest productivity (estimated using both the Mayfield method and using a colony average) failed to reveal significant effects of habitat or colony size but showed a stronger year effect. Storm tide flooding and egg chick disappearance (presumably predation by Herring Gulls Larus argentatus and Laughing Gulls L. atricilla nesting nearby) accounted for most nest failures. Losses due to both these mortality factors were unpredictable from year to year. Nest spacing in salt marsh colonies was much closer than it was on barrier beaches. In mixed-species colonies with Black Skimmers (Rynchops niger), distances between tern and skimmer nests were also much smaller in marsh colonies than they were on beaches. The limited amount of wrack (windrows of dead, matted vegetation) preferred by marsh-nesting terns probably explains these spacing differences. Several lines of evidence suggest that terns prefer beaches to marshes for nesting, however, the uncertainty of predation and flooding may often obscure any intrinsic differences in habitat quality. Long-term field studies are essential for testing hypotheses related to differential fitness of individuals among habitats.