Nearly all previous studies of saltmarsh-nesting Common Terns on the east coast of the United States have concluded that tidal saltmarshes were suboptimal or marginal breeding habitats. Questioning that conclusion, we analyzed patterns of both saltmarsh and nonmarsh colony use (stability, movement, establishment, abandonment, and size) obtained during 5 yr of annual helicopter censuses of all Common and Roseate terns breeding on Long Island, New York. We found 1900–3600 pairs at 10–33 saltmarsh and 22–30 nonmarsh sites; there were few biologically important differences between Common Terns nesting at marsh and at nonmarsh sites. We did find that (1) marsh sites and colony sizes increased through the study period; (2) both marsh and nonmarsh colonies grew with duration of occupancy; (3) smaller marsh and nonmarsh colonies (<50 pairs) usually lasted only 1–2 yr, while larger colonies were equally likely to persist for 1, 2, 3, 4, or 5 yr; (4) numbers of marsh and nonmarsh sites used each year were generally unrelated to population sizes; (5) 5-yr sites composed only 10.6% of total marsh and 17.6% of total nonmarsh sites; (6) the mean sizes of both newly established and about-to-be-abandoned colonies were smaller than the mean sizes of all others when averaged between but not within years; (7) most previously occupied sites, once abandoned, remained so for only 1 yr, and most new sites were occupied for only a single year; (8) annual turnover rates were 32%–49% for both marsh and nonmarsh sites; (9) marsh and nonmarsh breeding populations were correlated each year, allowing estimation of the total Long Island population to within ±4% by censusing only the 20–25% in saltmarshes. Roseate Tern data were few, especially in marshes, obviating marsh-nonmarsh comparisons, except that Roseates failed to persist in saltmarshes, and their overall mean colony sizes across the same numbers of years' occupancy were usually smaller than Commons', although their turnover rates were roughly the same. We conclude that saltmarsh-nesting Common Terns are well adapted to marsh nesting and that they have probably been doing so for perhaps hundreds of generations. We hypothesize that it may have been a relict population of saltmarsh-nesters that saved the species from extirpation in the late 1800s. In contrast, Roseate Tern's failure to exploit extensive saltmarsh habitat seems yet another factor abetting its precarious status in northeastern North America.