In 1980 the State of Connecticut began a tidal marsh restoration program targeting systems degraded by tidal restrictions and impoundments. Such marshes become dominated by common reed grass (Phragmites australis) and cattail (Typha angustifolia and T. latifolia), with little ecological connection to Long Island Sound. The management and scientific hypothesis was that returning tidal action, reconnecting marshes to Long Island Sound, would set these systems on a recovery trajectory. Specific restoration targets (i.e., pre-disturbance conditions or particular reference marshes) were considered unrealistic. However, it was expected that with time restored tides would return ecological functions and attributes characteristic of fully functioning tidal salt marshes. Here we report results of this program at nine separate sites within six marsh systems along 110 km of Long Island Sound shoreline, with restoration times of 5 to 21 years. Biotic parameters assessed include vegetation, macroinvertebrates, and use by fish and birds. Abiotic factors studied were soil salinity, elevation and tidal flooding, and soil water table depth. Sites fell into two categories of vegetation recovery: slow, ca. 0.5%, or fast, more than 5% of total area per year. Although total cover and frequency of salt marsh angiosperms was positively related to soil salinity, and reed grass stand parameters negatively so, fast versus slow recovery rates could not be attributed to salinity. Instead, rates appear to reflect differences in tidal flooding. Rapid recovery was characterized by lower elevations, greater hydroperiods, and higher soil water tables. Recovery of other biotic attributes and functions does not necessarily parallel those for vegetation. At the longest studied system (rapid vegetation recovery) the high marsh snail Melampus bidentatus took two decades to reach densities comparable with a nearby reference marsh, whereas the amphipod Orchestia grillus was well established on a slow-recovery marsh, reed grass dominated after 9 years. Typical fish species assemblages were found in restoration site creeks and ditches within 5 years. Gut contents of fish in ditches and on the high marsh suggest that use of restored marsh as foraging areas may require up to 15 years to reach equivalence with reference sites. Bird species that specialize in salt marshes require appropriate vegetation; on the oldest restoration site, breeding populations comparable with reference marshland had become established after 15 years. Use of restoration sites by birds considered marsh generalists was initially high and was still nearly twice that of reference areas even after 20 years. Herons, egrets, and migratory shorebirds used restoration areas extensively. These results support our prediction that returning tides will set degraded marshes on trajectories that can bring essentially full restoration of ecological functions. This can occur within two decades, although reduced tidal action can delay restoration of some functions. With this success, Connecticut's Department of Environmental Protection established a dedicated Wetland Restoration Unit. As of 1999 tides have been restored at 57 separate sites along the Connecticut coast.