We compared colonization, growth and succession from 1989 to 2000 in a restored mangrove site and in gap and closed canopy sites in a natural mangrove forest. The restored site was created in 1982 and planted with Rhizophora mangle (???2 m-2) propagules. By 1989, Laguncularia racemosa, with densities up to 12.9 tree m-2, was a dominant in all plots, although densities were greater at edge plots relative to inner plots, and near open water (west plots) relative to further inland (east plots), and in tall mangrove plots relative to scrub plots. Rhizophora mangle (1989 tree densities about 2 m-2) was a codominant in inner and scrub plots, while Avicennia germinans had the lowest densities (<1 tree m-2) in all plots. From 1989 to 2000 L. racemosa experienced reduced recruitment and apparent density-dependent mortality of canopy individuals in plots with high initial densities. Scrub plots experienced high rates of colonization by R. mangle and L. racemosa, rapid growth in height of all species (1989-1996), followed by a dieoff of L. racemosa in later years (1997-2000) as the canopy came to resemble that of tall mangrove plots. Colonization and growth rates were lower in gap and closed canopy regions of the natural forest relative to rates in the restored site. After 11 years, densities of L. racemosa were 10-20x lower and R. mangle slightly less in the gap relative to densities in tall mangrove plots in the restored site at the same age. Although the restored stand had converged with the natural forest by 2000 in terms of some factors such as species richness, vegetation cover, litterfall, and light penetration, trees were still much smaller and stem densities much higher. Full development of mature structure and ecological function will likely require decades more development. ?? Springer 2005.