Native riparian cottonwood (Populus deltoides) dominated habitats are declining throughout the southwestern United States and being replaced by the exotic salt cedar (Tamarix ramosissima, T. chinensis) dominated systems. Restoration efforts have often involved removing salt cedar and simulating or restoring overbank flooding timed with seed rain of cottonwood and willow (Salix spp.). Success is then often assessed by examining seedling densities of the native species and their known competitive superiority over salt cedar at this stage. We measured plots in the Middle Rio Grande Valley, New Mexico, USA that had been cleared of salt cedar a decade earlier, compared the initial seedling community to the established community, and determined variables important in predicting establishment of cottonwood. During the decade since establishment, the proportion of cottonwood and seep willow (Baccharis glutinosa) increased, while salt cedar and coyote willow (S. exigua) decreased. Proportion of community foliar volume changed little for salt cedar (17.0 to 17.6%), while it increased for cottonwood (25.5 to 63.3%). Cottonwood dominated (>70%) the overstory. Initially established woody stem density (all species) was the only predictor of the percent of cottonwood canopy cover a decade later as declining coyote willow allowed for greater cottonwood growth. No abiotic variables were important in predictions of subsequent cottonwood community establishment, probably because these variables were similar among sites in this active floodplain. High initial cottonwood and salt cedar seedling densities were related to lower cottonwood foliar volume and smaller individual tree diameters a decade later. Low initial densities of cottonwood allowed individuals to grow rapidly to a greater height, dominating the canopy. Restoration biologists should therefore consider areas that have low initial cottonwood seedling densities as potentially very successful if their goal is to produce a stand of large cottonwoods in a relatively short time.