We measured survival and growth of three shade—tolerant species of maple (Acer pensylvanicum, A. rubrum, and A. saccharum) in response to understory and experimentally created small canopy gaps of two sizes (8 x 12m, 75 m2; 16 x 24 m, 300 m2) in central New England. Seedlings of the three species (2160 total, 720 per species) were transplanted into five plot locations (center plus northwest, northeast, southwest, and southeast gap edges) within all gap and understory sites. Measurements of microclimates, architecture, photosynthetic performance, survival, and growth were made over 1 yr before, and 2 yr following, gap release. Red maple (A. rubrum) survived better overall across the study due to greater persistence in the north and center plots of large gaps. The small gaps and understories showed no differences among the species. Survival rates exceeded 80% in most sites and plots, with low values (30—65%) only in the exposed plots of large gaps. There were no relationships between post—gap survival and previous age, height, or basal diameter. By the end of 2 yr of gap release, both gap sizes induced greater distinctions among the species in all growth variables than the understory. Striped maple (A. pensylvanicum) exhibited greater leader extension, absolute stem height, net height change, absolute basal diameter, and net basal diameter change than red maple and sugar maple (in that order) in nearly all sites and plots. The exception was large—gap center and north plots, where red maple equalled or exceeded striped maple in net basal diameter change, but not net height increase. Sugar maple (A. saccharum) was the least responsive of the species to the gap—understory gradient. As with survival, there were no predictable relationships between pregap age or size and post—gap growth. Photosynthetic performance paralleled growth by these species across the gradient, particularly for shoot assimilation. When growth variables were plotted against irradiance and temperature measured at seedling plot positions, there were consistent and clear distinctions among species across the gap—understory gradient, providing limited evidence for gap partitioning in our system. Striped maple appears to be a superior generalist, red maple is a weaker generalist, and sugar maple shows the poorest performance in a manner that is nearly insensitive to the gap—understory gradient, in our experimental system.