Single-tree selection cutting is sometimes believed to be similar to the natural gap disturbance regime of hardwood forests, but few studies have specifically compared the compositional and structural characteristics of old-growth hardwood stands, undergoing natural gap dynamics and hardwood stands previously subjected to partial cuts. This study characterized and compared the composition (saplings and trees) and structure (gaps, foliage distribution, tree diameter and density, snags and coarse woody debris) of old-growth stands (OG), 12-year-old selection cuts (SC), and 28-33-year-old diameter-limit cuts (DLC) in sugar maple (Acer saccharum)-dominated northern hardwood stands.Results showed marked structural differences between OG and harvested stands, with stronger differences between DLC and OG than between SC and OG. The synchronized formation of numerous canopy openings in harvested stands induced a massive post-harvest recruitment of advance regeneration in both SC and DLC that created a dense foliage layer in the understory. Large living trees (dbh > 39.1 cm) and defective trees were less numerous in SC than OG, which can have a detrimental impact on species dependent on these structural elements, and on the future availability and characteristics of coarse woody debris. Relatively few compositional differences were noticed among stand types, although a greater proportion of mid-tolerant species was found in the post-harvest recruitment cohorts of harvested stands compared to OG, and a lower proportion of beech (Fagus grandifolia Ehrh.) saplings was observed in DLC compared to OG and SC.We argue that even if selection cutting is closer to the natural disturbance regime of hardwood forests than diameter-limit cutting, and therefore representing progress toward the development and implementation of a natural-disturbance-based management, a recurring application of selection cutting might lead to a homogenization of forest structure and composition, a reduction of key structural features and a reduction in biological diversity at both the stand and landscape scales. Some management recommendations are proposed. #