In northern hardwood-conifer forests, alternatives to conventional forest management practices are being developed in order to maintain biodiversity and ecosystem functioning while providing for timber revenue generation. The understory layer of vegetation encompasses the majority of plant species diversity in forested ecosystems and may be sensitive to timber harvest disturbance. Thus, monitoring the response of forest understories to new silvicultural techniques may provide a means for evaluating their intensity. In this study, we hypothesize that i) uneven-aged, low-intensity silvicultural systems can maintain understory plant diversity and support latesuccessional species through harvest disturbance; ii) retaining and enhancing stand structural complexity can increase understory plant diversity in northern hardwoodconifer forests; and iii) plant responses are influenced by interactions between canopy structure, soils, and exogenous climate processes. Experimental treatments include two conventional uneven-aged prescriptions (single-tree selection and group selection) modified to increase structural retention, and a third technique designed to promote late-successional forest structure and function, termed structural complexity enhancement (SCE). Four replications of each treatment were applied to 2 ha management units at three sites in Vermont and New York, U.S.A. Understory vegetation was monitored over 2 years pre- and 4 years post-treatment. We used a linear mixed effects model to evaluate the effects of treatment, soil properties, and drought stress on understory diversity and abundance. Compositional changes among treatments were assessed with non-metric multidimensional scaling (NMS), an ordination technique. Model results show that over time, understory responses were strongly affected by overstory treatment and less influenced by soil chemistry and drought stress. All treatments were successful in maintaining overall composition and diversity. However, late-successional diversity increased significantly in SCE units compared to group selection units. These results indicate that while conventional uneven-aged systems are capable of maintaining understory plant diversity, variations that retain or enhance structural complexity may be more efficient at retaining latesuccessional species. Increased microsite heterogeneity as a result of these techniques may also increase understory plant diversity, at least during the initial post-harvest recovery period.