Relationships among stand composition, stemwood productivity, and canopy structure were investigated for 55 study areas in northeastern Minnesota. Tree species composition among study areas was deliberately allowed to vary. Aspen, primarily quaking (Populus tremuloides Michx.) and to a lesser degree bigtooth (Populus grandidentata Michx.), was a significant component of every study area. The two most common associates were paper birch (Betula papyrifera Marsh.) and balsam fir (Abies balsamea (L.) Mill.). Productivity was defined as mean annual increment of tree stemwood volume and ranged from 3.3 to 12.6 m3·ha-1·year-1. A multiple-regression approach was used to investigate the relationships between productivity, stand composition, and canopy structure. Base models relating productivity to aspen site index, aspen cohort age, and total basal area were developed. Measures of stand composition and canopy vertical structure were added to the base models, and their significance in explaining residual variation in productivity was tested. Productivity was found to be negatively correlated with stand composition and canopy vertical structure, with all other factors held constant. Pure aspen and single-canopied stands were expected to be the most productive. Exceptions were present in the data: the two most productive stands were vertically stratified, aspen - balsam fir - paper birch mixtures.