Errors in estimates of tree ages from increment cores can influence age-class distributions, affecting inferences about forest dynamics. We compare methods of height correction of increment cores taken above ground level by examining how resulting errors affect age-class distributions of ponderosa pine (Pinus ponderosa Dougl. ex P. & C. Laws.) and Douglas-fir (Pseudotsuga menziesii var. glauca (Beissn.) Franco). We compared the sapling (corrections based on the average basal age of breast high saplings) and the ground methods (corrections based on the average difference in age between ground and coring height) with a regression model we developed to overcome traditional assumptions of temporal and spatial homogeneity in early growth. Where early growth differed among mature trees or between modern saplings and mature trees, the regression method estimated age better than the two other methods. All methods of height correction over- or under-estimated tree age by at least 10 years and up to 30 years, indicating that age cannot be related to independent events of periodicities less than 10-20 years, such as El Niño, without accounting for error. Monte Carlo simulations demonstrated that error from height corrections affected the shape of age-class distributions by generating spurious regeneration pulses. We suggest that the magnitude of this error should govern the width of analytical age-classes to scale interpretations within the confidence of age estimates.©2001 National Research Council of Canada. Abstract reproduced by permission.