In mechanistic growth models, the description of assimilate allocation or dry matter partitioning plays a key role. Although theoretical concepts of allocation exist, they include many parameters that cannot be quantified. Therefore, many growth models use descriptive keys that represent the proportions of dry matter or carbohydrates assigned to each plant component. I have developed a model to describe the dynamic partitioning of dry matter in individual trees, and used it to investigate the effects of growth conditions on the partitioning pattern in Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) and beech (Fagus sylvatica L.). The model estimates the fractions of total available dry matter that should go to certain plant parts, based on the concept of structural balances. Both mechanistic and allometric relationships between tree components are used to model conditions for the dynamic distribution of dry matter. The model was to used to estimate the effects of dominance position, site conditions, and thinning on growth partitioning. The fractions of the annual current increment of total dry matter gradually changed with tree age, but the changes were relatively small, especially after age 20. Compared with beech, Douglas-fir invested more dry matter in foliage, especially at the cost of the branch and stem components. Trees of average basal area invested more dry matter in branches and less in stem than suppressed trees, and their estimated increase in stem diameter over time generally fitted the yield table data well. Stem diameter development was underestimated at higher ages only in the case of a Douglas-fir tree of average basal area on a poor site. Over time, the proportion of standing biomass in foliage and fine root fractions showed a gradual decline, whereas there was a gradual increase in the proportion of standing biomass in the stem fraction. These age-related changes were attributed to different loss rates among components. Analysis of the effects of thinning revealed that a discontinuous reduction in stem number results in a slow decrease in partitioning to the stem. The most obvious response to thinning consisted in a sharp decrease in partitioning to fine roots and foliage, and an increased investment in branches. Stem diameter growth appeared relatively constant in response to thinning, indicating that it will increase almost linearly with time. I conclude that the model is able to reproduce the development of an individual tree over time, both in terms of stem diameter and biomass. The model is thus suitable for simulating the effects of competition for resources on growth and development of forest stands.