Current knowledge on tree carbon (C) allocation to wood is particularly scarce in plants subjected to disturbance factors, such as browsing, which affects forest regeneration worldwide and has an impact on the C balance of trees. Furthermore, quantifying the degree to which tree rings are formed from freshly assimilated vs. stored carbohydrates is highly relevant for our understanding of tree C allocation. We used (13)C labelling to quantify seasonal allocation of stored C to wood formation in two species with contrasting wood anatomy: Betula pubescens Ehrh. (diffuse-porous) and Quercus petraea [Matt.] Liebl. (ring-porous). Clipping treatments (66% shoot removal, and unclipped) were applied to analyse the effect of browsing on C allocation into tree rings, plus the effects on tree growth, architecture, ring width and non-structural carbohydrates (NSCs). The relative contribution of stored C to wood formation was greater in the ring-porous (55-70%) than in the diffuse-porous species (35-60%), although each species followed different seasonal trends. Clipping did not cause a significant depletion of C stores in either species. Nonetheless, a significant increase in the proportion of stored C allocated to earlywood growth was observed in clipped birches, and this could be explained through changes in tree architecture after clipping. The size of C pools across tree species seems to be important in determining the variability of seasonal C allocation patterns to wood and their sensibility to disturbances such as browsing. Our results indicate that the observed changes in C allocation to earlywood in birch were not related to variations in the amount or concentration of NSC stores, but to changes in the seasonal availability of recently assimilated C caused by modifications in tree architecture after browsing.