In a chestnut high forest converted into coppice, 4 permanent plots were established in 1994. These plots are being managed according to the silvicultural models proposed by Bourgeois (1992) and adapted to our conditions. The treatments are: T1 = Model 1: small dimensions; T2 = Model 2: medium dimensions; T3 = Model 3: Large dimensions; Control = coppice without intervention. In 2003, a 2 nd thinning was applied in order to select the more straight and cylindrical shoots without defects. The principal dendrometrical parameters (such as: number of stumps per hectare; number of shoots per hectare; mean total height of the shoots; dominant height of the shoots; mean DBH of the shoots; dominant diameter of the shoots; basal area of the shoots per hectare and basal area of the shoots) were measured, before and after thinning. Results show that T3 presents greater vigour and the highest dominant height. Concerning dominant height, T2 was exceeded by T1, because it has a higher number of shoots and, consequently, strong competition in relation to available site resources. It was expected a T2 with dominant height superior to T1, but we believe that T2 will recover its dominant height leadership with the heavy thinning applied in the last intervention (about 75% reduction in the number of shoots). The control follows the other plots dominant height growth pattern, although presents inferior mean basal area per shoot in comparison to the other treatments. INTRODUCTIONCoppice is a very flexible system in relation to management, producing wood material with different dimensions. Now today, logs with small dimension do not have demand as before. Thus, it is necessary to improve the management of the existent coppice stands through application of appropriate silvicultural models. With this aim, an adaptation of the models proposed by Bourgeois (1992) and Bourgeois et al. (2004) has been performed. A model to produce material of small dimensions implies a first thinning at 5-9 years old, leaving 3000 shoots per hectare and later a second thinning at 10-14 years old reducing the number of shoots to 1500. The clear cut should be realized at 25-30 years of rotation period. Another model to produce medium dimensions proposes a clear cut at 30-35 years, with two previous thinnings: the first one at 7-9 years, leaving 2000-2500 shoots per hectare, and the second at 11-13 years old, leaving 600-800 shoots per hectare, respectively. A third model to produce large dimensions determines that between the 10 th and the 13 th year, only 250 future shoots will be designated, the target of the silvicultural interventions being to obtain in the clear cut 150-250 shoots per hectare for quality timber. The final clear cut is foreseen at 40-50 years of rotation period.

No abstract