Litter decomposition in two Nothofagus pumilio stands located in Magallanes Region (Chile) was studied during two years. One stand was a primary forest (SI), without silvicultural treatment, and the other had a regenerative felling under a shelterwood system (CR). Weight loss and litter decomposition rates were obtained with the litter-bag method. At each stand, 36 litter-bags were systematically placed on the forest floor. A fraction of the litter-bags was then randomly extracted at months 2, 5, 9, 11, 12, 15, 21, 22 and 24. The weight of the collected litter-bags was measured at each sampling period. Litter weight-loss presented differences between the stands only after 12 months since the litter-bag set up. After 24 months, the litter weight-loss was similar between the stands, reaching 60 % of the initial weight. After one year of the experimental set-up, litter decomposition rates (k) were 0.76 and 0.44 for CR and SI stands, respectively. At the end of the trial, these values changed to 0.44 and 0.46 for CR and SI stands, respectively. The model that best fitted the relative litter mass according to the initial litter weight was the quadratic regression model, using the number of months, when the litter was on the forest ground, as the independent variable (R² = 0.95 for both stands). These results suggest that the effect of the regeneration felling on the litter decomposition rates decreases with time.