As fast-growing, short-rotation plantations are being planted in the tropics on low fertility soils, the problem of sustaining soil fertility becomes an important management issue. Above-ground biomass, nutrient concentration of above-ground tree tissues, and soil nutrients were examined in two young plantations of eight indigenous tree species grown in pure and mixed designs in a low fertility site in the humid lowlands of Costa Rica. The goal was to assess the role of nutrient accumulation in above-ground biomass on potential site nutrient decline, and to draw recommendations to conserve site nutrients in the long term. In Plantation 1, Jacaranda copaia pure stands had higher above-ground tree N, P, and Mg than the other treatments, while Vochysia guatemalensis had the greatest accumulation of K and Ca. For J. copaia, stem harvest would remove about 54% of total above-ground tree N, but about 80% of P, K, Ca and Mg. For V. guatemalensis, stem harvest would remove less than 30% of N but from 50 to 60% of total above-ground tree Ca, K, Mg and P. Branches and foliage summed together were 25 to 35% of total above-ground tree biomass, but they generally represented about 50% of above-ground tree nutrients. In Plantation 2, the mixed stands had the highest above-ground nutrient content for all nutrients, and both the mixture and Terminalia amazonia pure stands had the highest stem P and Mg. Five years after planting, decreases in soil P, K and Ca were apparent in pure plots of the fastest growing species with the largest accumulation of nutrients in above-ground biomass, such as J. copaia and V. guatemalensis. However, in other cases, beneficial effects on some soil nutrients were noted: for example, increases in soil Ca under T. amazonia and Virola koschnyi, both species with high Ca content in foliage and high rates of annual litterfall. The mixed plots showed intermediate values for the nutrients examined, and even improved soil conditions, as for P in Plantation 1. This suggests that in mixed conditions it may take longer to deplete soil nutrients than in monospecific stands of fast-growing species. Results of continued sampling will be needed to assess the long term effects of plantation treatments on soil chemistry, especially near the end of the rotation (estimated at 12–15 years, depending on the species). The calculation of whole-stand nutrient budgets can help in the selection of tree species and plantation management strategies to favor nutrient recycling mechanisms and site nutrient conservation.