ABSTRACT. The spatial heterogeneity of element fluxes was quantified by measuring litterfall, throughfall and litter decomposition for 1 y in 30 randomly located sampling areas in a lowland dipterocarp rain forest. The idea tested was that turnover of elements is more variable than turnover of dry matter in a forest with extremely high tree species diversity. In spite of the low fertility of the soil (an ultisol), total litter production (leaves, trash, and wood <2 cm in diameter) was high (1105 g m−2 y−1) with inputs to the forest floor of carbon, nitrogen, phosphorus, calcium, magnesium, potassium, manganese and iron of 550, 15.3, 0.47, 6.26, 2.49, 4.75, 0.95 and 0.14 g m−2 y−1 respectively. Throughfall was 81% of the annual rainfall and transferred 22.2, 1.37, 0.14, 1.07, 0.67, 0.39, 7.92, <0.06, and <0.06 g m−2 y−1 of organic carbon, nitrogen (all forms), phosphorus, sulphur, calcium, magnesium, potassium, manganese and iron, respectively. Average turnover rates of nutrients in litter were highest for potassium and decreased in the sequence calcium, magnesium, carbon, nitrogen and phosphorus. Concentrations of phosphorus, nitrogen and potassium in litterfall, litter mass and topsoil were closely correlated with each other. Concentrations of calcium and manganese were positively correlated with each other and with elevation. Variations in leaf chemistry and total litterfall caused the spatial heterogeneity of element input to the forest floor to have a coefficient of variation of 30 — 70%, depending on the element. Due to the strong positive correlation between element fluxes and pools, the spatial variability of turnover rates (CV c. 20%) was lower than that of element input. Turnover rates for K varied by a factor of 4, and for Ca by a factor of 2.8 when the different sites were compared. The results strongly suggest that the most important factor determining spatial heterogeneity of organic matter and element dynamics on the forest floor is the site-specific amount of leaf fall, rather than spatially variable decomposition rates.