1. Stand thinning is being used increasingly by forest managers to maximize wood production from regrowth eucalypt forests in Australia. Many of the ecological effects of this intensive method of management, including changes in nutrient cycling and the susceptibility of sites to fire due to addition of slash residues to the forest floor, are not well understood. 2. Decomposition of slash residues was studied at one site in regrowth karri (Eucalyptus diversicolor) forest, a tall, open forest growing in south-western Australia. 3. Thinning reduced tree density from 4000 to 380 stems ha -1 , reduced tree basal area from 32 to 17m 2 ha -1 and removed 2000 stems ha -1 of understorey plants. Approximately 78 t ha -1 slash residues from the thinning operation were added to the existing 40 t ha -1 of forest floor litter. 4. Rates of decomposition of 27 components of slash were determined during a 2-year period, following thinning using the mesh bag technique. Components studied consisted of 5 leaf fractions from the overstorey and main understorey species, 15 twig and stem components from the understorey, 6 twig and stem components from the overstorey and bark from the overstorey. 5. The rate of weight loss of each slash fraction was assessed using single and double exponential decay models. The single exponential model adequately described the decay of karri stem and twig wood, and of understorey stem wood. Decay of leaf fractions, bark and small-diameter understorey wood were best described by the double exponential model. 6. Fractions of labile components in leaf, bark and small understorey wood were lost rapidly during the initial decay phase (half-lives 1-6 weeks); thereafter decomposition proceeded more linearly and was dominated by the rate of decay of the resistant fraction of each slash component. Model predictions of the fraction of each slash residue present as labile compounds were significantly related (r 2 = 0.94) to the amounts determined in laboratory extractions of the residues. 7. The decay rate of wood components was slower than for leaf fractions and was dependent on the size of the wood and its source. Rate of decay decreased with wood diameter and was more rapid for understorey wood than overstorey wood. 8. Half-lives of leaf components of slash ranged from a few weeks up to about 18 months, whereas half-lives of wood in the slash residues were much greater and exceeded 30 years for the largest diameter overstorey wood fraction. 9. Fire control and nutrient conservation are critical considerations in formulating strategies for management of thinning residues in eucalypt forests. Thinning residues deposited on the forest floor increase the risk of fire. However, where this hazard can be tolerated for limited periods, a delay in fuel-reduction burning for up to 2 years following thinning will allow time for much of the leaf material to decompose. This will reduce the fire risk and enhance conservation of the nutrient capital.