Tree stem diameters were measured in a stratified random sample of plots in each of 19 forest plantation estates in nor-them New South Wales. The estates were 0.2-170 ha in area and the plantations were 1-10 years old, growing eucalypts, sub-tropical rainforest species or an exotic conifer, both in single-species and mixed-species plantations. A sample of trees was measured also for biomass and estate- and region-specific allometric relationships developed to predict tree biomass from tree diameter. With the stratified random sample data, the allometric relationships were used to predict the total amount of carbon sequestered in tree biomass, and its 95% confidence limit, across each estate. These results were compared with estimates made using an allometric relationship from Nova Scotia, Canada, and a stand-based relationship from Australia. Bias in estimates made using the region-specific or Nova Scotia relationships appeared to be unimportant if the confidence limit was > 10% of the estimate. Bias was greater using the stand-based relationship and was unimportant only if the confidence limit was > 30% of the estimate. It was concluded that using sampling intensities of around 2-4% of the estate area, the total carbon sequestered by an individual small plantation estate in the region could generally be estimated satisfactorily with a 95% confidence limit of about 30-40% of the estimate or better, with a minimum of about 10%. Generally, the younger the plantation estate, the higher was the sampling intensity necessary to achieve the same precision of estimate as in older estates. If small plantation owners in the region are intending to offer sequestered carbon for sale as carbon credits, it appeared it will be difficult for them to estimate carbon sequestration by their trees with a 95% confidence limit as low as the present Australian recommendation of 10% of the estimate. If owners pooled their estates in a cooperative, they should be better able to achieve the required confidence limit from the much larger, pooled estate. Analysis of organic carbon contents of soil below plantations and adjacent pasture paddocks suggested there was a decline in soil carbon with time following plantation establishment. Growers will need to account for these losses when estimating the total carbon sequestered by their plantations ecosystems. (C) 2003 Published by Elsevier Ltd.