Abstract(#br)We investigated whether the establishment of forests on organic grassland soils leads to significant losses in soil carbon, due to the site preparation for the planting of trees and other disturbances. We also investigated whether subsequent biomass accumulation over two consecutive rotations separated by a clearfell leads to a recovery in the initial soil carbon losses.(#br)Soil respiration, soil carbon stocks and annual litter fall were measured, while allometric equations were used for the estimation of above- and belowground biomass, in a replicated forest chronosequence of Sitka spruce ( Picea sitchensis ) on peaty gley soil, in Harwood Forest (NE England). The sites chosen were: 40-year-old first rotation stands, 12, 20 and a 30-year-old second rotation stands, one... 18-month-old clearfelled site and unplanted natural grassland sites. The soil carbon stock of the first-rotation 40-year-old stands was 140±15tCha −1 , far lower than in the surrounding unplanted grasslands (274±54tCha −1 ), while clearfelling caused a further decline (100±13tCha −1 ). Soil carbon accumulated again as the forest grew during the second rotation (147±8, 181±17 and 249±40tCha −1 at the 12, 20 and 30-year-old stands, respectively). Soil respiration was the highest at the grassland site (14.2±3.1tCha −1 year −1 ), possibly due to high respiration by the grasses. After forest establishment, soil respiration was: 2.3±0.9, 2.2±0.7, 5.4±0.7 and 5.0±0.4tCha −1 year −1 at the age of 12, 20, 30 and 40 years, while in the clearfelled site soil respiration was 5.6±1.6tCha −1 year −1 .(#br)The inputs and outputs of carbon into the soil were combined to evaluate the long-term effects of forest management on soil carbon dynamics. It was concluded that the establishment of coniferous forests on peaty gley grassland soils in the uplands can lead to a net accumulation of soil carbon during the second rotation.