vast areas of forests have been degraded in ghana. the degraded areas are generally invaded by chromolaena odorata, a strong competing species for trees. planting is the major primary means that is currently used to restore these areas back to forests. however, natural regeneration or human-assisted natural regeneration may provide an alternative way to regenerate these areas. to evaluate the potential of this alternative, the density and species composition of tree seedlings under the canopy of chromolaena odorata and the seeds of both trees and chromolaena odorata in a degraded area were estimated. the response of tree seedlings to the removal of chromolaena odorata was also examined. one hundred and eight plots were established 20 m apart for assessing the tree seedlings. chromolaena odorata was removed from so % of the plots to release tree seedlings and left the other half intact. seedling height, the number of leaves per seedling, and seedling mortality were assessed in both released and unreleased plots immediately after the release treatment (june 1998) and again three months later (september 1998). smaller plots (0.2s m2) next to half of the seedling plots were used to sample the soil seed bank. the soil samples were taken from two different depths and germinated in a germination house. fifty five species of tree seedlings and seven species of tree seeds were found at the site. there were 11,780 seedlings ha-1 and the seedlings were well distributed in the degraded area. the density of tree seeds (46,000 seeds ha-1) was 3.8 times higher than tree seedlings, but the species composition o f tree seedlings was 7.8 times higher than the species composition of seeds. the spatial distribution of tree seedlings was more even than that of tree seeds. the tree seed density did not vary significantly with soil depth but the seed density of chromolaena odorata (73.89 million seeds ha*1) decreased with increasing depth from the surface. there were 1,606 times more chromolaena odorata seeds in the soil seed bank than tree seeds. tree seedlings responded positively to the removal of chromolaena odorata. the height increment and the increase in number of leaves per seedling were three times greater in released seedlings than the unreleased ones three months after the release treatment. sixty four percent o f the species suffered various levels of mortality in the unreleased plots, but all the seedlings of all the species survived in the released plots. the results suggest that there are enough tree seeds and tree seedlings to restore the degraded area back to forest. however, tree seedlings will have to be released from the competition of chromolaena odorata.