To test the prevailing premise that cedar-swamp succession is controlled through replacement by hardwood species, we analyzed the size-class structure of Atlantic white cedar (Chamaecyparis thyoides) swamps that were established in the New Jersey Pinelands (Pine Barrens) prior to 1930. Cedar size-class distribution patterns in most stands were characteristic of self-thinning, even-aged stands. Hardwood species, which included red maple (Acer rubrum), blackgum (Nyssa sylvatica), sweetbay (Magnolia virginiana), and gray birch (Betula populifolia), occurred at low densities and contributed little to total basal area. For red maple, the most common hardwood species, relative basal area, canopy cover, density of large-diameter ( >10 cm) stems, and relative density of small (< 10 cm) and large ( >10 cm) diameter stems (expressed as a percentage of total small-diameter and large-diameter cedar and maple stems) increased in relation to increasing cedar-stem size and decreasing cedar-stem density. Although red maple abundance and recruitment may be related to cedar size-class structure, the absolute density of 5 - 9.9 cm diameter cedar was greater than that of red maple and we found no significant difference in the absolute density of <5 cm diameter maple and cedar stems. It is doubtful that any hardwood species occurred at densities sufficient to allow it to quickly exploit canopy gaps created by the death of overstory trees. We concluded that hardwood replacement of cedar in swamps is not a certain outcome of cedar-swamp succession. If it does occur, it may take centuries and only result in conversion to mixed stands dominated by cedar.