Abundances of forest birds in an unfragmented, undisturbed, and relatively mature temperate deciduous forest at the Hubbard Brook Experimental Forest, New Hampshire, changed markedly between 1969 and 1998. Total numbers of birds (all species combined) declined from 210–220 individuals/10 ha in the early 1970s to 70–90/10 ha in the 1990s. Of the 24 regularly occurring species, 12 decreased significantly (four to local extinction), three increased significantly, and nine remained relatively constant in abundance. Nine of the 12 declining species were Neotropical migrants. Most species exhibited similar trends on Breeding Bird Survey (BBS) routes in New Hampshire during the same 30 year period and on three replicate study sites in nearby sections of the White Mountains from 1986–1998. Probable causes of trends were diverse and differed among species. Most could be accounted for by individual species' responses to events occurring primarily in the local breeding area. The most important local factor affecting bird abundance was temporal change in forest vegetation structure, resulting from natural forest succession and local disturbances. Four species that declined markedly and in some cases disappeared completely from the study plot (Least Flycatcher, Empidonax minimus; Wood Thrush, Hylocichla mustelina; Philadelphia Vireo, Vireo philadelphicus; and American Redstart, Setophaga ruticilla) appear to attain peak abundance in early or mid successional forests. Species preferring more mature forests, such as Black-throated Green Warbler (Dendroica virens) and Ovenbird (Seiurus aurocapillus), increased significantly in abundance over the 30 year study. Other important factors influencing bird abundances were food availability and events in the migratory and winter periods. Nest-predation rates, although varying among years, showed no long-term pattern that would account for population declines, and brood parasites were absent from this forest. Findings from this study demonstrate that major changes in bird abundances occur over time even in undisturbed and relatively mature forests, and illustrate the need for considering habitat requirements of individual species and how habitat suitability changes over time when trying to assess the causes of their long-term population trends. The results also imply that any conclusions about the effects of other factors affecting forest bird abundances, such as increased nest predation or brood parasitism associated with habitat fragmentation, must also account for successional changes that may be affecting habitat suitability.