Herbivory, lighting regimes, and site conditions are among the most important determinants of forest regeneration success, but these are affected by a host of other factors such as weather, predation, human exploitation, pathogens, wind and fire. We draw together > 50 years of research on the Huntington Wildlife Forest in the central Adirondack Mountains of New York to explore regeneration of northern hardwoods. A series of studies each of which focused on a single factor failed to identify the cause of regeneration failure. However, integration of these studies led to broader understanding of the process of forest stand development and identified at least three interacting factors: lighting regime, competing vegetation and selective browsing by white-tailed deer (Odocoileus virginianus). The diverse 100-200 year-old hardwood stands present today probably reflect regeneration during periods of low deer density (< 2.0 deer/km super(2)) and significant forest disturbance. If this hypothesis is correct, forest managers can mimic these 'natural windows of opportunity' through manipulation of a few sensitive variables in the system. Further, these manipulations can be conducted on a relatively small geographic scale. Control of deer densities on a scale of 500 ha and understory American beech (Fagus grandifolia) on a scale of < 100 ha in conjunction with an even-aged regeneration system consistently resulted in successful establishment of desirable hardwood regeneration.