Regeneration of tree populations following stand-replacing wildfires is an important process in the multi-century development of Douglas-fir- western hemlock forests. Temporal patterns of tree establishment in naturally regenerated, mid-aged (100 to 350 years) Douglas-fir-dominated forests have received little study in comparison with the abundance of research on regeneration in older Douglas-fir stands (>400 years of age). Increment cores were obtained from 1455 trees in 18 mature and early old-growth forests in western Washington and northwestern Oregon USA in order to determine temporal patterns of natural Douglas-fir regeneration following stand-replacing wildfire. Continuous regeneration of Douglas-fir for many decades following initiating fire was evident in all of the stands. The establishment period averaged 60 (range 32 to 99) years. The pattern observed contrasts both with the view of rapid (one- to two-decade) regeneration of Douglas-fir portrayed in early forestry literature and with reports of establishment periods exceeding 100 years in older (>400 year) Douglas-fir- western hemlock stands. Current intensive production forestry practices directed toward rapid and uniform stand closure following logging have no precedent in the historic natural patterns of Douglas-fir regeneration documented in this study. Conversely, results of this study provide evidence that early seral ecosystems persisted for several decades following wildfires. Patterns of structural development in mid-successional Douglas-fir dominated forests - a period in which forest structures evolve from the relatively simple conditions found in young forests to the complex old forests - is poorly understood. Stand structure and composition was analyzed in nine early old-growth (200 to 350 year old) Douglas-fir-dominated stands in western Washington and Oregon, all of which originated following a single stand-replacement wildfire. Structure and composition of live tree populations (density, diameters, and heights) as well as dead tree structures (snags and logs) were quantified and compared with conditions in previously reported studies of older (400-600-year old) forests. Stand-level attributes were analyzed using descriptive statistics, nonlinear regression, principal components analysis, and two old-growth indices. Variability among stands in specific structural features was large but consistent with the current conceptual model of Douglas-fir forest development. Diameter distributions generally exhibited a reverse-J shape, a characteristic of >450-year-old forests. Douglas-fir populated the tallest height classes and shade-tolerant species (e.g., western hemlock and Pacific silver fir) were present in lower (co-dominant and intermediate) canopy positions. Coarse woody debris was abundant in early old-growth stands in the form of both snags (42-140 m3ha[superscript -1]) and logs (172-584 m3 ha[superscript -1]). Scores for early old-growth stands calculated using existing old-growth structural indices were comparable to those in older (400 to 600year old) forests. The structural conditions and variability in these early old-growth forests are useful guides for managers seeking to accelerate development of complex structures in young Douglas-fir forests.