A population of American woodcock (Scolopax minor) was studied on a 3,401-ha area of the Moosehorn National Wildlife Refuge in northeastern Maine from 1976 through 1985. During 1976-83, from 4 to 64 clearcuts were created each year, opening up large contiguous blocks of forest. A combination of mist nets, ground traps, nightlighting techniques, and trained dogs were used to capture and band 1,884 birds during the first 5 years. Capture and recapture data (totaling 3,009 observations) were used with both demographically closed and open population models to estimate population size and, for open population models, summer survival. Flying young, especially young males, represented the greatest proportion of all captures; analysis showed that young males were more prone to capture than young females. Male courtship began about 24 March each year, usually when there was still snow in wooded areas. Males ~2 years old dominated singing grounds during April each year, but this situation changed and first-year males dominated singing grounds in May. Singing males shifted from older established singing grounds to new clearcuts soon after we initiated forest management. Many males were subdominant at singing grounds despite an abundance of unoccupied openings. Three hundred adult females were captured and, except for 1978, the majority were ~2 years old. The year in which female homing rate was lowest(1979) was preceded by the year with the largest number of l-year-old brood female captures and a summer drought. Summer survival of young was lowest in 1978 and was attributed to summer drought. The year 1979 had an abnormally cool and wet spring, and was the poorest for production of young. Capture ratios of young-to-adult females obtained by nightlighting could be used to predict production on our study area. Closed population model estimates did not seem to fit either young or adult data sets well. Instead, a partially open capture-recapture model that allowed death but no immigration seemed to fit best. Only the number of males in the population changed significantly during the study. An increase from 88 males in 1976 to 156 in 1980 was attributed to habitat management. Singingmale surveys on our area detected little change in the number of singing males, but our independent population estimates from mark-recapture data showed a larger total male population by 1980. Annual density estimates for all age and sex classes ranged from 19 to 25 birds/l00 ha. A hypothesis on the breeding system of the American woodcock is presented as well as a discussion of management implications, including the importance of creating high-quality habitat on private lands.