During 17 years of study at the Woodworth, North Dakota study area, the percent of 548 wetland basin with water during 1-15 May ranged from 8 to 87 and averaged 56; waterfowl pair densities varied from 19 to 56/km2 and averaged 40/km2. Pond occupancy by duck pairs averaged 37% during mid-May counts and 48% for late May and early June counts. A positive linear relation occurred between the estimated number of duck pairs and the percent of basins with water during 1-15 May.There were 3,339 duck nests found in grassland habitats from 1966 through 1981. Approximately 66% (85% Mayfield) of these were depredated or abandoned. Mammals caused 88% of nest failures. Half or more of the eventually successful clutches were unhatched by 10 July in 9 of 16 years. Haying would have disturbed or destroyed an average of 43%, 33%, 22%, 15%, and 9% of the duck nests if initiated on 10 July, 15 July, 20 July, 25 July, and 1 August, respectively.The total average size of completed clutch for all species was 29% smaller at the end of the nesting season than at the beginning, underscoring the importance of protecting early clutches.Production averaged 30 broods per 100 pairs of ducks and ranged from 15 to 61 broods per 100 pairs. Brood densities ranged from 10 to 63/km2 and averaged 12/km2. Mean brood size averaged 6.4 for all species. July broods averaged 7.2 ducklings and August broods 5.7 ducklings. Duckling loss averaged 2.6 per brood and 85% (2.2 ducklings) of this loss was estimated to occur during the first 14 days after hatch.Wetlands of all sizes and classes were important at some time to one species of duck or another. With the exception of some diving ducks, all species used a complex of sizes and classes of wetlands for space, food, and shelter necessary for nesting and brooding. Pair counts during 20 May-7 June were most indicative of the breeding population. A combination of two brood counts resulted in the best estimate of annual production. An average of only 50% of the total duck broods per year was counted during the 1-15 July surveys, which approximated the average time of the Service's July aerial surveys. During this study the area produced an average of 1 duck per 4 ha of upland and had a nest density of approximately 1 nest per 14 ha. Nest success rates averaged 35.1% (16.3% Mayfield). Predation was significantly reduced by good vegetative cover at nest sites. Seeded grasslands (dense nesting cover) yielded better production than native prairie or croplands. Seeded grasslands also produced 3 times more ducklings per unit area than adjacent native prairie and more than 14 times as many as adjacent, annually tilled croplands.Ducks generally showed higher nest densities and better nesting success when using growing grain crops than when nesting in standing or mulched stubble fields. Among native mixed-grass prairie and seeded grassland, production was enhanced by leaving fields idle or by treating them with periodic burning. Duck production was generally lowered by grazing field of native prairie but duck production on grazing lands was higher than in annually tilled croplands.