During 1987-90, we used high-altitude photography, aerial videography, counts, and models to estimate sizes of breeding populations of dabbling ducks (Anatinae) and duck production and to identify duck habitat on U.S. Fish and Wildlife Service land and easements and on private land in the prairie pothole region of the United States. The study area contained about 3.1 million wetland basins (28,490 km2). Wetland area (ha per km2) was highest on service-owned land; wetland-basin density was greatest on service easements. Temporary and seasonal wetlands were underrepresented and lakes were overrepresented on service-owned land. Seventy-eight percent of all basins were less than 0.41 ha. Cropland dominated private land. Pond density decreased from 4.4/km2 in 1987 to 3.4/km2 in 1990 and pond area, from 7.2 ha/km2 to 2.7 ha/km2. The density of the blue-winged teal was greatest (3.4 pairs/km2) and was followed in magnitude by those of the mallard (2.1 pairs/km2), the gadwall (1.8 pairs/km2), the northern pintail (0.8 pairs/km2), and the redhead (0.8 pairs/km2). Duck density was consistently highest on service-owned land. The decline of breeding-population sizes in 1987-90 closely corresponded to losses of pond numbers and pond area. The density of breeding pairs per pond was inversely related to pond density, suggesting that breeding ducks tended to concentrate on the remaining ponds as drought intensified. The production of recruits followed the same pattern as breeding-population sizes. We estimated that 2.5% of the ducklings hatched on service-owned land, which was 1.3% of the study area; 19.6% hatched on service easements, which were 14.2% of the study area; and 77.9% hatched on private land, which was 84.6% of the study area. Various sources of bias and sampling error and improvements to the system are discussed.