Dramatic changes in the food habits and distribution of canvasbacks (Aythya valisineria) in Chesapeake Bay during the 1970's generated a need to evaluate the nutritional value of food items and the requirements of this species on its wintering grounds. Groups of captive canvasbacks were maintained ad libitum on 5 diets during the winters of 1978-79 and 1979-80 to evaluate the effects of varying protein and energy levels on feed intake and condition. Feed intake during the 1979-80 winter was 42% greater for those ducks fed the low energy (1,543 kcal/kg) diet than for those fed the high energy (3,638 kcal/kg) diet. Canvasbacks fed the high energy diet, however, consumed 317 kcal/bird day, whereas those fed the low energy diet consumed only 191 kcal/bird day. Body weight of males and females did not differ among groups fed different diets, but there were seasonal differences (P < 0.05) for both sexes aggregated across diets. Data from this study indicate that canvasbacks may be unable to adjust intake rates to compensate for low energy foods and subsequently may store less fat or modify behavior. However, decreased weight, feed intake, and activity of ducks fed ad libitum rations occurred in mid-winter irrespective of diet quality and appeared to be an endogenous component of their annual cycle that persists in captivity. These changes apparently have a selective advantage of increasing the probability of survival in ducks by decreasing energy expenditure during periods of winter stress. J. WILDL. MANAGE. 50(3):427-434 Aerial winter waterfowl surveys conducted by the U.S. Fish and Wildlife Service from 1955 to 1979 indicated that the winter population of canvasbacks in North America has ranged from 481,000 in 1955 to 179,000 in 1972 (Perry et al. 1981). The winter population in Chesapeake Bay (Md. and Va.) averaged 31% of the North American population and 62% of the Atlantic Flyway population. During 1955-79, numbers of canvasbacks declined substantially in Chesapeake Bay, the Atlantic Flyway, and North America (Perry et al. 1981). During the 1970's in the Atlantic Flyway, the percent of canvasbacks wintering in Chesapeake Bay decreased, whereas percentage increases occurred in New Jersey and North Carolina. Increases in New Jersey remain unexplained (Perry 1982), but increases in North Carolina may be related to better habitat conditions due to a larger number of freshwater impoundments (Perry and Uhler 1982). These wetlands provide abundant aquatic vegetation, which has traditionally been an important winter food for canvasbacks. Vegetation surveys from 1971 to 1979 showed Present address: 7 Crestwood Drive, Tolland, CT 06084. a decline in the frequency of occurrence and biomass of submerged aquatic plants in Chesapeake Bay (Kerwin et al. 1976, Bayley et al. 1978, Orth and Moore 1981, Haramis and Carter 1983). Turbidity and herbicides have been implicated in the decline of vegetation (Stevenson and Confer 1978). Two exotic aquatic plants, waterchestnut (Trapa natans) and spiked watermilfoil (Myriophyllum spicatum) were introduced into Chesapeake Bay during the 1900's and competed with native food plants (Gwathney 1945, Rawls 1975, Bayley et al. 1978). Some areas of the bay became completely covered by exotics, which restricted growth of wildcelery (Vallisneria americana) and other submerged native plants. Native waterfowl plant foods declined in abundance during the exotic plant infestations but recovered with the demise of the exotics. Waterfowl abundance on the Susquehanna Flats, in the upper bay, seemed to be inversely related to the abundance of exotic plants (Bayley et al. 1978). Habitat changes in Chesapeake Bay were reflected in a precipitous decline in vegetation as a food of canvasbacks, which occurred concurrently with an increase of mollusks as a food item (Munro and Perry 1981, Perry et al. 1981). The dramatic change in canvasback food habits