An important topic of waterfowl ecology concerns the relationship between the physiological condition of ducks during the nonbreeding season and fitness; i.e., survival and future reproductive success. We investigated this subject using direct band recovery records of mallards (Anas platyrhynchos) banded in autumn (1 Oct-15 Dec) 1981-83 in the Mississippi Alluvial Valley (MAV). A condition index, weight (g)/ wing length (mm), was calculated for each duck, and we tested whether condition of mallards at time of banding was related to their probability of recovery during the hunting season. In 3 years, 5,610 mallards were banded and there were 234 direct recoveries. Three binary regression models were used to test the relationship between recovery probability and condition. Likelihood-ratio tests were conducted to determine the most suitable model. For mallards banded in autumn there was a negative relationship between physical condition and the probability of recovery. Mallards in poor condition at the time of banding had a greater probability of being recovered during the hunting season. In general, this was true for all age and sex classes; however, the strongest relationship occurred for adult males. J. WILDL. MANAGE. 50(2):177-183 In recent years many studies have focused on documenting the physiological condition of waterfowl, where condition often is defined poorly but generally refers to the quantity of energy reserves. This interest in condition results primarily from the assumption that individuals in good relative condition (i.e., those with high levels of energy reserves) exhibit greater probabilities of surviving and reproducing than individuals in poor condition. Investigations of waterfowl condition primarily have concentrated on: (1) estimating physical condition (Bailey 1979, Wishart 1979, Chappell and Titman 1983, Piersma 1984, Whyte and Bolen 1984b, and others); (2) documenting changes in physical condition in relation to various energetic costs such as reproduction (Harris 1970, Raveling 1979, Drobney 1982, Tome 1984), molt (Ankney 1979, 1984; Young and Boag 1982), and weather (Owen and Cook 1977, Bennett and Bolen 1978); and (3) relating physical condition to some measure of fitness such as breeding success (Ankney and MacInnes 1978, Krapu 1981, Cowardin et al. 1985) and time of pairing (Wishart 1983). There are relatively few empirical results, however, that support the assumption that being in good physical condition is advantageous. A noted exception occurs for waterfowl during the breeding season. Studies show that females of some waterfowl species depend on the accumulation of energy reserves (lipids, calcium, and protein) to breed successfully. For example, development of nutrient reserves before arriving at breeding areas (McLandress and Raveling 1981a,b; Gauthier et al. 1984) is important for the reproductive success of arcticn sting geese (Newton 1977, Raveling 1978). In some species of ducks fat reserves also are critical for development of initial clutches, but protein demands can be satisfied through exogenous sources (Siegfried et al. 1976, Krapu 1981, Drobney 1982). Winter generally is considered to be energetically stressful for waterfowl (Reinecke et al. 1982) and other species of birds that overwinter in temperate regions. Individuals often are faced with reduced food availability (Jorde et al. 1984) and severe weather conditions (Bennett and Bolen 1978) which may cause a negative energy balance (Whyte and Bolen 1984a). It seems reasonable, therefore, that individuals entering 'Present address: Savannah River Ecology Laboratory, Drawer E, Aiken, SC 29801.