Predictions of the time-allocation hypothesis were tested with several a posteriori analyses of banding data for the mallard (Anas platyrhynchos). The time-allocation hypothesis states that the critical difference between resident and migrant birds is their allocation of time to reproduction on the breeding grounds and survival on the nonbreeding grounds. Residents have higher reproduction and migrants have higher survival. Survival and recovery rates were estimated by standard band-recovery methods for banding reference areas in the central United States and central Canada. A production-rate index was computed for each reference area with data from the U.S. Fish and Wildlife Service May Breeding Population Survey and July Production Survey. An analysis of covariance was used to test for the effects of migration distance and time period (decade) on survival, recovery, and production rates. Differences in migration chronology were tested by comparing direct-recovery distributions for different populations during the fall migration. Differences in winter locations were tested by comparing distributions of direct recoveries reported during December and January. A strong positive relationship was found between survival rate, and migration distance for 3 of the 4 age and sex classes. A weak negative relationship was found between recovery rate and migration distance. No relationship was found between production rate and migration distance. During the fall migration, birds from the northern breeding populations were located north of birds from the southern breeding populations. No pattern could be found in the relative locations of breeding and wintering areas. Although our finding that survival rate increased with migration distance was consistent with the time-allocation hypothesis, our results on migration chronology and location of wintering areas were not consistent with the mechanism underlying the time-allocation hypothesis. Neither this analysis nor other recent studies of life-history characteristics of migratory and resident birds supported the timeallocation hypothesis.