Current estimates of annual survival, an important process affecting population dynamics, are lacking for breeding-waterfowl populations in Washington. I used hunting-season recoveries in conjunction with band-recovery models (program MARK) to estimate survival and recovery probabilities of Mallards (Anas platyrhynchos) and Gadwalls (A. strepera) banded in eastern Washington during 1981–1998. I also evaluated hypotheses about sources of variation in these rates and described the geographical and temporal distribution of band recoveries. Mallard survival and recovery probabilities were sex and age-specific, and recovery rates were year-specific but not strongly correlated with harvest regulations. Survival probability of Mallards was 0.661 for adult males, 0.660 for immature females, 0.606 for adult females, and 0.560 for immature males. Average recovery rates were generally highest for immature males (0.083) followed by adult males (0.050), immature females (0.050), and adult females (0.029). Survival and recovery rates of Gadwalls were 0.576 and 0.054, respectively, but sample size was small (52 recoveries from 436 banded birds) and sex-age classes were pooled prior to analysis. Seventy-two percent of Mallards banded in eastern Washington were recovered in the Columbia Basin of central Washington. The proportion of adult-male recoveries decreased in the Columbia Basin (from 75% in 1981–84 to 54% in 1995–98) and increased in California (from 4% to 22%). The distribution of direct recoveries of Gadwalls was similar to Mallards. My data suggest that annual survival of Mallards and Gadwalls banded in eastern Washington was similar to or slightly higher than survival probabilities in other North America populations.