We monitored lead poisoning from the ingestion of spent lead pellets in sentinel mallards (Anas platyhrynchos) at the Sacramento National Wildlife Refuge (SNWR), Willows, California for 4 years (1986-89) after the conversion to steel shot for waterfowl hunting on refuges in 1986. Sentinel mallards were held in 1.6-ha enclosures in 1 hunted (P8) and 2 non-hunted (T19 and TF) wetlands. We compared site-specific rates of lead exposure, as determined by periodic measurement of blood lead concentrations, and lead poisoning mortality between wetlands with different lead pellet densities, between seasons, and between male and female sentinels. In 1986, the estimated 2-week rate of lead exposure was significantly higher (P < 0.005) in P8 (43.8%), the wetland with the highest density of spent lead pellets (>2,000,000 pellets/ha), than in those with lower densities of lead pellets, T19 (18.1%; 173,200 pellets/ha) and TF (0.9%; 15,750 pellets/ha). The probability of mortality from lead poisoning was also significantly higher (P < 0.01) in sentinel mallards enclosed in P8 (0.25) than T19 (0) and TF (0) in 1986 and remained significantly higher (P < 0.001) during the 4-year study. Both lead exposure and the probability of lead poisoning mortality in P8 were significantly higher (P < 0.001) in the fall of 1986 (43.8%; 0.25), before hunting season, than in the spring of 1987 (21.6%; 0.04), after hunting season. We found no significant differences in the rates of lead exposure or lead poisoning mortality between male and female sentinel mallards. The results of this study demonstrate that in some locations, lead exposure and lead poisoning in waterfowl will continue to occur despite the conversion to steel shot for waterfowl hunting.