We investigated the dynamics of a wintering population of Black-capped Chickadees (Parus atricapillus) using data from a long-term capture-recapture study. Goodness-of-fit and likelihood-ratio tests indicated that the standard Jolly-Seber model was inadequate for the data, so we explored different parameterizations of a model in which survival probability for new captures differed from that of previously marked birds. One possible explanation for the lower local survival rate of new captures is that some of the birds were transients that had no chance of being recaptured. Average survival probability over the entire period was 0.62, and the estimated fraction of transients among unmarked birds was 0.27. We found evidence that chickadee survival rates were lower in years immediately following establishment of the Tufted Titmouse (Parus bicolor) as a territorial species (after 1967) than during years immediately preceding this event. We found no evidence that chickadee survival rates were lower in the years immediately following establishment of territorial raptors (after 1984) than during preceding years. Evidence suggested a long-term decline in survival probability from 1959 to 1991.