Annual survival probability estimates for songbirds are generally biased low because dispersal is not differentiated from mortality. Presently, knowledge of between-year breeding dispersal is lacking for most songbirds. To assess adult survival probabilities and dispersal, we color-banded and resighted 436 Yellow Warblers (Dendroica petechia) over five breeding seasons at 11 study sites in the Bitterroot Valley, Montana. During the last two of those seasons, we searched extensively for marked warblers between and surrounding those sites. We compared survival probabilities estimated with and without this added dispersal information and assessed the effectiveness of adjusting survival probabilities with transient-type models. Survival probabilities were calculated using open population models, and model selection was based on Akaike's Information Criterion (AIC) within program MARK. The best model indicated that survival probabilities differed between males and females and varied among years. We found that dispersal off the study site was common (in 1999, 30% of resighted birds were found off their original study site), and survival probabilities increased by 6.5–22.9% (0.02 ± 0.07−0.106 ± 0.06) with the inclusion of dispersed birds. We suggest that emigration can have substantial effect on survival probabilities and advise against the use of return rates from small study areas or spatial or temporal comparisons of return rates because of spatial and temporal variation in extent of emigration. In addition, our results suggest that additional assessments of the reliability of transient-type models under some sampling schemes may be warranted.