The decline of freshwater mussels in the southeastern US emphasizes the need to evaluate the current status of mussel populations. We used the Robust Design, which is a capture–recapture sampling design, to estimate demographic parameters (apparent survival and temporary emigration) and capture probabilities of Alasmidonta arcula, Lampsilis dolabraeformis, Lampsilis splendida, and Pyganodon gibbosa in a large lowland river in Georgia. Mussels were sampled in individual habitat units using line-transect methods at ∼6-wk intervals from summer 2006–2007. We used an information-theoretic approach to evaluate the relative importance of maximum river discharge, habitat characteristics, mussel species, and season on temporary emigration (i.e., proportion of mussels not at the surface), apparent survival, and capture probability. The best-supported models indicated that apparent survival and capture probability varied positively with mussel shell length and among habitat types. Apparent survival (6-wk interval) ranged from 94 to 99% and was greatest in slackwater and lowest in swiftwater habitat. Capture probability ranged from 8 to 20% and was greatest in slackwater and lowest in swiftwater habitat. Temporary emigration also varied among species and season and appeared to be related to reproductive behavior, with the largest proportion of mussels occurring at the surface when mussels appeared to be reproductively active. A comparison of catch-per-unit-effort indices to population estimates suggested that the reliability of catch-per-unit-effort indices was influenced by vertical migration behavior and other factors affecting mussel capture probability.