Patterns of migratory connectivity are a vital yet poorly understood component of the ecology and evolution of migratory birds. Our ability to accurately characterize patterns of migratory connectivity is often limited by the spatial resolution of the data, but recent advances in probabilistic assignment approaches have begun pairing stable isotopes with other sources of data (e.g., genetic and mark–recapture) to improve the accuracy and precision of inferences based on a single marker. Here, we combine stable isotopes and geographic variation in morphology (wing length) to probabilistically assign Wood Thrush (Hylocichla mustilena) captured on the wintering grounds to breeding locations. In addition, we use known-origin samples to validate our model and assess potentially important impacts of isotopic and morphological covariates (age, sex, and breeding location). Our results show that despite relatively high levels of mixing across their breeding and nonbreeding ranges, moderate levels of migratory connectivity exist along an east–west gradient. In addition, combining stable isotopes with geographic variation in wing length improved the precision of breeding assignments by 10% and 37% compared to assignments based on isotopes alone or wing length alone, respectively. These results demonstrate that geographical variation in morphological traits can greatly improve estimates of migratory connectivity when combined with other intrinsic markers (e.g., stable isotopes or genetic data). The wealth of morphological data available from museum specimens across the world represents a tremendously valuable, but largely untapped, resource that is widely applicable for quantifying patterns of migratory connectivity.