The recent application of stable-isotope analyses, particularly the use of stable-hydrogen-isotope (deltaD) measurements of animal tissues, has greatly improved our ability to infer geographic origins of migratory animals. However, many individual sources of error contribute to the overall error in assignment; thus likelihood-based assignments incorporating estimates of error are now favored. In addition, globally, the nature of the underlying precipitation-based deltaD isoscapes is such that longitudinal resolution is often compromised. For example, in North America, amount-weighted expected mean growing-season precipitation deltaD is similar between the boreal forest of southwestern Canada and areas of northern Quebec/Labrador and Alaska. Thus, it can often be difficult to distinguish objectively between these areas as potential origins for broadly distributed migrants using a single isotopic measurement. We developed a Bayesian framework for assigning geographic origins to migrant birds based on combined stable-isotope analysis of feathers and models of migratory directions estimated from band recovery data. We outline our method and show an example of its application for assigning origins to a population of migrant White-throated Sparrows (Zonotrichia albicollis) sampled at a Canadian Migration Monitoring Network station at Delta Marsh, Manitoba, Canada. We show that likelihood-based assignments of geographic origins can provide improved spatial resolution when models of migration direction are combined with assignments based on deltaD analysis of feathers.