The Mesoproterozoic (1.47–1.40 Ga) Belt basin in western North America contains a thick succession of siliciclastic strata dominated by sediment derived from source areas that lay to the west and southwest of the present basin. The age and identity of this source area is important for reconstruction of the Mesoproterozoic history of western North America. Over 400 new SHRIMP U-Pb ages for detrital zircons, supplemented by Ar-Ar ages of detrital muscovite from sandstones within the Belt basin, document distinctive provenances of the basin fill; the predominant sediment influx came from a western craton, whereas subordinate input came from Laurentian sources to the east and south of the basin. During deposition of the lower two-thirds of the Belt basin, the western source was composed predominantly of Proterozoic rocks with 1920–1460 Ma ages; abundant grains have 1610–1500 Ma ages. These latter ages are unrepresentative of Laurentian sources, falling within the North American magmatic gap (1610–1490 Ma), but are widespread in central and eastern Australia and Baltica, the former being the most likely candidate for a former fluvial connection to western North America. The presence of detrital-zircon grains (1480–1440 Ma) with ages contemporary with sedimentation in the Belt basin suggests that the western source area hosted active magmatism during Belt sedimentation. The Missoula Group in the upper Belt records the sedimentologic effects of basin reorganization, which introduced sediment from a new southwestern source that, in contrast to earlier Belt history, resembles crust of the southwestern United States and lacks grains within the North American magmatic gap. Evidence for extension is widespread in the Belt basin in the form of mafic volcanic rocks, rapid rates of sedimentation, and intrabasinal syndepositional faults. The predominant contribution of detrital material from the western source area throughout Belt deposition, in concert with the large-scale facies patterns and evidence for rapid rates of subsidence, is taken to suggest that the Belt basin represents an extensional basin with a tectonically active western side. Although a wholly extensional environment has been suggested by previous workers, we suggest that an alternative model, wherein the Belt basin developed within an extensional domain along a collisional/convergent plate margin, analogous to Tethyan sedimentary basins such as the Black and Caspian Seas, is equally viable and paints a different picture of western North America at ca. 1450 Ma.