Improvement in the accuracy of late Paleozoic correlations requires close integration of a broad spectrum of methods and concepts. Comparison of fossil zonation and radiometric ages shows about a dozen ages from upper Paleozoic rocks that are reliable and helpful in delimiting ages for system, series, or stage boundaries, and most of these ages have some uncertain stratigraphic relations to fossil zones. Ages of 345 m.y. for Devonian-Mississippian, 320 m.y. for Mississippian-Pennsylvanian, 280 m.y. for Pennsylvanian-Permian, and 224 m.y. for Permian-Triassic boundaries are reasonable approximations. In order to establish more precise fossil zonations between faunal provinces and to relate fossil zones to radiometric ages in upper Paleozoic strata, we need to know much more about phylogeny and changes in paleogeography, their effect on faunal dispersal, their structural causes, and changes in other environmental factors. For example, comparison of lithologic correlation and fossil zonation shows fusulinacean species commonly range through several cyclothems and suggests rapid transgression and regression of biofacies in cyclothems. Comparison of worldwide zonations based on ammonoids, corals, and foraminiferids indicates development of semi-isolated and isolated faunal provinces, particularly during middle Chesterian through Desmoinesian time, Virgilian time, late Wolfcampian through middle Guadalupian time, and Dzhulfian time (latest Permian).