Abstract. Directions of magnetization observed from rock-units from a given continental block are conveniently summarized by a path of apparent polar wander (apw) relative to that block. This article documents apw paths for the maj or continental blocks spanning the interval from Late Carboniferous (approximately -300 Ma) to the present. It is shown that by passing moving windowsof 30 or 40 Ma duration over the data an apw path can be obtained with little or no loss of features through smoothing. It also is shown that mean pole positions obtained using Fisher's statistics or by calculating modes do not differ significantly. In the analysis 'poor' results are screened out by applying certain minimum reliability criteria. Unit weight is then given to each of the remaining results. No attempt has been made to weight the results differentially because of the risk of added subjectivity.The paths contain many interesting features. One is an interval of rapid apparent polar motion relative to all major land-masses that occurred sometime in the Late Permian and Triassic. The general form of these motions is roughly established, but their timing and detailed form is poorly known because of inadequacies in the paleomagnetic record. Although it is clear that the motions occurred sometime around the Paleozoic-Mesozoic boundary it is not known whether the motions are or are not contemporaneous from continent to continent.Apw paths are of great importance for many problems in geology and geophysics. As an example, the new paths are applied to the question of how should Gondwana be correctly assembled. The results show that the best available reconstruction is the 1937 assembly of Alex du Toit as quantified by Smith and Hallam, with a minor modification by Barton and Molnar for the fit of Australia and Antarctica. Assemblies, that place Antarctica to the south and wrap the Antarctic Peninsula around the southern Andes, are not supported by the paleomagnetic data. There are, however, several obscurities, which derive, it is argued, from the uncertainties in timing the Late Permian to Triassic transitions alluded to above.

No abstract