We have reviewed paleomagnetic data available for the Eurasian, African, North American and Indian plates over the last 200 Ma. Selection criteria are those generally accepted, with an emphasis on evidence for lack of remagnetization, accurate dating and proper structural analysis. This results in 23, 35, 51 and 2 poles for Eurasia, Africa, North America and India, respectively. We believe that this limited set of higher quality data is more likely to reveal key features of apparent polar wander (APW) paths than averaging of larger data sets involving less stringent selection criteria. We propose and describe revised APW paths, but more importantly, we next use relative motion models to transfer all data in a common reference frame.

We find good agreement between transferred data, when they are averaged separately for each plate in independent 20-Ma windows. This is a check on consistency of paleomagnetic data. Kinematic models, and of the geocentric dipole hypothesis. Transferred data from all plates are averaged in 20-Ma windows to generate synthetic APW paths for all plates studied. These synthetic paths are in agreement with the original (revised) APW paths that use only data from a single plate. Moreover, both geographic and time resolution are improved (spatial confidence intervals are of the order of 5¿). The synthetic paths display interesting features, such as a previously ill-recognzed APW loop for Eurasia. Paleomagnetic and hotspot APW are next compared, and a determination of true polar wander (TPW) is derived. We find significant TPW, amounting to over 20¿ in the last 200 Ma. TPW appears to be episodic, with a standstill between 180 and 110 Ma. There is general agreement with a previous study of Livermore et al. (1984). The TPW standstill appears to correlate with a time of decreasing reversal frequency, ending with the Cretaceous Long Normal Superchron. Other periods of fast TPW would seem to correspond to increasing reversal frequency. However, it is suggested that the major TPW hairpin at 50 Ma might correspond to the collision of India with Eurasia. Tentative correlations with core or mantle indicators can be understood as a result of couplings between the core, mantle and lithosphere, we believe primarily related to episodic evolution of the D' layer. ¿ American Geophysical Union 1991