A method of making paleocontinental reconstructions based on ocean floor magnetic anomaly data and land-based paleomagnetic data is briefly described, and its limitations are assessed. Synthetic polar wander paths for the period 20--200 Ma are presented for Australia, Antarctica, India, Africa, Eurasia, North America, and South America. Comparison is made with observed paths, except for Antarctica, for which the data are too sparse. There are no significant differences between observed and synthetic paths. A plot of inclination anomaly against paleolatitude shows a large scatter that limits the precision of the paleogeographic grid on continental reassemblies. The absence of any systematic trend to this scatter with time suggests that the field has had the same average behavior over the past 180 Ma and that large-scale expansion of the earth cannot have occurred in this period. South America has undergone the least change in position relative to the poles in the past 160 Ma.