The results of numerous palæomagnetic measurements made on rocks during the past decade have been interpreted according to various schemes based upon either or both of the assumptions of polar wandering and the relative movements of continents. Recently, however, certain criticisms have been directed against this approach and a new analysis has been made to reassess the available information, relying on as few preconceptions as possible about the eventual explanation of the data. It can be demonstrated that for older rocks there is a well defined deviation of the measured remanent magnetic vectors from the present geomagnetic field, which increases progressively with geological age. For the four continents in which the results are best attested this deviation shows a systematic trend, and some genuine phenomenon is established. A combination of polar wandering together with continental drift provides an explanation of the results to date. On the other hand, on the present evidence, the possibility cannot be ignored that the observed effects may be due to strong multipolar components in the ancient magnetic field of the earth: alternatively it is conceivable although improbable that they can be explained by some widespread geological activity such as tectonic stresses of continental extent. The most plausible conclusion appears to be that continental drift has taken place, and on this tentative assumption the magnitudes of the land movements can be estimated from the data now available.