The considerable palaeomagnetic data from the Gondwanic continents, Africa, Antarctica, Australia, India and South America, are critically examined and shown to support the hypothesis of continental drift. Palaeomagnetic latitudes for Australia, traced by Irving and his colleagues from the late Precambrian to Recent times, indicate that Australia was near the equator in the Devonian and close to the Pole during the Permo-Carboniferous. During the Mesozoic and Tertiary it drifted northwards to its present position. Gough and his colleagues in Southern Rhodesia have recently concluded that since the middle of the Mesozoic the palaeomagnetic latitude of Africa has remained appreciably unchanged, but they find evidence of a marked drift northwards to its present latitudes from polar latitudes during the Permian and early Mesozoic. From my own studies of South American rocks I have deduced that there has been little movement of South America relative to the pole since the Triassic-Jurassic, but that a sharp change in magnetic latitude took place during the Lower and Middle Permian. Results from Devonian and Silurian rocks indicate that in those times northeast Brazil was closer to the pole than Tierra del Fuego. The movement of South America relative to the pole during the Permian was thus a continuation of an Upper Palaeozoic polar shift. In the Triassic-Jurassic, basaltic lavas and diabase dykes were extruded and intruded into parts of all the Gondwanic continents. Palaeomagnetic studies have been made on these rocks from all five continents. The palaeolatitudes and palaeoazimuths so deduced are consistent neither with the present positions of these continents, nor with the suggestion that they were then adjacent to one another. A possible reconstruction satisfying the restrictions imposed by the palaeomagnetic data shows these continents occupying positions between those suggested by geologists for the Permo-Carboniferous and their present positions, and it is inferred that the continents as we know them today had separated and had started moving towards their present positions when this igneous activity occurred. For the Palaeozoic era reliable palaeomagnetic data have, as yet, been obtained only for the Devonian, Silurian and Cambrian of Australia and South America. There is one not very well established Silurian result for South Africa. The consistency of these data with reconstructions of Gondwanaland based on geology is examined. Both du Toit's and J. T. Wilson's reconstructions are considered.