The mobility of Morocco relative to Africa during the early history of the Atlantic has long been debated: arguments have been developed from Atlantic kinematic considerations and from paleomagnetic results. Both types of arguments are reexamined here. Using a new model of the Atlantic Ocean evolution described elsewhere, it is shown that the reconstructions of the positions of Africa relative to North America before the Atlantic opening, and at the times of magnetic anomalies M22 and J, do not imply a major motion of Morocco independent of Africa during these periods of time. The corresponding geomagnetic paleopoles have been recomputed from sample sites located on both "mobile" Morocco and "stable" Africa. The results indicate that the virtual geomagnetic pole of "mobile" Morocco for the Liassic falls within the 95% confidence cone of "stable" Africa. It is thus concluded that no major movement has occurred between "mobile" Morocco and "stable" Africa during the early phases of opening of the central Atlantic Ocean. This is in accordance with the field geological observations on the South Atlas fault; however, limited motion along this lineament, as observed in the field, is still compatible with the above conclusion, owing to the limited resolving power of both kinematic and paleomagnetic methods.