The Pangea configuration for the interval of Late Carboniferous to Early Jurassic is still debated: was it similar to the well known Wegenerian Pangea or did Gondwanaland move with respect to Laurussia along a major right-lateral lineament of several thousand kilometres during the Triassic, as suggested by Trying? We try here to answer this important question with the paleomagnetic study of well dated Permo-Triassic sediments from a borehole in Saudi Arabia. Sediments of the same age have also been sampled at the surface, but proved to be remagnetized, probably by a strong desert alteration. The paleomagnetic study of the bore-hole core yielded a high temperature component magnetization thought to be of primary origin carried either by magnetite (Late Permian limestones) or hematite (Early Triassic red shales) with antipodal normal and reversed inclinations. Its co-latitude and pole directions compare well with those from a selection of South American, African, Madagascar, and Moroccan poles. A new reliable pole for west Gondwana with a mean age of 244 +/- 11 m.y. is derived (lambda = 53.4 degrees N, phi = 259.4 degrees E, A(95) = 3.6 in West African coordinates), acid its comparison with the Laurussian poles of same age window strongly suggests a Pangea in which the northern part of South America was facing the east of North America (Pangea B of Irving) at the Permo-Triassic boundary, A large right-lateral strike-slip movement of some 3500 km during the Triassic is thus required to reconstruct Pangea to its Jurassic pre-Atlantic opening position.