Previous workers have used stratigraphic studies to identify three potential marine gateways that connected the Atlantic and Mediterranean during the Messinian salinity crisis (MSC): the Strait of Gibraltar that remains a 300-900-m-deep channel to the present-day and the Betic and Rifian corridors now exposed on-land in southern Spain and northern Morocco, respectively. Comparison of deepsea cores from the Atlantic and Mediterranean have shown that there was no significant or sea-level rise during the Messinian leaving a tectonic or climate control as the most likely cause for Messinian drying of the Mediterranean and that was followed in the early Pliocene by the re-flooding of Atlantic waters in the dessicated and evaporite-filled Mediterranean basin. In this study, we integrate bathymetric, GPS data from the Tangier Peninsula and its offshore areas with paleostress measurements at 25 sites ranging in age from Jurassic to Miocene. Offshore data from the Strait of Gibraltar indicates that the main ENE-lineament on the seafloor is a major right-lateral strike-slip fault whose sense is consistent with: 1) WNW-trending GPS vectors; 2) the arrangement of positive restraining and negative releasing bends; 3) formation of a 15-20-km-wide syncline within the Strait that deepened with continued compression; and 5) the right-lateral offset of the Mesozoic Calcaire Dorsale Ridge by ~7 km. Paleostress sites on-land in rocks of Oligocene to early Pliocene age indicate three events: 1) east-west compression of Miocene age inferred to record the formation and eastward motion of the Gibraltar arc; 2) NW-SE compression inferred to record the closure of Nubia and Iberia with compression of the Gibraltar arc; and 3) NE-SW compression inferred to represent continued compression of the Gibraltar arc that accompanied continued formation of the large syncline within the strait. We postulate that the offset of the highly resistant and 1-km-thick Calcaire Dorsale allowed the initial deep channels to open between the Atlantic and Mediterranean. We see no evidence for north-south-striking normal faults as postulated in strait-opening models based on roll-back of the Gibraltar slab.