Abstract
Studies of ophiolitic complexes and thermal and mechanical speculations indicate a highly complex history of basaltic magmas at mid‐ocean ridges. A thin molten region about 3 km wide probably exists at the top of the magma chamber. Primitive material enters the uppermost part of the chamber from below. Fractionated material is removed by the upward extrusion of dikes and settling of crystals to form a mush which probably fills the bulk of the chamber. The roof of the magma chamber is probably cooler than the melting temperature of primitive basalt. Secondary melts are probably produced most voluminously in the cumulate mush and also at the distal ends of the magma chamber and the bottom portions of dikes which cannot immediately solidify. The volume of the uppermost molten region can in principle be determined from chemical heterogeneities in basalts over a limited area and time. For mechanical reasons, the volume of additions of material from below is comparable to the volume of crustal dikes. Weaknesses along the line of dike intrusion and limited plating of material onto the roof of the chamber may restrict dike intrusion to the exceeding narrow zone inferred from ophioltic complexes.