Summary: Experimental data in the system CaO-MgO-FeO-SiO₂ suggest that there may be a plateau on the liquidus and solidus of the multicomponent system basalt-peridotite. If this is so, fusion of peridotite would produce only basaltie magmas over a wide temperature range; when the temperature reached a value such that the liquid crossed the threshold of the plateau, there would be a rapid increase in the amount of fusion for small temperature increases, with the formation of picritic magmas; basaltic magmas containing suspended forsteritic olivine crystals could dissolve them if the temperature rose slightly above that of the plateau threshold; a high proportion of a picritic magma would crystallize in a small temperature interval, with the precipitation of forsteritic olivine that was only slightly zoned. These possibilities are compared with current theories, and it is concluded that several petrological axioms may require critical re-examination. An experimental procedure is outlined to determine the shape of the liquidus and solidus in the basalt-peridotite system.