Early debate on MORB petrogenesis focused on a low-pressure (<1.1 GPa, basaltic primary melts) vs. high-pressure (1.5-3.0 GPa, picritic primary melts) origin. This unresolved issue was later supplanted by the widely accepted model of Klein & Langmuir (1987) involving polybaric melting columns of varying length, with the initial pressure of melting ranging from 1.2 to 4.0 GPa (Langmuir et al., 1992). The KL87 and other models (e.g. McKenzie & Bickle, 1988) rely on phase relations free of H₂O and CO₂. For H₂O, this appears to be a good approximation at all pressures because of the low H₂O concentration in the MORB source (~90-230 ppm) and the probability that these concentrations are less than the bulk solubility of H₂O in nominally anhydrous mantle phases. At low pressures, CO₂ can also be neglected, to first order, because of its low solubility in melts. However, at P > ~2.3 GPa, CO₂ dramatically modifies solidus phase relations and abruptly lowers solidus temperatures by ~ 300°C (e.g. Falloon & Green, 1989; Dalton & Presnall, 1998). Presnall et al. (2002) addressed the issue of CO₂ by using phase relations in the CaO-MgO-Al₂O₃-SiO₂-Na₂O-FeO (CMASNF) system at 1-1.5 GPa (plagioclase/spinel lherzolite transition interval) to model the major-element characteristics of MORBs and phase relations in the CaO-MgO-Al₂O₃-SiO₂-CO₂ system to produce very small amounts of carbonatitic melts at 2.3-7 GPa, which would mix with the more shallow melts and produce some of the trace element signatures. Asimow et al. (1995, 2001) showed the importance of near-isentropic melting and found that strong variations in isentropic melt productivity occur as a function of pressure. Although Asimow et al. found from MELTS calculations that melting is suppressed by isentropic decompression through the plagioclase/spinel lherzolite transition, Presnall et al. showed that the CMASNF experimental data require enhanced melting. KL87 explained the global inverse correlation of Fe(8) with Na(8) by large variations of potential temperature (1260-1530°C) in a relatively homogeneous peridotitic mantle, but Presnall et al. explained them by a much narrower range of potential temperatures (average ~ 1260°C) in a heterogeneous peridotitic mantle. Thus, the original disagreement over a low- vs. high-pressure origin for primitive MORBs has been transformed into a debate over the magnitude of potential temperature variations and the role of mantle heterogeneity. The outcome of this debate has important implications for temperature variations in the Earth's interior and for geodynamics.