Luna 24 ferrobasalt as a low-Mg primary melt

Luna 24 very-low titanium (VLT) ferrobasalts, metabasalts, brown glasses and impact melts form a tight compositional cluster with no gradation to other groupings postulated for the Luna 24 core components. This suggests that the Luna 24 VLT ferrobasalt was extruded as a liquid of its own composition and was not derived by fractional crystallization from a more magnesian parent in a surface flow. Furthermore, the characteristics of the core lithologies are not easily visualized as components of such a differentiated flow, e.g. brown glasses. Gravitative settling models purporting to demonstrate the validity of the flow differentiation model are merely permissive. Subsurface fractionation requires that plagioclase, not olivine, be the liquidus phase. The high-Mg component in the Luna 24 core can be constrained, though not identified, chemically, and it has neither the major element, trace element, isotopic, nor mineralogical characteristics required of a possible parent to the Luna 24 VLT ferrobasalt. Thus models of fractionation lack a physical expression of the less differentiated compositions, contrary to the belief that the high-Mg component in the core is the parent material. The Luna 24 VLT ferrobasalt is probably a primary low-Mg melt from a plagioclase-bearing source region, and may have undergone little or no fractionation prior to eruption. Such a model is compatible with, and suggested by, chemical and experimental data. Caution against postulating that all Mg-poor melts are fractionated products, based on terrestrial models, is advised. The terrestrial oceanic situation of 'primary melts' with similar Mg/Fe is probably not valid for the Moon.