The lunar crust: a product of heterogeneous accretion or differentiation of a homogeneous Moon?

The outer portion of the Moon including the Al-rich crust and the source regions of mare basalts was either accreted heterogeneously or was the product of widespread differentiation of an originally homogeneous source. A number of constraints make the heterogeneous accretion model unlikely; the differentiation model appears more plausible.If the differentiation model is correct, a series of cumulate rocks complimentary to the Al-rich crustal rocks must exist. The mare basalts may have been derived from such a complimentary cumulate for several reasons. For example, Philpottset al. (1973) on the basis of REE studies, suggest that Apollo 11 and 17 mare basalts were formed by partial melting of a cumulate rich in a phase(s) containing high Ti and heavy REE. The high Ti of Apollo 11 and 17 basalts is not readily explained in terms of partial melting of an undifferentiated mantle, but is consistent with partial melting of a pyroxene cumulate enriched in Fe, Ti oxides. The characteristic Fe-rich nature of mare basalts would be partly a consequence of melting of oxide cumulate minerals. It is postulated that the plagioclase-poor source region of mare basalts was enriched in an intercumulus residual liquid. During the partial melting that produced mare basalts, this material was largely incorporated into the melt, thus explaining the ancient model ages observed in most mare basalts. If the cumulate model is correct, then samples derived from the true (undifferentiated) lunar mantle have not been identified.