Is plagioclase removal responsible for the negative Eu anomaly in the source regions of mare basalts?

The nearly ubiquitous presence of a negative Eu anomaly in the mare basalts has been suggested to indicate prior separation and flotation of plagioclase from the basalt source region during its crystallization from a lunar magma ocean (LMO). Are there any mare basalts derived from a mantle source which did not experience prior plagioclase separation? Crystal chemical rationale for REE substitution in pyroxene suggests that the combination of REE size and charge, M2 site characteristics of pyroxene, fO2, magma chemistry, and temperature may account for the negative Eu anomaly in the source region of some types of primitive, low TiO2 mare basalts. This origin for the negative Eu anomaly does not preclude the possibility of the LMO as many mare basalts still require prior plagioclase crystallization and separation and/or hybridization involving a KREEP component.