The earliest evolution of the Moon likely included the formation of a magma ocean and the subsequent development of anorthositic flotation cumulates. This primary anorthositic crust was then intruded by mafic magmas which crystallized to form the lunar highlands magnesian suite. The present study is a compilation of petrologic, mineral‐chemical, and geochemical information on all pristine magnesian‐suite plutonic rocks and the interpretation of this data in light of 18 “new” samples. Of these 18 clasts taken from Apollo 14 breccias, 12 are probably pristine and include four dunites, two norites, four troctolites, and two anorthosites. Radiogenic isotopic whole rock data also are reported for one of the “probably pristine” anorthositic troctolites, sample 14303,347. The relatively low Rb content and high Sm and Nd abundances of 14303,347 suggest that this cumulate rock was derived from a parental magma which had these chemical characteristics. Trace element, isotopic, and mineral‐chemical data are used to interpret the total highlands magnesian suite as crustal precipitates of a primitive KREEP (possessing a K‐, rare earth element (REE)‐, and P‐enriched chemical signature) basalt magma. This KREEP basalt was created by the mixing of ascending ultramafic melts from the lunar interior with urKREEP (the late, K‐, REE‐, and P‐enriched residuum of the lunar magma ocean). The trace and major element compositions of nearly all magnesian‐suite cumulates can be generated by 0–55% fractional crystallization of a primitive KREEP basalt combined with the trapping of varied proportions (generally ≤20%) of instantaneous, residual, KREEP‐basalt liquid. A few samples of the magnesian suite with extremely elevated large‐ion lithophile elements (5–10× other magnesian‐suite rocks) cannot be explained by this model or any other model of autometasomatism, equilibrium crystallization, or “local melt‐pocket equilibrium” without recourse to an extremely large‐ion lithophile element‐enriched parent liquid. It is difficult to generate parental liquids which are 2–4× higher in the REE than average lunar KREEP, unless the liquids are the basic complement of a liquid‐liquid pair, i.e., the so‐called “REEP‐fraction,” from the silicate liquid immiscibility of urKREEP. Scarce age information on lunar rocks suggests that magnesian‐suite magmatism was initiated at progressively more recent time from the northeast to the southwest on the lunar nearside from 4.45 to 4.25 Ga. This magmatic “event” could be due to melting of the lunar mantle beneath these regions and could have been generated either by latent heat during crystallization of the final, KREEP‐rich (and, thus, Th‐ and U‐rich), residual, lunar magma ocean liquid or heating due to radioactive decay of K, Th, and U.