Primary stratiform platinum-group element (PGE) mineralizations associated with chromitites in the Critical zone of the Bushveld Complex, South Africa, are confined to cumulate layers that are regionally persistent over tens of kilometers. Four categories ofchromitite are recognized: layers at the bases of cycles in the Lower Critical zone (type I), layers at the bases of cycles in the Upper Critical zone (type II), thin layers in the intermediate parts of cycles (type III), and stringers associated with orthopyroxene pegmatoids (type IV). These latter may not be of primary magmatic origin, nevertheless, probably each type of chromitite contains anomalous concentrations of PGE.Chromitites at the bases of cycles, comprising the Lower Group (LG), Middle Group (MG), and Upper Group (UG), are further categorized into types Ia (LG-1/LG-4), Ib (LG-5/MG-1), IIa (MG-2/UG-1), and IIb (UG-2 and above) on the basis of lithostratigraphy, chromite and PGE chemistry, and sulfide content. Type IIb chromitites contain significant quantities of sulfides and are not investigated here. New whole-rock and electron microprobe chemical analyses of the sulfide-poor type Ia, Ib, and IIa chromitites (Cu contents <50 ppm), however, record regular variations with height that may elucidate our understanding of the economically important PGE layers. Type Ia chromitites occur in harzburgite-pyroxenite cycles, record Cr/Fe ratios >1.8, low PGE contents (<1,000 ppb), and (Pt + Rh + Pd)/(Ru + Ir + Os) ratios <1. Type IIa chromitites occur in pyroxenite-norite-anorthosite cycles and record Cr/Fe ratios <1.5, moderate PGE contents (1,000-5,000 ppb), and (Pt + Rh + Pd)/(Ru + Ir + Os) ratios >2. Type Ib chromitites occur in pyroxenite cycles and form an intermediate group.These regular variations with increased height are interpreted as a differentiation trend. PGE are concentrated in sulfide-poor chromitites as a result of chromitite control, a process not fully understood but thought to involve a combination of direct nucleation of platinum-group minerals and localized S saturation. It is the crystallization of copious quantities of chromite from hybridized melts close to the crystal-liquid interface that triggers this process, but magma mixing, essential to drive the melt into the chromite field, is the overall cause. Replenishment of the magma chamber was from a single magma batch, because the chromitites below the UG-2 are not associated with major isotopic inflections. Huge volumes of dense hot magma were fed as basal flows along the crystal-liquid interface. The magma chamber was compartmentalized at an early stage, possibly in response to basement tectonics, and lateral mixing over distances of > 150 km is an alternative mechanism for achieving sufficiently high R values.Repeated replenishment of the magma chamber resulted in mixing with resident liquid that became increasingly differentiated with height. Type Ia chromitites are ascribed to replenishment and mixing with the relatively primitive Lower zone residue, type Ib crystallized as a result of mixing with Lower Critical zone resident liquid, and type IIa, due to mixing with more evolved Upper Critical zone resident liquid. Differentiation of PGE with height is a critical part of the process which culminates in the occurrence of higher grade PGE mineralization in sulfide-bearing chromitites (UG-2, Merensky reef) in the uppermost part of the Upper Critical zone.