Whole-rock geochemistry and Sr–Nd isotope data are presented for a 500-m-thick section at Winnaarshoek in the Eastern Limb of South Africa’s Bushveld Complex. Extending from the Upper Critical Zone, including the UG2 and UG3 chromitites and the Merensky and Bastard Reefs, into the Lower Main Zone (LMZ), which has a bimodal lithostratigraphy. The regionally persistent anorthosites and associated norites are intercalated with subordinate pyroxenites. Lithological units are therefore descriptive rather than assigned a genetic origin. Across the UG2 to Merensky Footwall interval, feldspathic rocks show nearly constant initial Sr isotope ratios (SrR0) over 400 m, broadly overlapping the more variable pyroxenite values. SrR0 values are higher in the Merensky–Bastard Reef pyroxenites than in the UG2–UG3 pyroxenites, increase stepwise upward through the anorthosites capping the Merensky–Bastard units, and reach their highest levels in the LMZ. In contrast, ƐNd remains tightly constrained throughout the profile. This decoupling of the Sr and Nd isotope systems, together with Sr-isotopic differences between anorthosites and pyroxenites, supports intrusion as thin magma sheets rather than differentiation within a high-level magma chamber. The pyroxenites and chromitites are interpreted as composite sills emplaced along contacts between pre-existing noritic and anorthositic strata. The magma sheets retained isotopic heterogeneity acquired as their parental magmas evolved in a deep-crustal staging chamber. Up-section isotope trends indicate increasing crustal contamination or local assimilation of an isotopically enriched contaminant. Interpreting these layered igneous rocks as a stacked sill complex directly links chromitites and PGE mineralisation to the staging chamber. Pyroxenitic sill magmas transported chromite slurries and PGE from depth. This model explains both the metre- to centimetre-scale alternation of mineralised chromitites and pyroxenitic reefs with compositionally similar, PGE-poor layers and the lateral persistence of richly mineralised chromite stringers.

Deserts must be supplied with sediment in order to accrete. The Thar Desert, lying east of the Indus River in South Asia, might be expected to be largely supplied with sediment from that drainage. In this study, we use a combination of major and trace element bulk-sediment geochemistry, together with Sr and Nd isotopes, to constrain the provenance of postglacial dune sand. Our data indicate a stronger influence from mafic source rocks in the Sindh Desert compared to that in Cholistan. Nd isotopes imply sediment was largely derived from the lower Indus River during the early and pre-Holocene post-glacial time. The sand is coarser grained in Sindh and retains higher ϵNd values in sediment that eroded from mafic rocks in Kohistan and the Karakorum as a result of deflation of deltaic and floodplain areas in the lower reaches by southwesterly summer monsoon winds. The composition of Cholistan dunes, like that in the Eastern Thar Desert, reveals instead more supply from Himalayan sources and more negative ϵNd values. The greater Himalayan influence in Cholistan and the Eastern Thar Desert largely reflects finer grain size, a result of the longer transport from the delta source and a preference for more Himalayan supply in the form of finer sediment.