Apatite in most igneous intrusions has a high Cl/F ratio. However, chlor-apatite has been reported in the lower portions of the Bushveld and Stillwater Complexes. This has been used as evidence supporting the early separation of a Cl-rich discrete hydrous fluid in these intrusions. Mineralogical evidence is presented here to demonstrate that the Bushveld Complex, at least, formed from a nearly anhydrous magma, and did not release a hydrous fluid before apatite began to crystallize. It is suggested that apatite in the earliest cumulates equilibrated with trapped interstitial liquid, which converted it from the typical F-rich composition of cumulus apatite to a Cl-rich composition. This is an analogous process to that in which cumulus mafic minerals may become more Fe-rich on cooling and reaction with interstitial liquid.

The geology of the 2.7 Ga Stillwater Complex in South-Central Montana is reviewed with a focus on the role of volatiles in locally modifying both the crystallization sequence of the evolving parent magma and the initially precipitated solid assemblages to favour olivine ± chromite. A secondary origin for these two minerals is particularly probable for the olivine-bearing rocks of the Banded series and, at a minimum, also increasing their modal abundance in the Peridotite zone of the Ultramafic series. Direct evidence for volatiles includes the presence of high-temperature fluid inclusions in pegmatoids and hydrous melt inclusions (now crystallized) in chromite and olivine from both the Ultramafic and the Banded series rocks. Indirect evidence includes the boninitic character of the parent magma, the presence of volatile-bearing minerals including high-temperature carbonates, rock textures, and Cl / F variations in apatite. Mechanisms which favour the formation of olivine (± chromite) over pyroxene include volatile phase boundary shifts induced by added H2O, incongruent melting of pyroxene by hydration of a partly-molten mush, and the near- to sub-solidus replacement of pyroxene by olivine and chromite by silica-undersaturated fluids. These mechanisms cast doubt that magmas with different liquid lines of descent were involved in the crystallization of the Stillwater Complex. A dry Stillwater magma would have been mineralogically and modally much less varied and lacking in high-grade platinum-group element and chromium deposits.