Widespread, large-scale polymetallic W–Sn mineralization occurs throughout the Nanling Range (South China) dated 160–150 Ma, and related to widely developed coeval granitic magmatism. Although intense research has been carried out on these deposits, the relative contribution of ore-forming elements either from granites or from surrounding strata is still debated. In addition, the factors controlling the primary metallogenic element in any given skarn deposit (e.g., W-dominated or Sn-dominated) are still unclear. Here, we select three of the most significant skarn-deposits (i.e., Huangshaping W–Mo–Sn, Shizhuyuan W–Sn–Mo–Bi and Xianghualing Sn), and compare their whole-rock geochemistry with the composition of associated granites and strata. The contents of Si, Al and most trace elements in skarns are controlled by the parent granite, whereas their Fe, Ca, Mg, Mn, Ti, Sr and REE patterns are strongly influenced by the wall rock. Samples from the Huangshaping skarn vary substantially in elemental composition, probably indicating their varied protoliths. Strata at the Shizhuyuan deposit exerted a strong control during metasomatism, whereas this occurred to a lesser degree at Huangshaping and Xianghualing. This correlates with increasing magma differentiation and increasing reduction state of granitic magmas, which along with the degree of stratigraphic fluid circulation, exert the primary control on dominant metallogenic species. We propose that wall rock sediments played an important role in the formation of W–Sn polymetallic mineralization in South China.

Phase equilibrium experiments were used to determine conditions of melt evolution and phenocryst growth in high-Mg andesite magmas that were erupted at Whakaari (White Island) in New Zealand between 1976 and 2000. The high-Mg andesites are both mafic (7.21–10.3 wt% MgO) and silica-rich (55.3–58.6 wt% SiO₂) with phenocrysts of plagioclase, orthopyroxene, clinopyroxene, olivine, Cr-spinel and Fe–Ti oxides contained in a glassy to fine-grained matrix of mostly dacitic composition. Experiments were conducted on one of the most primitive samples available (the high-Mg andesite TRW34) at conditions ranged from 1 atm to 500 MPa at temperatures of 950 to 1200 °C with total water concentrations of 0 to 10 wt%. Except for the 500 MPa experiments, ƒO₂ was buffered at 1 or 2 log units above Ni–NiO. Consistent with earlier thermodynamic modelling, our results demonstrate that residual Whakaari melts (now represented by matrix glasses) evolved along a plagioclase + two-pyroxene cotectic (± magnetite ± ilmenite) under comparatively low-pressure, shallow conditions (< 200 MPa or < 6 km) and were relatively hot (> 950 °C) and dry (≤ 3 wt% melt-H₂O), with oxygen fugacities either at, or slightly above Ni–NiO + 1 log unit. Although the bulk-rock trends of Whakaari volcanic rocks are clearly calc-alkaline, those of the residual matrix glasses are only weakly so. A likely explanation for this contrast is that the primary magmas were relatively hydrous, but became dehydrated when intruded at shallow depths. The effectiveness of water in this role, combined with the demonstrable presence of primitive calc-alkaline magmas in the upper-crust, highlights the importance of magmatic water, in place of deep crustal fractionation, for shaping the calc-alkaline evolutionary trend.

Whakaari (White Island) is an active andesite-dacite volcano located on New Zealand’s northernmost continental shelf. During a series of Strombolian eruptions in 1976–2000 it produced high-Mg andesite magmas (Mg#: 65–74; SiO₂: 55–58 wt. %). It has already been suggested that the production of these magmas involved complex interaction between stored and ascending magmas in a mid-crustal magma chamber that forms part of a larger trans-crustal plumbing system. Here we re-examine this proposal by employing Rhyolite-MELTS 1.2.0 and Magma Chamber Simulator to simulate liquidus relationships for one of the 1976–2000 high-Mg andesites from Whakaari (P41600). It was found that production of the main phenocryst assemblage (olivine + Cr-spinel + orthopyroxene + clinopyroxene + plagioclase + magnetite), mineral compositions, and liquid line of descent (as determined from matrix glasses) requires 30–60% fractional crystallisation at comparatively low pressures (< 100 MPa) and melt-H₂O concentrations (< 2 wt. %) with moderate fO₂ (from Ni-NiO to one log unit above Ni-NiO) and temperatures of 1140°C to 1000°C. At least 0.5 wt.% water is required to stabilise olivine at 60 MPa although original magmatic water concentrations may have been significantly higher. The early loss of magmatic water is a predictable consequence of both the low pressures of fractionation and the partitioning of H₂O into a CO₂-rich vapour phase. Polybaric crystallisation does not improve the simulations when compared to the isobaric model. Considering these observations, it is unlikely that evolution of the Whakaari high-Mg andesites involved significant fractionation within the mid- to lower-crust. However, there are aspects of magma evolution (particularly for FeO_T and TiO₂) that are not well modelled by any of our simulations. This may be because our model could not accurately simulate the effects of mixing between magmatic products or interaction between evolved melts and earlier crystallized mineral phases.