(Supplemet S1) Silicon oxide and Potassium + Sodium oxide from the PACManus hydrothermal area

This study presents a systematic analysis and interpretation of autonomous underwater vehicle-based microbathymetry combined with remotely operated vehicle (ROV) video recordings, rock analyses and temperaturemeasurements within the PACManus hydrothermal area located on Pual Ridge in the Bismarck Sea of eastern Manus Basin. The data obtained during research cruise Magellan-06 and So-216 provides a framework for understanding the relationship between the volcanism, tectonismand hydrothermal activity. PACManus is a submarine felsic vocanically-hosted hydrothermal area that hosts multiple vent fields locatedwithin several hundredmeters of one another but with different fluid chemistries, vent temperatures and morphologies. The total area of hydrothermal activity is estimated to be 20,279m**2. Themicrobathymetrymaps combinedwith the ROV video observations allow for precise high-resolution mapping estimates of the areal extents of hydrothermal activity.We find the distribution of hydrothermal fields in the PACManus area is primarily controlled by volcanic features that include lava domes, thick andmassive blocky lava flows, breccias and feeder dykes. Spatial variation in the permeability of local volcanic facies appears to control the distribution of venting within a field.We define a three-stage chronological sequence for the volcanic evolution of the PACManus based on lava flow morphology, sediment cover and lava SiO2 concentration. In Stage-1, sparsely to moderately porphyritic dacite lavas (68-69.8 wt.% SiO2) erupted to form domes or cryptodomes. In Stage-2, aphyric lava with slightly lower SiO2 concentrations (67.2-67.9 wt.% SiO2) formed jumbled and pillowed lava flows. In the most recent phase Stage-3, massive blocky lavaswith 69 to 72.5wt.% SiO2were erupted throughmultiple vents constructing a volcanic ridge identified as the PACManus neovolcanic zone. The transition between these stages may be gradual and related to progressive heating of a silicic magma following a recharge event of hot, mantle-derived melts.

40Ar-39Ar investigation of volcanic clasts in glaciogenic sediments at ODP Sites 1097 and 1103

Three selected diamictite samples recovered within sequence group S3 at Sites 1097 (Sample 178-1097A-27R-1, 35-58 cm) and 1103 (Samples 178-1103A-31R-2, 0-4 cm, and 36R-3, 4-8 cm) of Ocean Drilling Program Leg 178 have been investigated by scanning electron microscope, electron microprobe, and 40Ar-39Ar laser-heating techniques. They contain variable proportions of fragments of volcanic rock groundmass (mostly in the range of 100-150 µm) with textures ranging from microcrystalline to ipocrystalline. Their rounded shapes indicate mechanical reworking. Fresh groundmass glasses, recognized only in grains from samples of Site 1103, show mainly a subalkaline affinity on the basis of total alkali-silica variations. However, they are characterized by relatively high TiO2 and P2O5 contents (1.4-2.8 and 0.1-0.9 wt%, respectively). Because of the small size of homogeneous grains (100-150 µm), they were not suitable for single-grain total fusion 40Ar-39Ar analyses. The incremental laser-heating technique was applied to milligram-sized samples (only for Samples 178-1097A-27R-1, 35-58 cm, and 178-1103A-36R-3, 4-8 cm) and to various small fractions (each consisting of 10 grains for the sample from Site 1097 and 30 grains for samples from Site 1103). The latter approach resulted in more effective resolution of sample heterogeneity. Argon ages from the small fractions show significantly different ranges in the three samples: 75-173 Ma for Sample 178-1097A-27R-1, 35-58 cm, 18-57 Ma for Sample 178-1103A-31R-2, 0-4 cm, and 7.6-50 Ma for Sample 178-1103A-36R-3, 4-8 cm. Ca/K ratios derived from argon isotopes at Site 1103 suggest that the data mainly refer to outgassing of groundmass glass. At Site 1103, we observe an overall apparent age increase with decreasing sample depth. This is compatible with glacial erosion that affected with time deeper levels of a volcanic sequence previously deposited on the continent. The youngest apparent age of 7.6 ± 0.7 Ma detected close to the bottom of Hole 1103A (340 meters below seafloor [mbsf]) is compatible with the age range of the diatom Actinocyclus ingens v. ovalis Zone (6.3-8.0 Ma) determined for the interval 320-355 mbsf and with the maximum ages derived from strontium isotope composition of barnacle fragments obtained at 262-263 mbsf at the same site. Nevertheless, this age cannot be taken as the maximum youngest age of the volcanic sequence sampled by glacial erosion or as the maximum age for the deposition of the Sequence S3 at 340 mbsf unless validated by further research.

Summary of phase compositions of experimental silicate melt, un-normalised analyses

The solubility of Re and Au in haplobasaltic melt has been investigated at 1673-2573 K, 0.1 MPa-2 GPa and IW-1 to +2.5, in both carbon-saturated and carbon-free systems. Results extend the existing, low pressure and temperature, dataset to more accurately predict the results of metal-silicate equilibrium at the base of a terrestrial magma ocean. Solubilities in run-product glasses were measured by laser ablation ICP-MS, which allows for the explicit assessment of contamination by metal inclusions. The Re and Au content of demonstrably contaminant-free glasses increases with temperature, and shows variation with oxygen fugacity (fO2) similar to previous results, although lower valence states for Re (1+, 2+) are suggested by the data. At 2 GPa, and Delta IW of +1.75 to +2, the metal-silicate partition coefficient for Re (DMet/Sil) is defined by the relation LogD[met/sil][Re] = 0.50(±0.022)*10**4/T(K)+3.73(±0.095) For metal-silicate equilibrium to endow Earth's mantle with the observed time-integrated chondritic Re/Os, (and hence 187Os/188Os), DMet/Sil for both elements must converge to a common value. Combined with previously measured DMet/Sil for Os, the estimated temperature at which this convergence occurs is 4500 (±900) K. At this temperature, however, the Re and Os content of the equilibrated silicate is ~100-fold too low to explain mantle abundances. In the same experiments, much lower Dmet/sil values have been determined for Au, and require the metal-silicate equilibration temperature to be <3200 K, as hotter conditions result in an excess of Au in the mantle. Thus, the large disparity in partitioning between Re or Os, and Au at core-forming temperatures argues against their mantle concentrations set solely by metal-silicate equilibrium at the base of a terrestrial magma ocean.

(Table 1) Geochemistry of glass shards at DSDP Site 87-582 and 87-583

In the Leg 87A holes, 45 ash layers were sampled in Recent to upper Pliocene strata. The main volcanogenic deposits came from single eruptions or subcontemporaneous eruptions of cognate volcanoes. Some of them are mixed ashes produced from multiple eruptions and accumulated in reworked sediments. The petrographic and geochemical patterns indicate rhyolitic and dacitic compositions; andesitic glasses are scarce. We infer a magmatic affinity with calc-alkaline sources and a possible origin from the volcanic arc of southwestern Japan. A few samples may originate from the alkaline volcanism of southwestern Japan or the area south of Korea and the Sea of Japan.