The geologic formations drilled through Unit S3 at Sites 1097 and 1103 are composed mainly of massive clast-rich diamict. At Site 1103, stratified diamicts, sandstones, and siltstone indicate deposition on a glacially influenced slope. At Site 1097, there is a wider range of sedimentary environments from subglacial deformation till to proglacial sediment gravity flow to open marine deposition (Fig. F3). Recovery was generally poor.
Following the smear slide descriptions provided by shipboard sedimentologists (Shipboard Scientific Party, 1999b, 1999c), we reexamined the diamict matrix at 10 locations that had volcanic glass contents reported above 5%. With the exception of a single sample at Site 1097 that contains up to 75%-80% volcanic glass, all other samples contain 5% to 10% glass by visual estimation in smear slides. All samples were soaked in water for several months in an attempt to soften the matrix before disaggregation. The disaggregation was then done by hand in a mortar, by gentle wet crushing after 2 min of ultrasonic bath. Size separation was done by wet and dry sieving at sizes >2 mm, 2 mm-210 µm, and 210-150, 150-75, and 75-63 µm. Not all the resulting samples were suitable for further treatment because of the difficulty in separating individual volcanic clasts and because often volcanic clasts resulted in too poor concentration. Finally, we selected three samples, 178-1097A-27R-1, 35-58 cm, 178-1103A-31R-2, 0-4 cm, and 178-1103A-36R-3, 4-8 cm, which were treated with a Frantz Isodynamic separator and heavy liquids to concentrate homogeneous grains. The highest concentration of homogeneous grains was obtained in the range of 100-150 µm for all samples. Petrographical investigation of concentrates mounted with epoxy resin was carried out using a Philips XL30 scanning electron microscope (SEM) operating at 20 kV equipped with an X-ray energy dispersive system (EDS) EDAX DX4. Glass analyses were carried out using a JEOL JX 8600 electron microprobe (EMP) fitted with four wavelength-dispersive spectrometers (accelerating voltage = 15 kV; sample current = 10 nA; and beam size = 5-10 µm). Natural standards were used for calibration.
SEM-EDS observations revealed that all three sample concentrates consist of mineral clasts (including quartz, plagioclase, minor alkali-feldspar, and rare clinopyroxene, amphibole, biotite, and epidote), unsorted mineral aggregates (Fig. F4A), and abundant fragments of volcanic rock groundmass. The latter range from ipocrystalline to microcrystalline rocks. Textures vary from hyalopilitic at Site 1103, with variable proportions of plagioclase and well-preserved glass (Fig. F4B, F4C, F4D), to "intergranular" at Site 1097, with plagioclase laths and unresolvable alteration products possibly replacing interstitial glass (Fig. F4E, F4F). Fresh glass was not recognized in the sample from Site 1097. Plagioclase from all three samples ranges in composition from andesine to labradorite. Minor clinopyroxene and Fe-Ti oxide microlites occur in some groundmass grains from Site 1103 (Fig. F4D). Individual clasts are subrounded to rounded. Because some rounded clasts include both volcanic groundmass and a fine-grained glaciogenic matrix (e.g., Fig. F4B), we infer that roundness is due to mechanical reworking.
Table T1 lists the electron microprobe data on groundmass glass of samples from Site 1103. Most of the analyzed glasses fall in the subalkaline field of a total alkali vs. silica diagram (Fig. F5), some of them at the boundary between the subalkaline and alkaline and a few in the alkaline field. They range compositionally from trachybasalt to trachyte and from basalt to dacite (Fig. F5). Present data do not allow a clear definition of their affinity (i.e., orogenic vs. anorogenic). However, the relatively high TiO2 and P2O5 contents (Table T1) point to a mature arc and/or to an anorogenic setting for some grains. The Ca/K ratio is variable and largely overlaps in the two samples (Table T1). It shows a negative correlation with SiO2. The Cl content is relatively high (generally higher than 0.05 wt%). Cl/K ratios range from 0.033 to 0.095 in Sample 178-1103-36R-3, 4-8 cm, and from <0.008 to 0.12 in Sample 178-1103-31R-2, 0-4 cm (Table T1).