Of 250 thin sections examined for sediments and rocks recovered from Leg 180, 119 sand/sandstone samples were point counted from Sites 1108, 1109, 1114, 1115, 1116, 1118, and 1112. In an attempt to eliminate compositional differences between samples arising from factors other than changes in provenance, most samples point counted were of medium- to coarse-grained sand grade (average grain size = 0.250-1.00 mm). Where such samples were unavailable, material in the fine sand range was used. All samples were air dried, set in epoxy, and ground to 30 µm in thickness. To minimize further uncertainties associated with grain-size variation, the Gazzi-Dickinson point-counting technique was employed (e.g., Dickinson, 1970; Zuffa, 1980; Ingersoll et al., 1984). The above approach differs from more traditional methods in that mineral grains of sand size (>0.0625 mm) that occur within larger rock fragments are counted as the minerals concerned rather than as larger rock fragments. The Gazzi-Dickinson method divides volcanic lithic fragments into various textural types (Dickinson, 1970), which is especially significant in magmatic arc settings where volcaniclastic components may constitute a large percentage of the sand fraction. Further divisions within the volcanic clasts were made on the nature of the intersertal material and glass color under plane-polarized light.
Reworked sedimentary rock clasts from within the formation were not counted in the modal analyses. Five hundred points per thin section were counted using an automatic digital point counter to determine the proportions of the components. Grid spacings were used that maximized the coverage of the thin sections and were larger than the maximum grain size. For modal analysis purposes, grains were divided into various compositional categories. These and derived parameters used in this study are listed in Table T1 and discussed below. Raw point-counted data and derived parameters for all samples are listed in Tables T2 and T3.
The nature and chemistry of minerals was confirmed on selected polished sections by electron probe microanalysis. This being a methodology not directly inherent to the purpose of this specific paper, the microprobe data are not reported or discussed here.