Samples were wet sieved to remove clay and dried in an oven before impregnation and thin sectioning. Thin sections were etched with concentrated hydrofluoric acid and stained with sodium cobaltinitride (yellow stain after K-feldspar) and amaranth (pink stain after calcic plagioclase). Albite is etched but not affected by staining and is distinguished from quartz by twinning and alteration in untwinned crystals. Staining also aids identification of silicic volcanic rock fragments and was particularly significant in the deeper samples from Site 1177. Where possible, sands with medium to coarse grain sizes were selected for point counting (39 slides from a total of 95 slides), but the remaining samples with fine to very fine grain size were examined. Stratigraphic setting of modally analyzed samples is shown on Figure F2. The Gazzi-Dickinson method of point counting was followed, where sand-sized crystals included in lithic fragments are counted as the mineral component to overcome the effect of grain size (Ingersoll et al., 1984). This method has the advantage that the results from different grain sizes can be compared and is particularly relevant given the variation in grain sizes between the different sites. A total of 500 points of sand grains were counted for each thin section. The numbers of raw points for each category can be calculated by multiplying the percentages in Table T1 by five. Although counted, carbonate, organic material, and intraformational clay debris were not included in the 500-point total (raw counts are shown in Table T1). Volcanic rock fragments with both lathwork and microlitic textures (Dickinson, 1970) are present, but given the abundance of compositionally indeterminate mafic to intermediate fragments at Site 1174 this classification as used in some previous studies (Marsaglia et al., 1992) was not attempted.