The objective of this study was to generate records of the mass accumulation rate (MAR) and lithologic composition of coarse sand-sized IRD deposited at Site 919 during the past 960 k.y. Samples were taken from the Site 919 cores at a spacing of ~50 cm, providing a temporal resolution of ~3-5 k.y. Samples were specifically chosen to avoid discrete ash layers and possible coarse-grained turbidites. The age of each sample was calculated using the sample's subbottom depth and the sedimentation rates derived from shipboard bio-stratigraphic and magnetostratigraphic data (Larsen, Saunders, Clift, et al., 1994), supplemented by the preliminary oxygen isotope stratigraphy of Flower (1998).

Samples were processed to provide both compositional and MAR data. In order to calculate the MAR of coarse sand-sized IRD for each sample, the concentration (in weight percent) of the coarse-sand IRD within the total sediment was determined as follows:

1. The sample was dried at 60ºC and weighed.
2. The sample was disaggregated in an ultrasonic bath, sieved at 250 µm and 2 mm to separate the coarse-sand fraction, and the coarse-sand fraction was weighed. The abundance of the coarse sand-size fraction (wt%) was then calculated using the weight of the total coarse-sand fraction and the initial dry sample weight.
3. The coarse-sand fraction was inspected through a binocular microscope to estimate the relative abundances of three grain types: IRD, biogenic components (radiolarians and foraminifers), and volcanic ash. In this examination, all coarse sand-sized grains that were not obviously either biogenic in origin or volcanic ash were assumed to be terrigenous IRD.
4. The weight percent of terrigenous coarse sand-sized IRD was calculated as the product of the weight percent of the total coarse-sand fraction and the relative abundance of IRD within the coarse-sand fraction. Visual inspection revealed that the majority of these samples are dominated by one grain type (i.e., either IRD, biogenic components, or volcanic ash). As a result, the temporal pattern of IRD importance could be illustrated without isolating IRD grains from samples with mixed grain populations. If more samples had contained such intermediate mixtures, however, the IRD would have been extracted by the appropriate use of chemical treatments (acid dissolution of carbonate and base dissolution of opal) and/or density separations (using sodium polytungstate).

The MAR of coarse sand-sized IRD for each sample was calcu-lated as follows:

IRD MAR = (%CS × %T) × LSR × DBD,

where

1. (%CS × %T) is the product of the weight percent of the total coarse sand-size fraction and the relative abundance of IRD within the coarse sand-size fraction (thereby giving the weight percent of terrigenous coarse sand-sized IRD in the total sample, as described above) at a particular level.
2. LSR is the linear sedimentation rate in that interval. Linear sedimentation rates were determined from shipboard biostratigraphic and magnetostratigraphic data (presented in Larsen, Saunders, Clift, et al., 1994), and were refined using the preliminary oxygen isotope stratigraphy of Flower (1998).
3. DBD is the dry bulk density of that interval (taken as the value from the nearest shipboard index properties sample).

The values used in this calculation for each sample, as well as the resulting IRD MAR values, are given in Table 1.

The composition of the IRD was determined by visual examination using a binocular microscope. If a sample contained >100 grains of IRD, then the compositions of 100 grains were used to calculate the abundances of the various grain types; if a sample contained <100 IRD grains, then all grains were counted. Grains were classified as quartz (monocrystalline + polycrystalline grains), basalt (black to dark green, fine-grained igneous grains), granite/coarse-grained acidic (coarse-grained, quartz-bearing, polycrystalline igneous and metamorphic grains), coarse-grained basic/coarse-grained mafic (black to dark green, coarse-grained igneous and metamorphic grains), sedimentary, and sedimentary carbonate. The abundances of each grain type, and the resulting MARs of each IRD grain type, are listed for each sample in Table 1.