METHODS

Samples from Holes 1002A and 1002C were obtained by pushing an oriented square-section stainless steel tube into the split core face. The resulting minicore was withdrawn and extruded into a 5-cm³ plastic box. Samples from Hole 1001A were obtained postcruise by cutting ~2-cm-long quarter-round samples from the working half using a nonmagnetic ceramic blade in a rock saw. The relatively large quarter-round samples were subsequently trimmed by hand with a titanium carving knife to fit into an AF demagnetizer.

Sample processing for magnetic measurements consisted of the following steps:

1. Measurements of the natural remanent magnetization (NRM) of all samples using a two-axis, 2G Enterprises cryogenic magnetometer.
2. Stepwise AF demagnetization with a DTECH demagnetizer and subsequent remeasurement of samples with the magnetometer.
3. Measurement of low field (0.1 mT) volume magnetic susceptibility (K) at a driving field frequency of 0.47 kHz with a Bartington Instruments susceptibility meter and MS1B dual-frequency sensor.
4. Demagnetization of all samples at 90 mT to remove the NRM followed by the acquisition of anhysteretic remanent magnetization (ARM) by subjecting samples to a gradually decreasing peak AF of 100 mT in the presence of a 0.1-mT direct current (DC) field. The ARM was measured using the cryogenic magnetometer. The ARMs are expressed as an anhysteretic susceptibility (K_ARM), obtained by dividing by the strength of the DC field.
5. All samples were given a saturation isothermal remanent magnetization (SIRM) in a 1-T DC field, measured with the cryogenic magnetometer, subjected to a reversed DC field of 0.3 T, and then remeasured. The two measurements were used to calculate HIRM ("hard" IRM), which is defined as (IRM _{-0.3 T} + SIRM)/2, and S, which is defined as -IRM_{-0.3 T}/SIRM.

In the following we briefly outline the rock-magnetic significance of these parameters. The K, SIRM, K_ARM, and NRM usually reflect variations in the concentration of ferrimagnetic minerals (i.e., members of the magnetite-ulvospinel and magnetite-maghemite solid solution series, henceforth described as "magnetite"). However, K and SIRM are biased toward coarser grained magnetite and K_ARM and NRM toward finer grained magnetite. The NRM also partly reflects the intensity of the ambient geomagnetic field at around the time of sediment deposition. The HIRM is proportional to the concentration of such coercivity, antiferromagnetic minerals as goethite and hematite.

For magnetic mineral assemblages dominated by magnetite, the ratio K_ARM/K is indicative of magnetite grain size within the magnetically stable fraction; that is, grains that span the range of stable single-domain through multidomain behavior (e.g., Hartstra, 1982a, 1982b; King et al., 1982; Ozdemir and Banerjee, 1982). The S values reflect the ratio of goethite/hematite to magnetite, with lower values indicating an increase in this ratio.

Age modeling of Hole 1001A data was done by comparison to the Leg 165 polarity time scale (Sigurdsson et al., 1997).