The working halves of the cores containing Units I, II, and III were sampled for paleomagnetic analyses using a tool shaped for standard Japanese paleomagnetic "cubes" (octagonal horizontal section; faces parallel to a 2-cm square are 1.5 cm wide; missing 0.25 cm at each corner of the square). The sediment samples were placed in standard ODP sample cubes (2 cm x 2 cm x 2 cm), but instead of a sample volume of 8 cm3, these were missing ~0.0625 cm3 from each corner, for a total sample volume of 7.75 cm3. The standard ODP magnetic direction conventions were used: the x-axis of the sample is positive downward from the split face of the working half of the core; the y-axis is positive to the right when looking upcore and parallel to the split face of the core; the z-axis is positive downcore and parallel to the split face of the core. Sample cubes are marked with an arrow that points upcore (–z).
Every core was sampled at an average sampling rate of one sample per 0.75 m of core available for sampling. Each identified magnetic anomaly was sampled at least once. All of the samples were measured for natural remanent magnetization (NRM) and were alternating-field (AF) demagnetized by steps of 5, 10, 15, 20, 25, 30, 35, 40, 50, 60, and 70 mT. Inclination, declination, and maximum angular deviation were determined from principal component analysis of Zijderveld plots of the discrete sample AF-demagnetization measurements. These analyses were carried out on the ship and the data are accessible through the Janus database.