Fig. 1. Location of Site 1202 in the Southern Okinawa Trough and path of the Kuroshio Current.

Fig. 2. Plot of calibrated ¹⁴C ages versus depth at Hole 1202B. Age depth values were fit by a third-order polynomial (black line). Age = 0.95987 + 375.05 Depth − 4.1649 Depth² + 0.027464 Depth³.

Fig. 3. Temperature-dependence of low-field magnetic susceptibility for a representative sample. The thermomagnetic behavior is characterized by low susceptibilities that increase at around 300 °C probably because of the formation of magnetite from sulfites.

Fig. 4. (a) Hysteresis loops show saturation fields <200 mT and coercivities typical for magnetite. (b) Hysteresis ratios display the pseudo-single domain state of the magnetic carrier (Day et al., 1977) and vary within a narrow range along a grain size mixing line.

Fig. 5. Lithology (pattern represents clayey silt) and downcore variations of magnetic parameters from 0 to 36 rmbsf at Hole 1202B.: (a) low-field magnetic susceptibility (k) measured on u-channels, (b) low-field magnetic susceptibility measured on whole cores directly after recovery, (c) ARM demagnetized at 20 mT, (d) median destructive field (MDF) of remanence, (e) kARM/k ratio, and (f) magnetic inclination. Concentration dependent parameters and grain size proxy remain fairly constant throughout the section.

Fig. 6. Vector component diagrams with normalized intensity decay plots of alternating-field (AF) demagnetization behavior for four representative samples from Holes 1202A and 1202B. Open (closed) symbols represent projections on the vertical (horizontal) plane.

Fig. 7. Comparison of the Hole 1202B paleosecular variation of magnetic inclination (21 point smoothed) with data from North America. Several large-scale features can be correlated across the records.

Fig. 8. (a) NRM/ARM after AF demagnetization of both NRM and ARM at peak fields of 20, 30, 40, 50, and 60 mT. (b) Mean values of NRM/ARM measurements and associated standard deviations. Error bars indicate ±1 standard deviation about each mean.

Fig. 9. NRM/k after demagnetization at peak fields of 20, 30, 40, and 50 mT.

Fig. 10. Comparison between the ARM and k normalized NRM. Both records are normalized to unity to allow direct comparison. The correlation between NRM/ARM and NRM/k is superb (r = 0.89).

Fig. 11. Power spectra of the paleointensity records, k, and ARM and coherence spectra between the two determinations of paleointensity and concentration-dependent rockmagnetic parameters. Neither of the paleointensity normalizations shows power at any frequency. ARM and k are dominated by millennial-scale cycles and the cross functions do not show power at the frequency of the ARM or k. Both paleointensity estimates are apparently free of any environmental signal.

Fig. 12. The relative paleointensity record at Hole 1202B. Data are ARM normalized. Relative paleointensity increased rapidly from 9000 to 8400 B. P., increased further from 6000 to 5300 B.P., and has been declining over the last 4500 years.

Fig. 13. Comparison of the relative paleointensity record from Hole 1202B with the Fish Lake, Oregon, record (Verosub, unpublished data). Also shown are two low resolution records, the dipole moment determined from archeological hearths (MS82) McElhinny and Senanayake (1982) and the NAPIS-75 paleointensity stack (Laj et al., 2004). OH93 is the synthetic record from Ohno and Hamano (1993). Arrows mark intensity lows that correlate across the records.