Records of magnetic field secular variation within Leg 172 cores were routinely recovered aboard ship by measuring the NRM of all cores (archive halves) at 5-cm spacing after 20-mT AF demagnetization. The 20-mT demagnetization removed a ubiquitous low-coercivity drill string magnetic overprint that has been noted previously during many ODP legs (e.g., Nagy and Valet, 1993; Weeks et al., 1993). Further AF demagnetization of selected core segments always showed good characteristic remanences that decayed toward the origin (Keigwin, Rio, Acton, et al., 1998). Stepwise AF demagnetization of discrete samples from selected horizons also displayed the same behavior (Keigwin, Rio, Acton, et al., 1998).
The patterns of directional variability observed after demagnetization could commonly be correlated between holes at individual sites for Sites 1060-1063. For example, inclination and declination variability at Site 1061 (Blake Outer Ridge; five holes) between ~15,000 and 45,000 radiocarbon years before present (BP) is shown in Figure F2; similar variability at Site 1063 (Bermuda Rise; four holes) is shown in Figure F3. The chronologies for these selected intervals are based on radiocarbon dates associated with other nearby piston cores (Keigwin and Jones, 1994) that were correlated to the Leg 172 cores using variations in magnetic susceptibility (Keigwin, Rio, Acton, et al., 1998; see also Schwartz et al., 1997). It is clear for both sites that selected inclination and declination features can be traced among the records from independent holes separated by distances of <1 km. The most diagnostic features are narrow intervals of relatively low or high inclination or easterly or westerly declination. In between these most easily correlated features, it is common to see more subtle patterns of variability that are also consistent between holes.
The directional secular variation in selected time intervals can also be correlated between sites up to 1600 km apart. For example, compare the inclination and declination variability within the same time interval (~15,000-45,000 radiocarbon years BP) as recorded at Site 1063 (Fig. F3) vs. Site 1061 (Fig. F2). The numbers that identify selected PSV features in Figures F2 and F3 are based on previously published piston-core studies (discrete sample measurements) from the same region (Lund, 1993; Lund et al., 1995, in press) summarized in Figure F4. Selected longer duration (104-105 yr) variations in the interval average inclination can also be correlated between sites (Keigwin, Rio, Acton, et al., 1998). Such variations are interpreted to indicate nonstationarity (time-dependent changes in the statistical character) of local secular variation.