RESULTS

A large (~50%–80% of the NRM) vertical drilling overprint was typically removed by 15 mT or 200°C (Figs. F2, F3). The remaining remanence decays to the origin of a vector endpoint diagram for most samples. The AF demagnetized samples, however, often appear to acquire a remanence at steps >40 mT that is not compensated for by the double-demagnetization procedures. This typically occurs when the remanence is less than a few percent of the NRM, and we disregard these steps in the calculation of a principal component. Results from thermally demagnetized samples are consistent with those from AF demagnetized samples (Fig. F3).

The remanence direction was calculated by principal component analysis (Kirschvink, 1980) for steps from 15 to 40 mT (five to seven points) for most samples. A few higher coercivity samples (e.g., Fig. F2B) were calculated at higher levels. Directions with a maximum angular deviation (Kirschvink, 1980) >15° (e.g., Fig. F2H) or samples that did not decay to the origin of the vector endpoint diagram (e.g., Fig. F2G) were rejected. The remaining inclinations were used, along with shipboard pass-through data, to determine polarity.

In most, but not all cases, the discrete sample inclination agreed in polarity with the pass-through data, but often provided a less ambiguous answer (Figs. F4, F5). See, for example, results from Chron C28r in Hole 1262C at ~210 meters composite depth (mcd), or Chron C23 at ~100–118 mcd (Fig. F4). Declination data suffer from bias (Fig. F6), though perhaps not to the same degree as the shipboard data (Zachos, Kroon, Blum, et al., 2004). Whereas the shipboard declinations from archive halves clustered around 0° in core coordinates, the discrete sample declinations from working halves tend to cluster around 180° (Fig. F6). This suggests some kind of radial overprint (e.g., Fuller et al., 1998). Because the declinations do not appear to provide meaningful results, we will focus discussion on the inclination data.

Inclination data remain quite scattered above the P/E boundary; below the boundary, scatter is suppressed, as in the pass-through cryomagnetometer data, and many of the data seem to cluster around the expected GAD value of 53°–56°. However, some inclinations appear excessively steep and may be related to an incompletely removed vertical drilling overprint.

Over the critical C24r/C24n boundary at Site 1262 (Fig. F4), the discrete sample inclinations show significant scatter, yet all three holes show a transition from reversed to normal polarity at ~115–116 mcd. This significant revision to the shipboard interpretation, which placed the boundary at ~120 mcd, was reported in Lourens et al. (2005). The discrete samples also allowed identification of this same boundary at Site 1267 (~201–207 mcd), although with less precision (Fig. F5). At Site 1266, the discrete sample inclinations were highly scattered and did nothing to help identify the C24r/C24n boundary.

At several boundaries in the Paleocene or Late Cretaceous, discrete samples resolved some ambiguities and allowed us to constrain most reversals to within ~10–30 cm at Sites 1262 and 1267. However, for the upper parts of the cores, the coarsely spaced sampling did not significantly resolve any magnetostratigraphic issues, and we do not plot these data. Principal components for all discrete samples are listed in the Table AT1. Tables T1, T2, T3, T4, T5, and T6 list all magnetostratigraphic picks and highlight boundaries that are new or changed from the Leg 208 Initial Reports volume (Zachos, Kroon, Blum, et al., 2004).

Hysteresis data were difficult to obtain on the weakly magnetized, carbonate-rich sediments. After removing the predominant diamagnetic signal, a few relatively noise free loops were obtained. Hysteresis parameters are summarized in a typical Day plot (Day et al., 1977) in Figure F7A and are reflective of multidomain (MD) to pseudo-single-domain size grains. Alternatively, on a squareness (M_r/M_s) vs. coercivity plot as in Tauxe et al. (2002), samples plot on a mixing trend between MD and vortex-state grains (Fig. F7B).

Results from the Lowrie three-axis IRM experiments can be summarized as being of three different types (Fig. F8). Type 1 sediments (Fig. F8A) are found above ~98 mcd in Hole 1267A and are characterized by a large soft (0.1 T) component that shows a partial unblocking at ~150°–200°C but otherwise decays with nearly constant slope to zero between 550° and 600°C. Minor medium and hard fractions also decay consistently to zero at 550°–600°C. These results are consistent with coarse-grained magnetite mixed with titanomagnetite. However, one sample of this type (Sample 208-1267A-1H-3, 108 cm [0663]) has slightly larger medium and hard components (combined ~25% of the total), which also show evidence for partial unblocking at ~250°C (Fig. F8B). It is possible that these hard components represent a contribution from pyrrhotite or greigite, although reported maximum unblocking temperatures are somewhat higher for these minerals at 325° and 333°C, respectively (Dekkers, 1989; Roberts, 1995).

The coercivity of remanence derived from the hysteresis loops for all samples is low (14–24 mT), which is more consistent with a hysteresis dominated by (titano)magnetite rather than greigite or pyrrhotite (Roberts, 1995; Peters and Dekkers, 2003). The ratio of saturation IRM to low-field susceptibility has been suggested as a way to discriminate among several magnetic remanence carriers (e.g., Peters and Dekkers, 2003). Greigite and pyrrhotite typically have higher values than titanomagnetite, but the fields for the three minerals do overlap. Calculated values for this ratio are slightly higher for the Type 1 sediments (average = 25.2 kA/m) compared to the Type 2 sediments (average = 12.9 kA/m). Although the higher Type 1 values fall within the lower range of both pyrrhotite and greigite, all values still fall within the range for titanomagnetite.

To further explore the possible presence of iron sulfides in the medium or hard components, a few magnetic grains were separated from the noncarbonate fraction for scanning electron microscopy (SEM) analysis (Samples 208-1267A-1H-4, 108 cm, and 1H-5, 108 cm [0664 and 0665]). Results from two grains are consistent with titanomagnetite (Fe_3–xTi_xO₄) with x being ~0.49 and ~0.35, respectively. Titanomagnetites with these compositions correspond to Curie temperatures ranging from ~250° to 350°C. Al and Mg impurities are also present, however, and will lower T_c somewhat (Dunlop and Özdemir, 1997). Although we cannot definitively exclude the presence of iron sulfides in these samples, it seems likely that titanomagnetite can explain the unblocking at 250°C, although coercivities as high as 0.3 T would be unusual.

Type 2 sediments (Fig. F8C) are very similar to Type 1 sediments but show no evidence for the low-temperature unblocking of the soft component. These sediments, therefore, contain predominately coarse grained magnetite and are lacking the titanomagnetite inferred to be present in the Type 1 sediments. Type 2 sediments are found below ~98 mcd, although two samples with profiles intermediate between Types 1 and 2 are found between 98 and 129 mcd and another is found at ~264 mcd.

Type 3 sediments are represented by Sample 208-1267A-8H-4, 88 cm (0712) (Fig. F8D), found in an interval of extremely low remanent intensity and susceptibility at ~70–83 mcd at Site 1267 (see figs. F21 and F22 of Chapter 8 in Shipboard Scientific Party, 2004). This sample has the same low-coercivity behavior as Type 1 but also has large medium- and high-coercivity fractions that are not completely demagnetized by 600°C. This is consistent with a large contribution from hematite, as well as magnetite and titanomagnetite.

The downhole transition from Type 1 to Type 2 sediments is accompanied by an increase in remanent intensity and a decrease in the average median destructive field. This change to lower coercivity sediments below ~98 mcd can also be seen in the pass-through data by calculating the fraction of the NRM removed by demagnetization to 15 mT. These same changes in coercivity and intensity can be observed at all sites in the pass-through data at the same stratigraphic horizon in the upper Miocene. At Sites 1262 and 1267, this horizon has also been characterized as the transition between lithostratigraphic Units I and II. In several holes, the horizon roughly correlates with local pore water maximums in Mn²⁺ and minimums in Fe²⁺ (e.g., fig. F1 of Chapter 8 in Shipboard Scientific Party, 2004), suggesting a possible diagenetic link to redox conditions that might precipitate iron sulfides. However, because of the local nature of these peaks, we suggest the consistent presence of Type 1 sediments above the upper Miocene horizon is reflective of a change in the source of the detrital grains (magnetite/titanomagnetite) at this time.