The natural remanent magnetization (NRM) and remanent magnetization after alternating-field (AF) demagnetization of the archive-half sections from Holes 1168A and 1168B were measured at 5-cm intervals using the pass-through cryogenic magnetometer. We measured the NRM of all cores up to Core 189-1168A-26X, but the remainder were only measured after 20-mT demagnetization because of time constraints. A nonmagnetic core barrel assembly was used for alternate cores in each hole. In Hole 1168A, even-numbered APC cores were taken with the nonmagnetic core barrel assembly. In Holes 1168B and 1168C, odd-numbered cores were taken, starting with Cores 3H. The comparison between results from cores collected with the nonmagnetic corer and standard cores is discussed (see the "Appendix" chapter), as are results of experiments investigating the effect of core splitting on magnetization and other coring-related magnetic experiments. We note here that a strong overprint was found to be induced by splitting sections from two test cores (Sections 189-1168B-3H-4, 3H-5, and 3H-6; 189-1168B-8H-2, 8H-3, 8H-4, 8H-5, and 8H-6). This was not demagnetized by 20 mT. In Hole 1168C, long-core sections were therefore run as whole cores, rather than as archive halves, to determine if a better magnetostratigraphy could be obtained.
The Tensor tool was used to orient the APC cores beginning with the third core at each hole. The variability in the declination values for Holes 1168A and 1168B precluded the orientation of the cores.
Discrete oriented samples were routinely collected from Hole 1168A; two samples taken from each working-half core. These were used to aid in the interpretation of the long-core record of magnetization by providing additional measurements of polarity and basic magnetic characterization. Most of them were demagnetized at 5, 10, 15, 20, 30, 40, and 50 mT to permit principal component analysis. For rock magnetic characterization, anhysteretic remanent magnetization (ARM) and isothermal remanent magnetization (IRM) were applied, measured, and progressively demagnetized up to 50 mT. A DTECH AF demagnetizer (model D-2000) was used to impart an ARM. A direct-current (DC) field at 0.2 mT combined with an alternating field of 200 mT was used. IRM was given in a DC field of 1 T with an Analytical Service Company (ASC) model IM-10 impulse magnetizer. Some discrete samples were progressively saturated to 1.3 T to study the hardness of the IRM.
Long-core measurements provided, for the most part, poor records of reversal sequences with shallow inclinations, anomalously long transitions between polarities, and the 0° declination artifact. After demagnetization of coring overprints, the magnetization was on the order 10-5 A/m, which is close to the sensitivity limit of the instrument. Moreover, measurements of empty core liners gave spot readings on the same order of magnitude of the intensity of magnetization. These facts explain the poor quality of the paleomagnetic signal. Hole 1168C sections were measured as whole cores and for the first ~30 mbsf yielded inclinations close to the expected value of 62° for the site latitude of 42°S (Fig. F22). The first 0.25 mbsf was oxidized and strongly remagnetized in the present field direction (see "Physical Properties"). Five well-defined magnetozones (three normal and two reversed) were identified between 0.25 and 28.1 mbsf. Unfortunately, the correlation with the geomagnetic polarity time scale (GPTS) (Berggren et al., 1995a, 1995b) is not completely satisfactory. Three biostratigraphic datums are available (see "Biostratigraphy")—the FO of E. huxleyi (0.26 Ma) at 0.64 mbsf, the LO of C. macintyrei (1.67 Ma) at 12.12 mbsf, and the LO of D. surculus (2.55 Ma) at 21.86 mbsf—to constrain the magnetozones. We interpret the termination of the Jaramillo Subchron (C1r.1n) at 3.75 mbsf, the onset of this subchron at 9 mbsf, and the termination of the Olduvai Subchron (C2n) at 28.1 mbsf.
At greater depth in the core, we attempted to use biostratigraphic markers to constrain the depth of such long chrons as C5n, C6n, C12r, and C13r. In addition, we used the measurements of discrete samples to find depths where a reliable paleomagnetic record was likely to be present. Using these techniques we were able to establish a magnetic polarity record between 250 and 287.6 mbsf; where the magnetization is relatively strong, inclinations approached the expected values, and the numerous biostratigraphic datums allowed us to constrain the correlation with the GPTS (Fig. F23). Magnetozones can be matched to Subchrons C5En, C5Dr, C5Dn, C5Cr, and C5Cn. Around 300 m, the sedimentation rate changes, reaching >40 m/m.y., which makes interpretation of the long positive-magnetization zone below C5Br problematic. However, it appears that C6n is indeed between 321.5 and 347.3 mbsf. Attempts to identify sequences of reversals around other long periods of constant polarity were unsuccessful. In the two depth ranges, within which we established the polarity sequences, the intensity of magnetization measured in the long-core mode was 10-3 A/m or more.
Examination of discrete samples indicated that the majority of samples carried a poor paleomagnetic record in their NRM and after demagnetization to 20 mT, although there were notable exceptions (Fig. F24). The preliminary rock magnetic analysis in terms of demagnetization of NRM, ARM, and saturation IRMs revealed a range in magnetic characteristics that afforded a basis for interpretation of the paleomagnetic record. For example, in the upper calcareous part of the hole, the ratio of ARM to IRM was high, suggesting the predominance of fine, single-domain material, probably magnetosomes, as the carrier of the paleomagnetic record. There was a particularly striking contrast in the characteristics of samples taken from Cores 189-1168A-30X through 34X compared with 35X and 36X. Cores 189-1168A-30X through 34X carried NRM, which gave principal component analysis (PCA) results with maximum angular deviation (MAD) angles of as low as 3°, indicating that a stable magnetization had been acquired (Fig. F24). In Cores 189-1168A-35X and 36X, the results were so scattered that no PCA could be carried out after the coring contamination had been separated. Figure F25 shows that Cores 189-1168A-30X through 34X have stronger NRM, ARM, and IRM ratios of ARM:IRM and have hard IRM relative to Cores 189-1168A-35X and 36X. These results indicate that Cores 189-1168A-30X through 34X have finer magnetic material than Cores 189-1168A-35X and 36X. Whether this reflects better preservation of magnetosomes or secondary fine particles produced in the oxidation region is not initially clear. However, the low ratio of NRM:IRMs suggests that this is not a chemical remanence and that the preservation of magnetosome magnetite may be better in this interval than above. Cores 189-1168A-34X and 36X appear to represent a change in sedimentation associated with coarser magnetic material.
As is evident from Figures F22 and F23 and the discussion of long-core measurements above, the magnetostratigraphy of this site is very difficult to establish. The principal problem lies in the weak intensity of magnetization of the sediments, which approaches the noise limit of the instrument. The weak intensity also means that magnetic contamination adding to the signal has a much greater effect than in more strongly magnetized cores. It is for these reasons that we have only given magnetostratigraphic results for the more strongly magnetized parts of the hole. The results obtained are shown in Figures F22 and F23 and in Table T13. They proved to be in good agreement with the biostratigraphic data and were included in the time-depth study. However, whereas the magnetostratigraphy at the top of the core can be recognized independently from the biostratigraphy, the magnetostratigraphy between 250 and 350 mbsf is dependent upon the biostratigraphy for the location of Chron C6n.