After alternating-field (AF) demagnetization to 20 mT, the natural remanent magnetization (NRM) of whole-round sections from Site 1172 was measured at 5-cm intervals using the pass-through cryogenic magnetometer. An exception was made for cores whose liners or end caps were deformed, because these could damage the magnetometer. These deformed cores were measured as archive-half cores. The nonmagnetic core barrel assembly was used for even-numbered cores in Holes 1172A and 1172C and for odd-numbered cores, starting with Core 3H in Hole 1172B. The comparison between results from cores collected with the nonmagnetic corer and with those from the standard corer is discussed in the "Appendix" chapter as are results of experiments investigating the effect of core splitting on magnetization and other coring-related magnetic experiments. The Tensor tool is usually employed to orient the APC cores beginning with the third core from each hole, but the poor determination of the declination of the cores at this site precluded such orientation.
Oriented discrete samples were routinely collected from Holes 1172A and 1172D. 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, measurements were made of the demagnetization behavior of anhysteretic remanent magnetization (ARM) given in 0.2 direct-current (DC) and 200-mT alternating-current (AC) fields and isothermal remanent magnetization (IRM) in a DC field of 1 T. Some discrete samples were progressively saturated up to 1.0 T to study the hardness of the IRM.
The long-core measurements for the APC cores are presented in Figures F16 and F17, which show inclination and intensity for Holes 1172A, 1172B, and 1172C. The magnetostratigraphic interpretation was based primarily upon Hole 1172B. The intensity of magnetization was between 10-5 and 10-4 mA/m, so that measurements again approached the noise level of the instrument and the background noise from core liners. However, in Holes 1172B and 1172C a useful magnetic record emerged, which allowed us to establish a continuous magnetostratigraphy from the late Pleistocene down to the middle Miocene. Using sequences of reversals, we also determined that, for the most part, the depth shift between Holes 1172B and 1172C was <1 m.
In the first 110 m, we identified successively at 12.5 mbsf the onset of the Brunhes Chron; at 19.5 mbsf, the onset of the Olduvai Subchron; at 46.25 mbsf, the onset of the Gauss Chron; at 78.9 mbsf, the onset of Subchron C3n.4n; and at 107.4 mbsf, the onset of Subchron C3An.2n. In both records, the Jaramillo Subchron was missing, which indicates the presence of an early Pleistocene hiatus. The duration of the hiatus must be <1 m.y. because Subchron C3n.3n was identifiable in Hole 1172C.
Figure F17 shows the long-core measurements and interpreted magnetostratigraphy for the depth interval of 100-200 mbsf. Unfortunately, weather conditions were poor during the drilling of Hole 1172A, resulting in substantial disturbed sequences, but conditions in the two other holes permitted the magnetostratigraphy to be completed, although it may require some refinement from postcruise analysis. From 200 to 300 mbsf, the magnetostratigraphy is poorly defined with only the onset of Chron C5n and the termination of Subchron C5An.1n possibly discernible (Fig. F18).
The interval from 300 to 400 mbsf is of particular interest because it contains the remarkable Eocene/Oligocene boundary section, which was recovered in Core 189-1172A-39X and subsequently in Core 189-1172D-2R. Curiously, the RCB core has a much better paleomagnetic record (Fig. F19). Chron 11n is identified between 356.2 and 358.5 mbsf, which is consistent with the diatom datum of 30.2-30.8 Ma for this interval and within the weaker constraints of 26-32 Ma from calcareous nannofossils and post-33.3 Ma from dinocysts. We were not able to trace the magnetostratigraphy through the boundary, but it seems likely that the predominantly reversed chrons between ~30 and 35 Ma are represented by the reversed interval between 350 and 360 mbsf. Immediately below this we interpret Subchron 16n.2n, which is consistent with a dinocyst datum of 35.5 Ma at a depth of 359 ± 1 mbsf. This then places Subchron 17n.1n below 367.7 mbsf, which is consistent with the dinocyst datum of 36.5 Ma at 361.8 ± 0.96 mbsf. Below this depth, the interval is dominated by the higher Eocene sedimentation rate.
In the Eocene section and below, we encountered severe difficulty in interpreting the magnetostratigraphy from the inclination data because there was a strong normal overprint. However, it was clear from the intensity record that intensity fluctuations correlated with the minor indications of magnetozones in the inclination. Therefore, we used the intensity of the z-component to interpret the magnetostratigraphy. Initially, we simply removed the bias that represented the normal overprint and recovered a more easily interpreted record of the magnetozones. Eventually, it should be possible to improve this method by first detrending the data and then applying a sum and difference calculation to separate the constant overprint from the opposed normal and reversed z-component magnetization. An example of the initial application of the technique is shown in Figure F20.
Samples were analyzed in a similar manner to that used at previous sites, and the results are presented in Figures F21 and F22. For the most part, the intensity of saturation IRM and ARM, and hence their ratio, were all consistent over much of the cored depth in the sampled holes. A notable exception is in an interval centered ~340 mbsf. There is a gradual increase in intensity associated with an increase in volcanic glass and the ratio of ARM:IRM increases, probably reflecting the fine grain size of the particles in the glass and devitrified products.
The combination of data from the APC part of Hole 1172A and the APC Holes 1172B and 1172C provides an excellent magnetostratigraphy and, hence, age-depth determination for this interval to 200 mbsf. The results are given in Table T13 and Figures F16 and F17. The record was sufficiently good enough to demonstrate that Subchron 1r.1n (the Jaramillo) was missing, indicating a lower Pleistocene hiatus. Although the quality of the record deteriorates considerably in the older intervals, a relatively complete magnetostratigraphy was achieved and, for the most part, found to be consistent with the biostratigraphic data (Fig. F21). The magnetostratigraphy across the Eocene/Oligocene boundary constrained the timing of events, but unfortunately across the suspected K/T boundary, there was no strong indication of Chron 29R in the inclination record.