We measured the natural remanent magnetization (NRM) of most archive halves from Hole 1138A with the pass-through cryogenic magnetometer using measurement intervals of 5 and 2.5 cm for sediment and basement rocks, respectively. Subsequently, sediment and basement core sections were demagnetized with peak alternating fields (AF) of 20 and 50 mT, respectively. Discrete sediment samples were stepwise AF demagnetized up to 60 mT. Discrete basement samples were stepwise AF demagnetized in a peak field of up to 60 mT or thermally demagnetized in temperatures of up to 620°C. We determined the anisotropy of magnetic susceptibility (AMS) of discrete basement samples to obtain information about the magnetic fabric.
We obtained stable and reliable paleomagnetic directions from the sediments of Hole 1138A. Correlation of biostratigraphic data and polarity reversals suggests that the reversed and normal chrons are Pliocene to Late Cretaceous in age. We compared magnetic properties with lithologic and basement units. The basement rocks yielded reliable paleomagnetic directions and a normal polarity. We found a difference of approximately 8° between the paleolatitude (46°S) and the present latitude (54°S) of Site 1138.
We measured the remanent magnetization of all archive halves from Hole 1138A, except for highly disturbed sections. One archive section for each core was stepwise AF demagnetized up to 30 mT. We took two discrete samples per section, and 42 samples were stepwise AF demagnetized up to 60 mT to confirm the reliability of whole-core measurements. We obtained reliable results in undisturbed cores, and correlated normal and reversed segments with biostratigraphic zones (see "Biostratigraphy"). The sediments of Hole 1138A generally have stable magnetization, which was obtained after AF demagnetization at 10 mT. Most discrete samples have a high median destructive field (MDF), and the remanent direction is stable in demagnetization steps between 10 and 30 mT (Fig. F72). We, therefore, used the remanent magnetization after AF demagnetization at 20 mT to correlate the paleomagnetic record with geomagnetic chrons. Furthermore, we used the data selection criteria described in "Paleomagnetism" in the "Explanatory Notes" chapter for magnetostratigraphic studies of Hole 1138A (Fig. F73). The selection criteria were that (1) the intensity of remanent magnetization after AF demagnetization at 20 mT was >2 × 10-4 A/m, (2) the inclination was > ±30°, (3) at least two consecutive values (which corresponds to a 10-cm length of split core) had the same polarity, and (4) there was no significant core disturbance. Characteristic inclinations from discrete samples generally agree well with selected inclinations from whole-core measurements (Fig. F73).
Correlation of biostratigraphic data and polarity reversals (Fig. F73) suggests that the reversed and normal chrons in Hole 1138A are Pliocene to Late Cretaceous in age (see "Biostratigraphy"). We obtained a relatively continuous paleomagnetic record from high-recovery core except in the upper part of Unit I (see "Lithostratigraphy"). Between 75 and 190 mbsf, 315 and 370 mbsf, 450 and 510 mbsf, and 620 and 665 mbsf, the paleomagnetic record appears continuous. We propose the following correlations with paleontological data from the core catcher of each core (see "Biostratigraphy"). We correlate the normal and reversed segment between 70 and ~120 mbsf to Pliocene Chrons C2r to C3n. The underlying normal and reversed segment between ~120 and 135 mbsf corresponds to late Miocene chrons between C3r and C4. We correlate the normal segment between 140 and 153 mbsf and the reversed segment between 153 and 175 mbsf with Chrons C5n and C5r, respectively. Dispersed normal and reversed segments between 180 and 330 mbsf correlate with Chrons C5An (middle Miocene) to C11r (early Oligocene). We correlate normal and reversed polarities between 343 and 442 mbsf, and between 448 and 470 mbsf to middle Eocene Chrons C17-C21 and early Eocene Chron C24 to late Paleocene Chron C26, respectively. We obtained a polarity record from early Paleocene Chron C27 to Late Cretaceous Chron C32 between 475 and 575 mbsf. The normal sequences at ~613 mbsf and between 620 and 692 mbsf may correlate with Campanian Chron C33n and Cretaceous long normal Chron C34n, respectively.
We obtained reliable paleomagnetic data from most of the sediments in Hole 1138A. The record is less reliable, however, between 0 and 70 mbsf and between 540 and 605 mbsf. We observed high susceptibilities (whole-core MST measurements, see "Physical Properties") and strong NRM intensities, but scattered inclinations in the upper part of Unit I (diatom clay and ooze, see "Lithostratigraphy") (Fig. F73). Scattered directions are probably caused by core disturbance. Among the sediment units of Hole 1138A, we observed the lowest susceptibilities and weakest NRM intensities in Unit III (white chalk) (Fig. F74). The lower part of Unit IIIB is characterized by negative susceptibilities and weak remanent magnetization, and the paleomagnetic results are unreliable. We observed high susceptibilities and strong NRM intensities in neritic sediments (Units V and VI, see "Lithostratigraphy") (Fig. F74).
We determined the magnetic properties of each basement unit (see "Physical Volcanology" and "Igneous Petrology") and the variation of magnetic properties within each unit (Fig. F75). Three independent types of susceptibility measurements, MST, AMST, and discrete samples generally show consistent results. We observed no significant differences in average susceptibility and NRM intensity among lava flows (basement Units 3-22) (Figs. F74, F75) and no significant variations within the units. In several parts of the lava flows, we observed scattered inclinations of whole-core measurements after AF demagnetization at 40 mT. However, we chose the data points obtained from core pieces that fit together and are longer than the effective sensitivity of the pass-through magnetometer (~15 cm) to obtain reliable inclinations without drilling disturbance. Scattered inclinations from the upper part of basement Units 9, 10, and 13 correspond to brecciated parts. Inclinations are also scattered in the lower parts of basement Units 9, 10, 11, 13, and 14. These rocks were probably rotated after acquisition of thermal remanent magnetization (< ~600°C). Basement Unit 2 (volcaniclastic sediments, see "Physical Volcanology") has lower susceptibilities and lower NRM intensities than the lava flows (Fig. F74). Susceptibilities and NRM intensities are lower in the uppermost 3 m (reworked volcanic sediments) and in the lowermost 2 m (altered pumice flow) of basement Unit 2 than in the middle part (mostly pumice lithic breccia or lithic breccia) (Fig. F75). Inclinations are scattered in the lowest 2 m.
From 30 discrete samples of basement Units 6, 7, 8, 12, and 13, we measured magnetic susceptibility in 15 different directions to determine the AMS and the magnetic fabric. The degree of anisotropy (ratio between maximum and minimum axes) was generally low and its magnitude ranged from 1.01 (basement Unit 6) to 1.05 (basement Unit 13). We found better developed shapes of magnetic fabric with increasing degree of anisotropy. All samples have negative shape parameters or shape parameters close to zero; hence, prolate (rod) shapes predominate. In basement Unit 13, which is described as channelized shelly/slab pahoehoe lava (see "Physical Volcanology"), we found a relatively high degree of anisotropy and a grouping of the minimum axes along the vertical axis (Fig. F76A). However, in the other basement units of different origins (e.g., basement Unit 6, inflated pahoehoe), we found lower degrees of anisotropy and we could not determine groupings of any (minimum and/or maximum) axes (Fig. F76B).
Five discrete samples from basement Units 3, 7, 9, 13, and 20 were stepwise thermally demagnetized up to 620°C, and two samples from Units 10 and 20 were progressively AF demagnetized up to 60 mT. We measured the susceptibility of the samples after each heating step to detect changes in their magnetic minerals. The two AF demagnetized samples have stable, single-component remanent magnetization (Fig. F77), and their MDFs were ~20 mT. We found two magnetic phases during stepwise thermal demagnetization. The high-temperature phase is characterized by an unblocking temperature of ~580°C (Fig. F78A) and the low-temperature phase by an unblocking temperature of ~300°C (Fig. F78B). The high-temperature phase probably corresponds to magnetite or titanium-poor titanomagnetite, and the low-temperature phase to (titano)maghemite as previous rock magnetism studies of the southern Kerguelen Plateau basalts have shown (Heider et al., 1992) (see "Paleomagnetism" in the "Site 1136" chapter). Most samples exhibit two magnetic phases and stable, single components of remanent magnetization.
We calculated the characteristic inclinations of discrete samples using component analysis (Table T16). Inclinations are negative, indicating normal polarity, and range from -53° to -80°. Variation of inclinations among samples is probably caused by secular variation. We calculated a mean inclination of -65°, which corresponds to a paleolatitude of 46°S assuming a geocentric dipole field. The difference between the present latitude of 54°S of Site 1138 and its paleolatitude during the Cretaceous (see "Biostratigraphy") is 8°, suggesting that either the central Kerguelen Plateau has moved south or basement has been tilted since Cretaceous time. Interpreted igneous basement contains some internal reflections with a slight apparent dip to the southwest. The top of basaltic basement, however, is flat lying (see "Background and Objectives"). We conclude that the central Kerguelen Plateau has moved ~8° southward since Cretaceous time, which is consistent with the contemporaneous southward movement of the southern Kerguelen Plateau (Inokuchi and Heider, 1992) and Elan Bank (see "Paleomagnetism" in the "Site 1137" chapter). Limited measurements, however, preclude definitive paleolatitude estimates at this time. We will investigate this problem further with shore-based studies.