All the archive-half sections from Hole 1167A (APC and XCB cores) were subjected to pass-through measurements, except for Sections 188-1167A-19X-2; 37X-1, 37X-2, and 37X-3; and 39X-1 and 39X-2 because they were sandy. The natural remanent magnetization (NRM) and remanent magnetization after alternating field (AF) demagnetization were measured routinely using the shipboard pass-through cryogenic magnetometer at 4-cm intervals. Three AF steps at 10, 20, and 30 mT were used for all core sections. A total of 252 discrete samples (standard oriented 8-cm3 cubes) were collected from the center of the working halves at a frequency of one or two per section. The lithofacies are mainly dominated by coarse-grained sediments with dispersed clasts and minor beds of coarse sands, clays, and sandy clays (see "Lithostratigraphy"). When possible, samples were selected from fine-grained horizons; however, there was often no alternative but to sample from the sandstone-dominated lithofacies.
A total of 175 samples, after measurement of NRM, were AF demagnetized at successive peak fields of 2, 7, 10, 20, 30, 40, 50, and 60 mT. Thermal demagnetization was conducted on 12 samples collected throughout the core at temperatures of 100°, 200°, 300°, 350°, 400°, 500°, 550°, 600°, 650°, and 700°C. Magnetic susceptibility was measured after each step to monitor for thermal alteration of the magnetic fraction.
Rock magnetic analyses were performed on a set of representative discrete samples after they had been subjected to AF demagnetization in order to obtain a quantitative estimate of downcore variation in the composition, concentration, and grain size of the magnetic minerals. These variations often provide valuable information about changes in paleoenvironmental conditions in a sedimentary basin and its surrounding regions (Thompson and Olfield, 1986; Verosub and Roberts, 1995). The mineral magnetic analyses followed the same approach that was utilized at Sites 1165 and 1166.
Low-field magnetic susceptibility (k) was routinely measured for each discrete sample (252), and the resultant data were compared with the whole-core susceptibility log (see "Physical Properties"). The frequency-dependent susceptibility, fd(%), was measured on 59 selected samples. Anhysteretic remanent magnetization (ARM) was measured for 172 samples using a 100-mT AF with a superimposed 0.05-mT bias field. On 164 samples, an isothermal remanent magnetization (IRM) was imparted in a direct-current field of 1.3 T. On 57 of these samples, the IRM was then demagnetized by inverting the sample and applying a backfield of 300 mT to determine the S-ratio (-IRM -0.3T / IRM 1T) (e.g., Verosub and Roberts, 1995). The progressive acquisition of IRM was studied for 12 selected samples.
Time constraints did not allow the investigation of the coercivity of remanence (Bcr) or the analysis of thermal demagnetization of the composite IRM (Lowrie, 1990).
Analyses of the rock magnetic properties from Hole 1167A suggest that the core can be divided into two main units (Units I and II) and a number of subunits based on the abundance and grain size of the magnetic minerals in the sedimentary sequence. The main unit boundary coincides with a lithologic change at 217 mbsf (see "Lithostratigraphy"), whereas the subunit boundaries cannot be directly related to visual lithologic variations in the core.
Magnetic Unit I (0-198.6 mbsf) can be divided into five subunits on the basis of changes in the concentration-dependent parameters (k, IRM intensity, and ARM intensity), which have similar patterns of variation (Fig. F21). Subunit IA (0-4 mbsf) is characterized by relatively low k, ARM, and IRM. At the boundary between Subunits IA and IB (4 mbsf), the magnetic susceptibility jumps from ~23 × 10-5 to 126 × 10-5 SI then rises in a quasi-linear fashion to ~300 × 10-5 SI at 55 mbsf. Between 55.0 and 78.5 mbsf (Subunit IC), the magnetic susceptibility (along with the other magnetic concentration parameters) drops to 128 × 10-5 SI, after which it remains approximately constant with a mean value of 160 × 10-5 SI. In Subunit ID (78.5-112.2 mbsf), susceptibility rises quasi-linearly to 224 × 10-5 SI. Lack of recovery from ~113 to 151.2 mbsf renders susceptibility levels at the base of this unit uncertain. Subunit IE (112.2/151.2-198.6 mbsf) is characterized by a quasi-linear rise in k from 178 × 10-5 to 214 × 10-5 SI.
Unit II (208.3-447.7 mbsf) can also be divided into subunits on the basis of downcore variations in k, IRM intensity, and ARM intensity. Subunit IIA (208.3-217.5 mbsf), corresponding to decimeter-scale beds of dark gray clay (lithostratigraphic Unit II-3; see "Lithostratigraphy"), is characterized by relatively constant values of k. The sharp rise in k at the base of this unit is probably related to the presence of igneous clasts in the samples. The susceptibility steadily increases downcore between 217.5 and 447.7 mbsf (Subunit IIB). The sandy horizon at ~325 mbsf is reflected in low susceptibility at this depth.
Preliminary mineral magnetic analyses, based on coercivity spectrum analyses and thermal demagnetization behavior, are consistent with a magnetite-dominated magnetic mineralogy in samples from much of the core. Plots of IRM acquisition display steep slopes at low magnetic inductions, and saturation is achieved between 200 and 300 mT (Fig. F22). S-ratio values (see "Paleomagnetism" in the "Explanatory Notes" chapter) are higher than 0.96. In three thin intervals, located at 3.40, 80.0, and 431.0 mbsf, lower S-ratios and resistance to AF demagnetization indicate the prevalence of a high-coercivity mineral (e.g., hematite) as a major magnetic carrier in these horizons.
With the magnetic mineralogy constrained as magnetite, the ARM/IRM ratio can be used as a magnetic grain-size indicator because ARM is more effective in activating finer magnetite grains than IRM. In Figure F23 the ARM/IRM ratio, plotted as a function of depth, shows a considerable variation in grain size between 0 and 15 mbsf. In the interval between 15 and 217 mbsf, the ratio of ARM/IRM is relatively constant, ranging from 0.029 to 0.038 with a mean value of 0.032. Below 217 mbsf, variation of the ARM/IRM ratio is relatively minor, ranging from 0.039 to 0.053 with a mean value of 0.046. The major boundary at 217 mbsf is consistent with a shift from relatively coarse-grained magnetite (above) to relatively fine-grained magnetite (below).
We suggest that the variations in both concentration and grain size reflect changes in sediment provenance. The "sawtooth" concentration fluctuation reflects the varying importance of two or more sediment sources: one or more magnetite rich and one or more magnetite poor. Similarly, we interpret the change in magnetite grain size at 217 mbsf to reflect a change from sources containing relatively fine-grained magnetite to sources containing relatively coarse-grained magnetite. Traces of authigenic pyrite are observed in a few horizons (see "Lithostratigraphy"); however, diagenetic alteration of the core is not evident from the magnetic signal, and geochemical data are inconsistent with significant diagenesis (see "Inorganic Geochemistry").
Coarse-grained sediments similar to those common in Hole 1167A have historically been deemed unsuitable for paleomagnetic investigations; however, recent paleomagnetic investigation of glaciogenic sedimentary units from the Victoria Land Basin (Ross Sea) evidence the presence of strong and stable magnetizations even in coarse-grained units (Wilson et al., 1998, in press; Roberts et al., 1998). The stability of these remanences was attributed to the presence of fine magnetic particles within the fine-grained sedimentary matrix of these otherwise coarse-grained units. Hole 1167A paleomagnetic analyses are also hindered by the presence of granules and pebbles dispersed throughout the hole. The presence of such clasts requires that care be taken in the interpretation of paleomagnetic data, especially for data originating from the archive-half sections in which it is not possible to verify the presence of clasts beneath the section surface. This emphasizes the importance of paleomagnetic analysis of discrete samples to verify the reliability of whole-core measurements.
Most of the analyzed cores display a low-coercivity, nearly vertical reversed polarity component that we interpret to represent a drilling-induced overprint (Weeks et al., 1995). In a few horizons, a normal polarity drilling-induced overprint was also observed. In most cases this component was removed with peak AFs of <10 mT. After the removal of this overprint, a stable characteristic remanence component (ChRM) is evident for a large portion of the analyzed archive halves and discrete samples (Fig. F24). In some cases, this component and the characteristic component of magnetization have completely overlapping coercivity spectra, rendering no demagnetization interval over which only one component is removed (Dunlop, 1979). In these situations it is not possible to isolate the two components.
For 76% of the discrete samples, stable paleomagnetic behavior was evident from the vector component diagram. In most cases the ChRM direction was determined using a best-fit line that was constrained, based on principal component analysis, through the origin of the vector component diagram. In some cases the best-fit line was not constrained through the origin of the plot. The ChRM directions are in excellent agreement with the directions obtained from long-core measurements, with the exception of the reversed polarity direction indicated by a discrete sample at 16.92 mbsf where, conversely, the long-core measurement indicated a normal polarity.
Preliminary magnetostratigraphic interpretation for Hole 1167A is shown in Figure F25. The uppermost 30-m interval of the polarity record is entirely normal. Transitional directions are recorded over a stratigraphic interval of ~4 m, between 30 and 34 mbsf. Considering that the process of polarity reversal occurs over periods of ~5-10 k.y. (Jacobs, 1994), it is possible to estimate a sedimentation rate for this interval (~0.4-0.8 m/k.y.). From 34 mbsf to the bottom of the hole, the polarity is reversed with two short normal polarity intervals at ~355 and at ~380 mbsf.
Biostratigraphic data are restricted to a diatom assemblage constraining the top of Core 188-1167A-1H to <0.66 Ma and two nannofossil assemblages: a Zone CN14a assemblage at 37.05 mbsf and a Zone CN13b assemblage between ~218 and 228 mbsf, with ages of 0.4-0.9 Ma and 0.9-2.0 Ma, respectively (see "Biostratigraphy and Sedimentation Rates"). Constrained by these nannofossil datums, the upper normal polarity magnetozone is correlated with the Brunhes (C1n) Chron and the long reversed interval, between 34.0 mbsf and the bottom of the hole (447.5 mbsf), is correlated with the C1r.1r and C1r.2r Chrons. We suggest that the Jaramillo Subchron (C1r.1n) is missing possibly because of unconformities in the record (e.g., at 55 m) that are suggested from sharp changes of the concentration-dependent parameters (Fig. F21).
Eight samples with adequate numbers of the planktonic species N. pachyderma have been collected for Sr dating (see "Biostratigraphy and Sedimentation Rates"). These samples, close to the interpreted Brunhes/Matuyama boundary and to the two thin normal polarity intervals, will further constrain this preliminary magnetostratigraphic interpretation.
Notably, the inclinations downcore are consistently shallower than expected for the site latitude (66°S). We suggest that observed shallow inclinations are due to sediment compaction immediately after deposition (e.g., Anson and Kodama, 1987; Arason and Levi, 1990).