Archive-half sections of cores from Holes 897A, 897C, and 897D were measured using the cryogenic magnetometer. Most sections were alternating-field (AF)-demagnetized at 10-cm intervals using a peak field intensity of 15 mT. When time permitted, sections of lithostratigraphic Unit IV were demagnetized progressively at 5-cm intervals using peak fields of 2,5,10, and 15 mT. A total of 31 discrete samples, taken from the working halves of cores recovered from the three holes, were AF demagnetized progressively up to a peak field of 50 mT, and 12 minicores from Holes 897C and 897D were thermally demagnetized to determine directional stability and to verify magnetostratigraphic results from the cryogenic magnetometer. Magnetic susceptibility was measured at 3- to 5-cm intervals on most of the recovered cores from Holes 897A, 897C, and 897D.
One of the major experimental requirements in paleomagnetic research is to isolate the characteristic remanent magnetization by selective removal of secondary magnetization. In Figure 22, we show the natural remanent magnetization (NRM) intensities and inclinations before and after 15-mT AF demagnetization observed in Cores 149-897D-4R and -5R. The NRM inclinations are strongly biased toward high positive inclinations (65° to 87°). AF demagnetization to 15 mT removes a soft, possibly drilling-induced component, as indicated by a change in inclination to values ranging from 10° to 60°. The expected normal inclination at Site 897 is about 60°. That the NRM inclinations are biased toward the vertical (+90°) in most cores suggests that drilling-induced magnetization is present.
In several cores, we observed a dramatic decrease in magnetic intensity by one or two orders of magnitude from the upper to the lower part of a single core. Examples of this phenomenon are found in Cores 149-897C-8R (117.9 mbsf), -12R (156.6 mbsf), -19R (228.6 mbsf), and -35R (379.0 mbsf). The magnetic susceptibility data of these cores also show extremely high-amplitude spikes (with values exceeding 1.88 -3 SI units). We suspect that this is also related to drilling. It may have been caused by rust flakes that fell into the uppermost portions of each core and/or exposure to a strongly magnetized core barrel.
Oriented samples (6-cm3 plastic cubes or 10-cm3 minicores) were collected from almost every core: (1) to verify the magnetostratigraphic results from the cryogenic magnetometer and (2) to understand the magnetic behavior of the sediments during stepwise demagnetization. Thirty-one of these samples were measured using the Molspin magnetometer, with progressive AF demagnetization steps to 50 mT on a Schonstedt GSD-1 AF-demagnetizer. During AF demagnetization of the discrete samples with this instrument, we noticed that several samples acquired spurious magnetizations above 50 mT (e.g., Fig. 23). Tests, conducted after demagnetization along each positive and negative axis, suggest that these spurious magnetizations were mainly an anhysteretic remanent magnetization (ARM), produced along the axis of demagnetization. Thus, the determination of the magnetization carried by high-coercivity magnetic grains was hindered, and higher demagnetization steps will have to be made in shore-based studies on a triaxial tumbling AF demagnetizer.
Progressive thermal demagnetization was performed on 12 mini-cores from representative lithified cores in Holes 897C and 897D to test the stability of the magnetization vs. temperature and to help identify the magnetic minerals. We found that removal of secondary magnetization was better accomplished through thermal demagnetization than through AF demagnetization (Fig. 24). A secondary magnetic component of magnetization was removed at low temperatures (300°C), and a component of magnetization having higher unblocking temperatures (up to 620°C) could be identified. Despite the apparent stability of the magnetization of these samples, the absence of core orientation and information about the attitude of the bedding planes precluded a more detailed interpretation of the paleomagnetic data.
In Hole 897A, paleomagnetic results were obtained from only three cores because of poor recovery. Evidence for remagnetization imparted by coring was observed. Clayey silts appear to be more easily affected by this type of remagnetization than the turbidite units recovered in Hole 897C. We observed only normal magnetic polarity.
The paleomagnetic properties of material recovered from Holes 897C and 897D reflect the observed lithologic variations. Thus, the paleomagnetic data from these two holes are discussed according to the major lithologic changes and are summarized below:
Below 677.5 mbsf in Hole 897C and below 693.8 mbsf in Hole 897D, all cores are extensively serpentinized peridotites. Pervasive brittle deformation with serpentinite and calcite veining has affected the entire section (see "Structural Geology" section, this chapter). AF and thermal demagnetizations in nine discrete samples revealed an apparently stable component of magnetization. Information is lacking about core orientation and bedding attitude of these rocks, which precludes a more detailed interpretation of the paleomagnetic data at present. It may be possible after further work to investigate whether a component of viscous remagnetization parallel to the present-day magnetic field has been recorded in these rocks that may be used to reorient these cores. Rotary coring often disrupts the core by breaking it into several pieces that rotate independently of each other within the core liner. These effects can sometimes be removed using magnetic declination data, assuming that each segment having a constant declination is a homogeneous drilling biscuit. Several sections in Cores 149-897C-71R and -73R, and in Core 149-897D-23R exhibit drilling-induced rotations. The orientations of the veins in different parts of these sections suggest that these can be divided into two groups, with a difference in orientation of about 90°. Paleomagnetic declinations measured between the groups are also about 90° apart, suggesting that they rotated about 90° with respect to each other during drilling.
Age control in the sediments at Site 897 is primarily from nannofossil and planktonic foraminiferal biostratigraphy. Preliminary magnetostratigraphic interpretation mostly agrees with biostratigraphic ages (Table 8). The normal polarity of Cores 149-897A-1R through -3R suggests that these sediments were deposited during the Brunhes Chron (i.e., age <0.78 Ma), which agrees well with the biostratigraphic age estimate (see "Biostratigraphy" section, this chapter). Although dates (based on foraminifers) suggest that Cores 149-897A-4R through -6R were deposited within the Matuyama reversed Chron, we observed no evidence for a reversed polarity in these cores. However, most sections of these cores were highly disturbed by drilling, making the polarity record less reliable. The Brunhes/Matuyama boundary was not observed in Hole 897C. The first evidence for a reversed magnetization occurs in Core 149-897C-10R (137.2-138.7 mbsf), which may correspond to the Olduvai (1.81 Ma), assuming that the biostratigraphic ages (Table 8) are correct. Samples between 137.2 and 181.8 mbsf (Cores 149-897C-1 OR to -14R) tentatively have been assigned to the late Pliocene on the basis of preliminary biostratigraphic data. Thus, the reversal sequence found in this interval suggests that these sediments were deposited within the Gauss Chron (2.6-3.55 Ma). The next reversal was found in Cores 149-897C-19R through -23R (224.1-262.7 mbsf). Preliminary biostratigraphic evidence suggests that sedi-ments from these cores are of early Pliocene age (NN15). Therefore, the shift in polarity from normal to reversed at Core 149-897C-19R (224.1 mbsf) may represent the Gauss/Gilbert boundary (3.55 Ma). A transition from normal to reversed polarity in Core 149-897C-28R (310.9-312.4 mbsf) might record the Nunivak Subchron (4.12-4.26 Ma); this is supported by the calcareous nannofossils found in this core (see Table 8).
The absence of a distinctive polarity sequence in cores of Oligocene age precludes independent correlation of polarity sequences with the geomagnetic polarity time scale. Based on the nannofossil and foraminiferal biostratigraphy, however, the polarity intervals found in cores from lithostratigraphic Unit II (see "Lithostratigraphy" section, this chapter) can be tentatively assigned ages. The reversed polarity found in Core 149-897C39R at 418.6 mbsf may contain the lower part of reversed Chron C9 (corresponding to nannofossil Zone NP24), and the polarity change in Core 149-897C-43R (455.6 mbsf) may correspond to the boundary between Chrons C9 and C10. The magnetostratigraphy for the entire Eocene and Paleocene is unknown at this site. The next reversed polarity interval was found in Core 149-897C-61R. The brown clays in the upper part of the core (Section 149-897C-61R-1) exhibit reversed magnetization on the basis of both pass-through and discrete sample measurements (Fig. 25). The dark brown clays in the middle and lower parts of the core (Sections 149-897C-61R-2 to -4) exhibit normal polarity. Hence, the evidence for reversed magnetization occurs over less than one section.
Cores from lithostratigraphic Unit IV (see "Lithostratigraphy" section, this chapter) and from the basement did not provide any reliable polarity results; and no magnetostratigraphic interpretation is possible for these cores at present.
Although the preliminary magnetic polarity reversals in Hole 897C, summarized in Table 8, are prominent, more detailed paleomagnetic and biostratigraphic studies will be needed to construct a definitive magnetostratigraphy.
The downhole profile of magnetic susceptibility at Site 897 is shown in Figure 26. Poor core recovery and drilling disturbance in Hole 897A resulted in an incomplete record of susceptibility that is difficult to interpret. The magnetic susceptibility of Unit I in Hole 897C ranges from 1.25 to 8.79 -4 SI units, with four large spikes (exceeding 1.88 x 10-3 SI units) in Cores 149-897C-8R (~120 mbsf), -12R (~160 mbsf), -19R (~230 mbsf), and -35R (~380 mbsf), respectively. These susceptibility spikes may have been caused by rust flakes from the drill pipes. The mixture of terrigenous sands and pelagic clays in this interval also produces a distinctive pattern of alternating high and low susceptibilities, which correlates well with turbidite layering (see Fig. 27 for an example). This characteristic susceptibility response might be useful for identifying turbidites using a magnetic susceptibility logging tool. Below 300 mbsf, susceptibility values gradually decrease, corresponding to a decreasing NRM intensity and with a change in color from brown pelagic claystone to green and gray pelagic claystone (e.g., Cores 149-897C-29R and -30R), observed in Unit II. The color change and decrease in susceptibility are consistent with a transition from oxidized iron phases in the Miocene brown clays to reduced iron phases in the late Oligocene age green and gray clays. The suggested boundary between oxidized and reduced iron phases corresponds to the top of a region of high concentration of sulfate, which is apparent in interstitial water analyses (see "Inorganic Geochemistry" section, this chapter). Besides the big spike (3.14 -3 SI units) near 380 mbsf, which is probably an artifact resulting from rust flakes, several smaller-scale susceptibility maxima were observed between 400 and 600 mbsf. These correspond to the clay-rich intervals of the pelagic host sediments and to the bioturbated zones at the tops of the calcareous turbidites in lithostratigraphic Subunit IIC.
Magnetic susceptibility increases from 1.26 -4 (SI units) in Subunit IIIA to 6.28 -3 (SI units) in Subunit IIIB, a trend that reflects the increasing abundance of opaque minerals observed in smear-slides (see "Lithostratigraphy" section, this chapter). A distinctive peak in susceptibility occurs at about 640 mbsf in Holes 897C and 897D (Fig. 27B; Cores 149-897C-62R and -897D-6R). Susceptibility peaks found in both holes exhibit similar magnitudes, and susceptibility plots are similar. Significant changes in rock type, magnetic polarity, and other physical properties at this depth suggest the need for a detailed study of this interval and for regional correlation with the other sites that were drilled during Leg 149.
Sharp increases in susceptibility occur at approximately 640 mbsf in Hole 897C and at 645 mbsf in Hole 897D. These increases correspond to the boundary between Subunits IIIA and IIIB and coincide with the appearance of conglomeratic material. The serpentinized peridotite basement (see "Igneous and Metamorphic Petrology and Geochemistry" section, this chapter) has much higher susceptibility values (averaging 1.88 -2 SI units) than the sedimentary section. The variability of susceptibility below 640 mbsf in Hole 897C reflects differences in the degree of alteration of the original peridotite. Two distinct groups of altered peridotite are indicated (Fig. 28). Moderately low susceptibility values (less than 1.26 -2 SI units) are associated with highly altered serpentinized peridotite, in which nearly all of the olivine and pyroxene have been replaced by serpentine with a clay texture. The least-altered peridotite has the highest susceptibility values (4.2 -2 SI units; see Fig. 28). A similar multimodal pattern was observed in the acoustic velocity, in which samples having different degrees of alteration fall into distinct groups with very different velocities (see "Physical Properties" section, this chapter). Susceptibility measurements in cores from Hole 897D do not follow this trend; instead, they indicate a monotonically decreasing susceptibility, with increasing alteration of the basement rock.