Archive halves of all cores recovered at Site 1095 were measured at 4-cm intervals. For Core 178-1095A-1H, the natural remanent magnetization (NRM) was measured after partial alternating field (AF) demagnetization at the 0 (NRM step), 10, 20, and 25 mT. For Cores 178-1095A-2H through 10H and Cores 178-1095B-1H through 8H, measurements were obtained after partial AF demagnetization at 0, 10, 20, and 30 mT (Tables T5, T6, T7, T8, T9, T10, T11, T12, T13, T14, T15 all also in ASCII format in the ASCII TABLES directory). For Cores 178-1095B-9H through 52X and Hole 1095C, the AF demagnetization commenced at 20 mT (Tables T11, T12, T13, T14, T15, T16, T17, T18 all also in ASCII format in the ASCII TABLES directory). The decision to start demagnetization at 20 mT was justified by the coercivity of the drill-string overprint and the need to maintain a constant core flow through the paleomagnetics laboratory. For Hole 1095D, the AF demagnetization scheme was altered to 0 , 10 , and 20 mT (Tables T19, T20, T21, T22, T23, all also in ASCII format in the ASCII TABLES directory) to preserve the remanence for future U-channel studies. Besides providing access to the unprocessed data, we include processed data from each hole (Tables T5, T6, T7, T8, T9, T10, T11, T12, T13, T14, T15, T16, T17, T18, T19, T20, T21, T22, T23), for which processing consists of removing results from within 8 cm of the ends of each core section (these are anomalous owing to magnetic edge effects) and from drilling-disturbed intervals. In addition, the raw data are available in the Ocean Drilling Program database.
In all holes, the magnetization of the cores contained a steep, downward component, which simulated a reversed polarity direction at low (0-10 mT) demagnetization levels (Fig. F16A). In addition, either a radial overprint or other biases, perhaps instrumental, resulted in nearly all of the declinations having a bias toward 0º (Fig. F16B), even after demagnetization.
Discrete samples were collected from the working halves of cores to carry out full AF demagnetization up to fields as large as 80 mT (Tables T24, T25, T26) all also in ASCII format in the ASCII TABLES directory). Stepwise demagnetization up to high fields is used (1) to reveal the peak field necessary to clean the remanence vector of the drill-string overprint, (2) to assess the stability of the remanence vector, (3) to determine how many magnetization components are present, and (4) to estimate the direction and uncertainty of the characteristic remanent magnetization (ChRM, i.e., the highest coercivity component that is stable) of each sample through principal component analysis (PCA). We also conducted very preliminary rock magnetic investigations, which mainly were restricted to anhysteretic remanent magnetization and susceptibility measurements. These measurements were made to estimate a preliminary relative paleointensity record at Site 1095 and to assess the quality of this record for postcruise research.
Demagnetization revealed that the drill-string overprint could be mostly or wholly removed by partial AF demagnetization at 20 or 30 mT and that the characteristic remanence was very stable up to the 60-mT demagnetization step (Fig. F17). A few discrete samples have three components of magnetization: a low-coercivity drill-string overprint that is removed with ~10-mT demagnetization, a second component that is removed between 10 and 50 mT, and a third high-coercivity component that is stable during demagnetization up to 80 mT (Fig. F18). These samples typically come from silty intervals, and the medium-coercivity component is consistent with adjacent samples that are fine grained. Hence, this medium-coercivity component is probably the primary remanent magnetization in these silty intervals. A small number of samples gave no stable magnetization direction.
Inclinations calculated from PCA of the discrete sample measurements agree very well with the inclinations from the magnetically cleaned split cores, except in a few samples as noted above (Fig. F19). A computer program was written to automate the PCA. The program iteratively searched for the demagnetization steps that minimized the size of the maximum angular deviation angle, which measures how well the vector demagnetization data fit a line. The program requires that at least four demagnetization steps be used, never uses steps lower than 10 mT, does not require that the best-fit PCA line pass through the origin (the "free" option of standard PCA), and generally favors high-coercivity components over low-coercivity components in samples with multicomponent magnetizations.
For the 302 samples analyzed, 201 give maximum angular deviation angles that fall between 0º and 5º, and 54 give angles between 5º and 10º (Table T27, also in ASCII format in the ASCII TABLES directory). These 255 samples have very well-constrained ChRMs, whereas those with higher maximum angular deviation angles should be treated with caution. In general, samples with low intensities give larger maximum angular deviation angles and more uncertain ChRM directions (Fig. F19C, F19D). An additional 36 samples were collected from Core 178-1095B-34X to examine the behavior of the field during Chron 4r.2r (see "Identification of Cryptochron 4r.2r-1 in Hole 1095B"), with 31 of these having maximum angular deviation angles <5º.
Results obtained from Holes 1095A and 1095B form a near-continuous paleomagnetic data set for the upper 450 mbsf. The magnetostratigraphy was constructed from records of the inclination and intensity of remanence. Declination was not used in the magnetostratigraphic record because the cores were unoriented, because of the bias in the declination mentioned above (Fig. F16B), and because the declination contributes little to the total paleomagnetic vector, which is very steep at Site 1095. The overall quality of the paleomagnetic record is very good, with clear magnetic polarities characterized by steep inclinations (averages = +73º, -71º) approaching the geocentric axial dipole value of ±78º expected at 67ºS.
The inclination record provided a reliable magnetostratigraphy down to 460 mbsf and comprised 208 m of APC and 252 m of XCB core (Figs. F20, F21; Table T28). The Brunhes/Matuyama (0.78 Ma) boundary occurs in Hole 1095A at ~17.1 mbsf (Section 178-1095A-3H-6), and the Jaramillo Subchron (0.99-1.07 Ma) between 23.1 and 29.1 mbsf. The Jaramillo Subchron is not observed in Hole 1095D, possibly destroyed by core flow-in or masked by the drill-string overprint. The short interval of normal polarity observed between 40.8 and 41.2 mbsf in Hole 1095A may be associated with the Cobb Mountain Event (1.2 Ma). The polarities of underlying sediments show a succession of two large normal periods interrupted by a shorter reversed interval. This in turn suggests a record of Chron C2An instead of C2n and the possibility of a hiatus in the sedimentary sequence between 50 and 65 mbsf. Seismic stratigraphy suggested an unconformity at ~60 mbsf coincident with a prominent lithostratigraphic boundary (see "Seismic Stratigraphy" and "Lithostratigraphy"). The reversal at ~58.8 mbsf has been interpreted as the termination of the Olduvai (1.77 Ma). It is probably underlain by a hiatus such that the onset of the Olduvai and all of Chron C2r was lost.
The discontinuous core recovery at the base of Hole 1095B prevented a confident identification of the sparse directions from 480 to 560 mbsf. Downhole logging using the GHMT provided a complete polarity stratigraphy at the base of Hole 1095B (see "Downhole Measurements"). When a 5-m offset was applied to the logging depth scale, the locations of polarity transitions observed in the core and in the drill hole agreed extremely well. The logging record confirmed that the polarity transition observed in the core at 460.7-461.04 mbsf is the termination of Chron C4Ar.1n. The deepest event that could be confidently identified in the paleomagnetic data is the termination of Chron C4Ar.2n (9.58 Ma) at 522.7-522.9 mbsf.
Interpretation of the magnetostratigraphy from the onset of Chron 4Ar up to the onset of Chron 2An is fairly simple, with the exception of an extra normal polarity event near the base of Chron 4r.2r (Fig. F22) that is not shown in most geomagnetic polarity time scales (GPTSs). Cande and Kent (1992a, 1995) have, however, placed a geomagnetic event in the lower part of Chron 4r.2r, which they refer to as Cryptochron C4r.2r-1.
Cryptochrons are geomagnetic events of short duration (<20 k.y.), which may represent full polarity reversals, directional changes in the field that are incomplete reversals (such as some excursions), geomagnetic intensity variations, or some combination of these. Most of the cryptochrons listed by Cande and Kent (1992a, 1992b, 1995) were identified from the small-scale magnetic anomalies (referred to as "tiny wiggles") observed in marine magnetic anomaly profiles.
Neither the nature nor the duration of a cryptochron can be uniquely ascertained from marine magnetic anomaly profiles alone. Hence, observing one of these events in a magnetostratigraphic record has important implications for geomagnetism and for future refinement of the GPTS.
By identifying the normal polarity interval between 399.9 and 404.8 as Cryptochron 4r.2r-1, we obtained a pattern of reversals that fits the GPTS very well and gives fairly constant sedimentation of 6-11 cm/k.y. for the interval from 450 to 325 mbsf. Within this interval, the interpreted magnetostratigraphy agrees very well with the biostratigraphy. Given these two observations, we were confident in the interpretation, and thus decided to investigate further the nature of the cryptochron.
We conducted detailed AF and thermal demagnetization experiments on 36 discrete samples that span the cryptochron (Figs. F23, F24, F25, F26). These confirmed that the magnetizations within the normal polarity cryptochron and the bounding reversed polarity intervals are fairly simple, with the drill-string overprint being removed by demagnetization at ~10 mT or 250ºC. Beyond this, the demagnetization data reveal univectorial behavior, except in transition intervals and a few other intervals that have low relative paleointensity.
Thermal demagnetization reveals that ~80% of magnetization decays between 275º and 425ºC, which indicates that the dominant remanence-carrying mineral is titanomagnetite. Another 10% of the total magnetization is removed between 525º and 625ºC, which suggests the presence of magnetite and possibly hematite. The remanence present above the 580ºC step could result from the temperature of the oven being inaccurate by 20º-40ºC because the sample is completely demagnetized above the 600ºC step. In any case, after removal of the drill-string overprint, the remanence-carrying minerals for a single sample all give the same direction.
The simple demagnetization behavior and the agreement between discrete and split-core measurements indicate that the split-core results after 30 mT provide an accurate, detailed record of the cryptochron. First, the inclination record indicates that the cryptochron is a full geomagnetic reversal. The declinations, on the other hand, are meaningless in these XCB cores because core deformation produces biscuits (pieces of core) that are azimuthally unoriented with respect to one another. Because of the steepness of the inclination at this site, however, the normal and reversed directions must be nearly antipodal even in the absence of declination observations. Second, the relative paleointensity record, derived by dividing the intensity (after 30-mT demagnetization) by the multisensor track (MST) susceptibility, shows that the cryptochron is also a zone of low paleointensity, at least with respect to adjacent chrons (Fig. F27). Third, the transition zones between the normal polarity cryptochron and adjacent reversed polarity intervals are relatively sharp, each spanning 20-30 cm or ~2-4 k.y. The paleointensity collapses to zero within the transition zones but recovers within a few thousand years. Fourth, given that sedimentation rates are between 6 and 11 cm/k.y., the duration of the cryptochron is 45-82 k.y., several times larger than the 16 k.y. estimated by Cande and Kent (1995) from marine magnetic anomalies. Of course, all these interpretations are preliminary, and more work is planned to investigate these properties and alternative magnetostratigraphic interpretations.