Archive halves of APC (down to ~170 mbsf) and XCB (down to ~588 mbsf) cores recovered at Site 1096 were measured at 5-cm intervals. Measurement of the natural remanent magnetization (NRM) of Cores 178-1096A-1H through 15H, 178-1096B-1H through 22X, and 178-1096C-1H through 9X was done after stepwise alternating field (AF) demagnetization at 0 (NRM), 10, and 20 mT (Tables T4, T5, T6, T7, T8, T9, T10, T11, T12, T13, T14, T15, T16, all also in ASCII format in the TABLES directory). Cores 178-1096B-23X through 32X and Cores 178-1096C-10X through 41X were demagnetized up to 30 mT because of the higher coercivity of the drill-string overprint for these cores. The distribution of inclination values obtained after demagnetization at 20 mT reflects the overall good quality of the magnetic record. It can be clearly defined as the combination of two log-normal distributions centered respectively at -74º and +76º and approaching the geocentric axial dipole values of ±78º expected at 67ºS (Fig. F20).
Analysis of the data collected from discrete samples followed the methods described in "Paleomagnetism" in the "Site 1095" chapter (Tables T17, T18, T19, all also in ASCII format in the TABLES directory). Stepwise AF demagnetization of discrete samples taken from the working halves of cores revealed that the drill-string overprint was dominantly vertically downward, as observed at Site 1095 (see "Paleomagnetism" in the "Site 1095" chapter). The drill-string overprint was mostly or wholly removed by partial AF demagnetization of the NRM at 20 or 30 mT. Samples from the upper 18 m of cores from Site 1096 had a very "hard" magnetization, as AF demagnetization up to 80 mT only removed 30%-60% of the intensity of remanence (Fig. F21). Samples from the underlying sediments had "softer" magnetization, with 80%-90% of the remanence removed after demagnetization at 80 mT (Fig. F22). In both cases, the direction of the characteristic remanent magnetization (ChRM) vector was very stable up to 60 mT. A small number of samples had no stable direction of magnetization.
Inclinations calculated from principal component analysis (PCA) of the discrete sample measurements agreed very well with the inclinations from the magnetically cleaned split cores (Fig. F23A; Table T20, also in ASCII format in the TABLES directory). For the 118 samples analyzed, 85 gave maximum angular deviations between 0º and 5º, 13 gave angles between 5º and 10º, and 20 gave angles >10º (Fig. F23C; Tables T8, T9, T10, T11, T12). In general, samples with low intensities gave larger maximum angular deviation angles and more uncertain ChRM directions (Fig. F23D).
Results obtained from Holes 1096A, 1096B, and 1096C provide a near-continuous paleomagnetic data set down to 588 mbsf. The magnetostratigraphy has been constructed from records of the inclination and intensity of remanence (Figs. F24, F25; Table T21). Declination was not used in the magnetostratigraphic record because the cores were not oriented. The Brunhes/Matuyama (0.78 Ma) reversal occurs in Hole 1096A between 54.60 and 55.40 mbsf, and in Hole 1096B between 54.94 and 54.99 mbsf. The Jaramillo (0.99-1.07 Ma) and the Olduvai (1.77-1.95 Ma) Subchrons are seen in Holes 1096A and 1096B, but their boundaries remain difficult to determine accurately and correspond to intervals of very low intensities of remanence (Fig. F26). These intervals of low intensity are also related to low values of magnetic susceptibility (Fig. F26) and may be understood in terms of local changes in the nature or concentration of the magnetic minerals. The rest of the stratigraphic column displays a succession of polarity intervals, starting in Hole 1096B with the termination of Chron C2An.1n (2.581 Ma) between 215.35 and 217.25 mbsf and ending in Hole 1096C with the onset of Chron C3n.2n (4.62 Ma) between 583.85 and 584.05 mbsf.