The natural remanent magnetization (NRM) of whole-round sections from Cores 202-1236A-1H through 20H and 202-1236B-1H through 9H was measured to help facilitate hole-to-hole correlation. The NRM of archive-half sections of each core was remeasured after alternating-field (AF) demagnetization at selected levels. Core 202-1236A-1H was demagnetized at peak alternating fields of 10, 15, and 20 mT. Section 202-1236A-1H-3 was additionally demagnetized at 25 mT. Core 202-1236A-2H was demagnetized at 15 and 20 mT. Section 202-1236A-2H-6 was additionally demagnetized at 25 mT. Core 202-1236A-3H was demagnetized at 15, 20, and 25 mT. Cores 202-1236A-4H through 24X, all cores from Hole 1236B, and Cores 202-1236C-1H through 11H were demagnetized at 20 and 25 mT. Cores 202-1236C-12H through 18H were demagnetized at 25 mT.
NRM intensities of the whole-round sections varied around 10-2 A/m for the upper ~120 mcd. At greater depths values were generally lower but more variable, alternating between 10-2 and 10-4 A/m. Discrete peaks were more commonly found from Core 202-1036A-9H to the base of the APC-cored interval (the whole rounds of the XCB cores were not measured), with values often >1 A/m. The peaks are generally restricted to core tops and reflect the accumulation of debris in the hole from drilling of the overlying section. After AF demagnetization at peak fields of 20 or 25 mT (Fig. F24), a substantial decrease in intensity is observed with values slightly more than an order of magnitude lower than their predemagnetized level. The downhole pattern of intensity variations is similar to that observed prior to demagnetization and to a large degree mimics the magnetic susceptibility profile. In general, the NRM intensity pattern reflects the major lithologic changes at Site 1236 (see "Lithostratigraphy"). Intensity values within lithologic Unit I are at or below 10-3 A/m for the upper 20 mcd and between 10-3 and 10-4 A/m at greater depths. Within Unit II beginning at ~120 mcd, the intensity values are generally lower (as low as 5 x 10-6 A/m) but are significantly more variable, with some intensities higher than those found in Unit I. Starting with Unit III, intensities begin to increase and continue increasing through Unit IV to a value >10-2 A/m (Fig. F24).
As observed at the previous sites, inclinations prior to demagnetization are steeply positive, characteristic of a drill string-induced magnetic overprint. After demagnetization at either 20 or 25 mT, negative and positive inclinations are observed, suggesting that much of this overprint is removed (Fig. F25). Inclinations steeper than expected are common in intervals of either negative or positive polarities (expected inclination is -37.5° at the latitude of Site 1236 during normal polarity) and may reflect slight coring deformation within these generally unlithified carbonate-rich sediments. A generally positive inclination bias is apparent (the overprint direction), particularly in Hole 1236B (Fig. F25). Because of the overprint problem observed at previous sites (see Lund et al., this volume), the nonmagnetic core barrel was used for even-numbered cores in Holes 1236A and 1236C and for odd-numbered cores in Hole 1236B. Although the effect of the nonmagnetic core barrel on NRM intensity was smaller than observed in the siliciclastic-rich sediments of Site 1235, it did result in more clearly recognizable polarities and reversals.
The directional data from the lower part of lithologic Unit I, beginning at ~100 mcd, and through all of Unit II displays no obvious polarity pattern (Fig. F25). These sediments contain coarse netritic carbonates (see "Lithostratigraphy") deposited by downslope processes. The reliability of any directional information in these transported sediments must, therefore, be questioned. However, the sediments cored with the XCB that comprise lithologic Units III and IV may preserve a reliable polarity sequence. Sediments from ~210 mcd to the base of the cored interval show clear reversed polarity.
Positive and negative inclinations are present within the upper 100 mcd of Site 1236, suggesting that correlation to the geomagnetic polarity timescale (GPTS) (Cande and Kent, 1995) may be possible (Fig. F26). Though normal and reversed polarities are common, few clear polarity transitions are recognized. Some transitions apparently recognized in one hole are not obvious in another hole at the same composite depth. Additionally, many intervals of indeterminate polarity are present, which complicates stratigraphic interpretation. A few unambiguous polarity transitions are observed, including the Gauss (2An)/Matuyama (2r) transition, the boundaries of the Olduvai (2n), the Jaramillo (1r.1n), and the Matuyama (2r)/Brunhes (1n) transition (Table T10).
In view of the difficulty in recognizing clear polarity transitions, we have attempted to provide magnetostratigraphic information for Site 1236 by using the distinct fingerprint provided by the GPTS. Rather than picking discrete polarity reversal boundaries, we identified the magnetic chron(s) to which a sequence of sediment belongs. Because of the low fidelity of records from each individual hole, we stacked the inclination data from Holes 1236A, 1236B, and 1236C based on the composite depth scale (see "Composite Section"). We then took a 10-point running mean of the stacked inclination data and interpreted the polarity sequence from the top down (Fig. F26). This stacked inclination record was correlated by anchoring the top of our record to the top of the GPTS and working downward. The stacked and smoothed record provides a relatively unambiguous match to the GPTS, with many individual polarity zones recognizable. The clarity of this record suggests that much of the noise within in each individual core is random and that stacking and smoothing significantly increased the signal-to-noise ratio. Based on the overall correlation to the fingerprint of the GPTS (Fig. F26), 18 polarity zone identifications were made within the upper 100 mcd of Site 1236. Ages for the polarity are given as midpoint averages of the polarity interval ages of Cande and Kent (1995). Depth errors are assumed to be plus or minus one-half the depth interval of the interpreted polarity zone (Table T11; Fig. F26). Other polarity events may be present in the smoothed stack record, suggesting that shore-based research may be able to resolve a more detailed magnetostratigraphy.