Core archive-halves from Hole 1121B were measured on the shipboard pass-through cryogenic magnetometer. Declination, inclination, and intensity of natural remanent magnetization (NRM) were routinely measured at 5-cm intervals at 10- and 20-mT alternating field (AF) demagnetization steps. In situ Tensor tool data were not collected as only three APC cores were collected before refusal, and coring continued with the XCB method. Therefore, only inclination could be used to determine the magnetic polarity of Hole 1121B. At least two oriented discrete samples were collected from the working half of each core interval for progressive AF and thermal demagnetization and rock magnetic studies. Whole-core magnetic susceptibility was measured routinely on all cores, using a Bartington susceptibility loop on the automated multisensor track (MST).
Magnetic properties vary with lithology downcore (Fig. F13). In lithostratigraphic Subunit IA (0-15.2 mbsf), a bioturbated, yellowish brown, silty and sandy clay, NRM intensities mimic susceptibility and are relatively high (about 10-2 A/m and about 10-4 SI volume units, respectively). Beneath a sharp contact, Subunit IB (15.2-32.7 mbsf, pale yellow clay) has much lower NRM intensities (about 10-4-10-5 A/m on average), but susceptibility remains moderately high (10-5 SI). Susceptibility and NRM intensities are both low (10-6 SI and 10-4-10-5 A/m, respectively) in Unit II, a mostly light green pelagic diatomaceous ooze (32.7-134.5 mbsf, base of hole).
In the upper 28 m of the core, NRM measurements displayed consistent, steeply positive (downcore) inclinations ranging between +70° and +80° , consistent with a drill-string overprint induced during coring. The 20-mT AF demagnetization step proved very effective in removing the overprint and revealing a well-defined polarity reversal stratigraphy (Fig. F14). Intensity of magnetization in lithostratigraphic Unit I was strong enough to subject discrete samples to stepwise AF demagnetization of up to 80 mT in the shipboard pass-through cryogenic magnetometer. In each case, the drilling-induced overprint accounted for 30%-80% of the NRM but was removed by the 10-mT step of demagnetization (Fig. F15). Further stepwise AF demagnetization was mostly effective in isolating a primary remanence direction, but the signal was generally noisy (Fig. F15). A single sample (26.04 mbsf, Fig. F15D), taken from the interval of "flow-in" from within Subunit IB, had a horizontal drilling-induced overprint but moderate inclination upon cleaning. Steep paleomagnetic inclination has been reported to distinguish "flow-in" from other intervals of core (e.g., Roberts et al., 1996); however, inclination was not distinctly different in this case. After AF cleaning, selected discrete samples from Unit 1 were subjected to progressive isothermal remanent magnetization (IRM) fields until saturation (Fig. F16A). Magnetization was "soft" and samples were generally saturated by 200 mT, although some samples did not saturate fully until 500 mT. Coercivity of remanence (Bcr) was less than 50 mT in all samples. AF and thermal demagnetization of the saturation IRM (SIRM) shows two components of magnetization with overlapping coercivity and blocking temperature spectra (Figs. F17A, F17B, F17C). The lower blocking temperature component is removed by 300° - 400° C and the high-temperature component remains until 580° - 600° C, although a very minor component may persist above 600° C.
The "soft" SIRM, low Bcr, and distributed unblocking temperatures up to 580° - 600° C demonstrate that multiple grain-size magnetite is the main carrier of remanence in samples from Unit I. However, a small component of remanence may be held by a sulfide mineral, such as pyrrhotite, which would explain slightly harder IRM acquisition curves and higher values of saturation magnetization for some samples (Fig. F16A), as well as the low unblocking temperature component of SIRM in all samples (Fig. F17). The presence of sulfides may also explain why AF demagnetization of NRM produces noisy results (e.g., Fig. F15C) as pyrrhotite is known to respond poorly to AF demagnetization methods (e.g., Turner and Kamp, 1990; Roberts and Turner, 1993; Wilson and Roberts, 1999).
The NRM inclination record in Unit II is very noisy and not obviously overprinted by the drilling process. This is most likely a result of the very low intensity of magnetization throughout Unit II. The 20-mT AF demagnetization step is still noisy; however, it is possible to distinguish intervals of dominantly positive vs. intervals of dominantly negative inclination. Because of the low intensities of magnetization and poor response to AF cleaning of archive core halves on the shipboard pass-through cryogenic magnetometer, further AF cleaning of discrete samples was not attempted, but several discrete samples were given an IRM. Typical IRM and backfield IRM acquisition curves are shown in Figure F16B. Sample magnetizations are "harder" than those from Unit I and did not saturate until applied fields of 500-1000 mT. Bcr values were between 75 and 100 mT. Alternating-field demagnetization to 60 mT only removed 50% of the SIRM. Thermal demagnetization of the SIRM shows two components of magnetization with overlapping blocking temperature spectra (Figs. F17D, F17E, F17F). The lower unblocking temperature component is again removed by 300° - 400° C, but it is a more prominent component in Unit II than it was in Unit I. The high-temperature component persists until 580° - 600° C, except in one case where it persists to 680° C (Fig. F17F).
The "harder" SIRM and higher Bcr in Unit II suggests that an iron sulfide mineral (possibly pyrrhotite) is more dominant in holding remanent magnetization than it was in Unit I. However, even though less dominant, magnetite of multiple grain sizes is still present, as demonstrated by the distributed higher-temperature unblocking spectra that generally persist until 580° -600° C (Figs. F17D, F17E, F17F). A minor component of magnetization may once again persist to higher temperatures in a few samples. This is most likely to be caused by hematite from rust contamination during the drilling process. Unblocking temperature spectra in Figure F17 clearly show that the lower blocking temperature spectra component is more dominant in Unit II (Figs. F17A, F17B, F17C) than in Unit I (Figs. F17D, F17E, F17F). The dominance of pyrrhotite may well explain the low NRM intensities for Unit II, the lack of a drilling-induced overprint component to the NRM, and the ineffectiveness of AF demagnetization in isolating a clear primary inclination pattern for Unit II. It is possible that the AF demagnetization was only effective in removing any drilling-induced component of the NRM in Unit II.
Tensor tool data were not collected in core from Hole 1121B because of early APC refusal, so only inclination could be used to determine polarity. In Unit I, after 20 mT of demagnetization, the inclination pattern allows a clear determination of magnetic polarity stratigraphy for the upper 22 mbsf of the drill hole (Fig. F14). Four major normal polarity magnetozones were identified (M2, 1.05-2.31 mbsf; M4, 3.71-9.44 mbsf; M6, 14.70-19.02 mbsf; and M8, 21.17-22.00 mbsf). Polarity could not be unambiguously determined in the uppermost meter of Hole 1121B, probably because this interval comprises saturated, unconsolidated, well-sorted, medium sand. However, it is possible that this is an interval of reversed polarity. Polarity could also not be determined for the lowermost part of Unit I because of a combination of "flow-in" during coring and loss of recovery. A Pliocene diatom assemblage was recovered from the top 3 m of Hole 1121B (see "Biostratigraphy"). Based on this, the uppermost normal polarity interval (1.05-2.31 mbsf) is tentatively correlated with the lower part of the Gauss magnetochron of the Geomagnetic Polarity Time Scale (GPTS) (Berggren et al., 1995; Cande and Kent, 1995). This correlation would imply that at 2.3 mbsf the sediments are already 3.6 Ma old and either the uppermost part of the hole is very condensed, or the Pleistocene is not recorded at Site 1121. The frequent polarity reversal pattern within Unit I (Fig. F14) also suggests quite slow sedimentation rates and a condensed record. Unfortunately, no biostratigraphic information is available to assist with correlation of any of the magnetozones in Subunit IB to the GPTS, and the magnetic polarity pattern is not unique and cannot be correlated independently. Furthermore, several disconformable horizons are identified where time may be missing (see "Lithostratigraphy"). However, only the reversal boundary between M6 (normal) and M7 (reversed) occurs in a lithostratigraphic break marked by a chert bed; all other reversal boundaries do not appear to occur in lithostratigraphic breaks.
A major lithostratigraphic disconformity is noted between Units I and II. Diatoms, radiolarians, nannofossils, and foraminifers recovered from Unit II are confined to the Paleocene and confirm a major unconformity in the vicinity of the chert horizons in Cores 181-1121B-3H, 4H, and 5H (19.0-32.7 mbsf, see "Biostratigraphy"). In Unit II, the inclination signal is noisy, because of the ineffectiveness of AF demagnetization, and the polarity record disjointed by the many intervals of poor or no recovery. Despite this, four distinct polarity zones can be identified (Fig. F13): a reversed-normal-reversed polarity sequence between 42.82 and 93.27 mbsf, and an interval of reversed polarity between 119.65 and 127.18 mbsf. All reversal boundaries lie within unrecovered intervals at core breaks, except for a possible transition to normal polarity at 93.27 mbsf at the base of the reversed-normal-reversed (R-N-R) sequence (Fig. F13). Radiolarian zones (Bryella tetradica, RP5, and Bekoma compechensis, RP6) and nannofossil zones (NP5-NP8) allow direct correlation of the R-N-R sequence with Chrons C26r-C26n-C25r (see "Biostratigraphy") (Fig. F18). The combination of the Bryella foremanae (RP4) radiolarian zone and the nannofossil NP4 zone beneath 120 mbsf suggests that the reversed polarity zone (119.6-127.18 mbsf) in the lower part of the hole must correlate with part of Chron 27r of the GPTS. If the polarity reversal at 93.27 mbsf is real, it must by definition correlate with the top of C27n of the GPTS. However, this is in disagreement with the nannofossil zone (NP5) throughout this interval (see "Biostratigraphy"). Because it occurs just above an interval of nonrecovery, more detailed work is necessary to be sure polarity has been correctly identified. Correlation with the GPTS is well constrained by biostratigraphic datums and suggests sedimentation rates of about 15 m/m.y. for the Paleocene interval in Hole 1121B (about 40-130 mbsf), as opposed to much lower sedimentation rates at the top of Hole 1121B (possibly about 1 m/m.y.). Thermal demagnetization will be undertaken on discrete samples from the Paleocene interval of Hole 1121B to confirm and better define the polarity stratigraphy reported here from shipboard studies.