Core archive halves from Holes 1123A, 1123B, and 1123C were measured on the shipboard pass-through cryogenic magnetometer. Declination, inclination, intensity of natural remanent magnetization (NRM), and a 20-mT alternating field (AF) demagnetization step were measured at 5-cm intervals. The first few cores of each hole were also measured after a 10-mT demagnetization step; this step added little extra information and, because of time constraints, only the 20-mT step was continued. In situ tensor tool data were collected for all APC cores, but a problem with the shipboard pass-through cryogenic magnetometer prevented the use of declination for polarity determination in the APC cores. Therefore, only inclination could be used to determine magnetic polarity in Holes 1123A, 1123B, and 1123C. At least two discrete oriented samples were collected from the working half of each core interval for progressive AF and thermal demagnetization and for rock magnetic studies. Whole-core magnetic susceptibility was measured on all cores using a Bartington susceptibility loop on the automated multisensor track (MST).
A composite paleomagnetic record was constructed for Site 1123 using data from Hole 1123C (0-140 mcd and 500-635 mcd) and Hole 1123B (140-500 mcd). Hole 1123A was not used in constructing a composite record, as paleomagnetic results for the uppermost 90 mcd were not easy to interpret and not reproduced in Holes 1123B and 1123C. Discussion with the drilling team suggested that possible early failure of the shear pin may have contributed to this difference in paleomagnetic signal. A slower rate of penetration of the piston core and, hence, more disturbance of the core occurred during piston firing. Intensity of magnetic remanence varied markedly with depth in the composite record (Fig. F19). The upper 280 mcd has average NRM intensities of 4 x 10-4 A/m and these intensities increase steadily downcore to 1 x 10-3 A/m at 550 mcd. The lowermost 38 m (597-635 mcd) has average NRM intensities on the order of 10-5 A/m.
NRM measurements displayed consistent, steeply positive (downcore) inclinations ranging between +70° and +80° , consistent with a drill-string overprint induced during coring. The single 20-mT AF demagnetization step proved very effective in removing the overprint and elucidating a polarity reversal stratigraphy (Fig. F19). Figure F20 shows representative acquired isothermal remanence magnetizations (IRM) to saturation (SIRM) and backfield SIRM.
All samples above 597 mcd exhibited very uniform behavior, were saturated by 500 mT, and had backfield coercivities of remanence (Bcr) values of 40-80 mT (Fig. F20A). Alternating field and thermal demagnetization of the SIRM of these samples demonstrated moderate to soft magnetization with 60%-90% of intensity of remanence lost by the 60-mT AF demagnetization step (Fig. F21). Samples from the Eocene/Oligocene limestones beneath the hiatus at 597 mcd (see "Lithostratigraphy") all had much harder magnetizations that did not saturate until IRM fields of 1 T (Fig. F20B). AF demagnetization was less effective in removing saturation magnetizations from discrete samples. Thermal demagnetization to zero intensity was not possible in samples above 597 mcd, as alteration of clay components of the sediment from stepwise heating caused new mineral growth above 500° C. This was detected by an increase in magnetic susceptibility, and further heating was discontinued. The general trend in each case was to zero remanence around 580° C (Fig. F21A, F21B). This, along with low coercivity and low Bcr values, suggests that the main remanence carrier is magnetite of distributed grain size. Below the hiatus at 597 mcd, however, intensity of remanence persisted above 600° C in all cases (Fig. F21C). The high unblocking temperature spectra, the high saturation IRM fields, and higher Bcr values (up to 100 mT) suggest hematite may contribute to the magnetic remanence of lithostratigraphic Unit IV.
Interpretation of magnetic polarity from the composite inclination record for Site 1123 (Fig. F19) is well constrained by key foraminiferal, nannofossil, and diatom datums from each of the core-catcher samples (see "Biostratigraphy," also see "Biostratigraphy" in the "Explanatory Notes" chapter). The inclination record for the upper 130 mcd of the composite record is taken from Hole 1123C and, when compared with the Geomagnetic Polarity Time Scale (Cande and Kent, 1995; Berggren et al., 1995), it provides a near-complete record of the Brunhes (C1n), Matuyama (C1r-C2r), and Gauss (C2An) magnetochrons (Fig. F19A). The uppermost 32.5 mcd of the polarity record is entirely normal, contains the last occurrences of the diatom Nitzschia reinholdii (0.65 Ma), nannofossil Pseudoemiliania lacunosa (0.42 Ma), and acme first occurrence (FO) of Hemidiscus karstenii (diatom, 0.42 Ma) and is assigned to the Brunhes (C1n) Chron. The characteristic normal-reversed-normal-reversed-normal pattern of the Gauss Chron occurs between 100 and 130 mcd, constrained by the LO of nannofossils Reticulofenestra pseudoumbilicus (3.8 Ma) and Discoaster tamalis (2.8 Ma), and foraminiferal datums Globorotalia puncticulata (LO, 3.8 Ma), G. crassaconica (LO, 3.2 Ma), G. crassula (FO, 2.6 Ma), and the range of the dextral form of G. crassaformis (2.1-3.2, Fig. F19A). Between 32.5 and 130 mcd, the polarity is mostly reversed with three short normal events. Two of these short normal polarity events immediately underlie the Brunhes Chron and are assigned to the Jaramillo (C1r.1n) and Cobb Mountain Subchrons, respectively. The other short normal polarity event occurs between 70 and 73 mcd and, constrained by nannofossils Gephyrocapsa (medium) (FO, 1.7 Ma) and D. brouweri (LO, 1.96 Ma), is assigned to Chron C2n (Olduvai). The inclination record between 94 and 100 mcd is ambiguous and not able to be designated as normal or reversed polarity. It is labeled "unknown" in Figure F19A, though it most likely lies within the upper part of Subchron C2An.1n (upper Gauss).
The inclination record between 130 and 260 mcd (Fig. F19B) is taken from Hole 1123B. It is more noisy than that for the upper 130 mcd but still allows interpretation of polarity zones. Between 130 and 213 mcd, polarity is mostly reversed with five short normal polarity events. Several foraminiferal events (short range of G. sphericomiozea, ~5.6 Ma, FO G. puncticulata, 5.2 Ma, FO G. crassaconica, 4.7 Ma, and LO G. mons, 4.8 Ma) and the FO of D. asymmetricus (nannofossil, 4.1 Ma) constrain this interval to the Gilbert Reversed Chron (C2Ar-C3n-C3r). The four short normal polarity subchrons between 150 and 185 mcd are characteristic of Chron C3n and are correlated with the Cochiti (C3n.1n), Nunivak (C3n.2n), Sidufjall (C3n.3n), and Thvera (C3n.4n) Subchrons, respectively. The lowermost of these (Thvera, C3n.4n, 177-185 mcd) is not well defined because most of the record at that depth was disturbed during the drilling process. The characteristic R-N-R-N polarity record of the latest Miocene (Chrons C3Ar and C3An) is observed between 213 and 265 mcd, and marks the base of the Gilbert Chron (Fig. F19B). It is constrained by the LO of the nannofossil D. quinqueramus (5.5 Ma) and by Bolivinita pentaspinosa (~7 Ma). The short interval of normal polarity between 200 and 201.5 mcd has no equivalent in the polarity time scale of Cande and Kent (1995).
The remaining polarity record of the late Miocene is very complex with many short reversals that make correlation with the GPTS difficult. However, the long normal polarity chron, C5n (9.74-10.95 m.y.), is quite characteristic and marks the early part of the late Miocene record. A thick interval of normal polarity is observed between 350 and 380 mcd in the composite magnetic polarity record of Site 1123 (Fig. F19B). Constrained by the nannofossil datums LO Minylitha convallis (9.34 Ma) and FO Coccolithus miopelagicus (10.9 Ma) and foraminiferal datums LO G. druryi (11.2 Ma) and LO Globoquadrina dehiscens (10.1 Ma), this interval of normal polarity is correlated with Chron C5n.2n of the GPTS. Between 260 and 350 mcd, nine normal polarity intervals of the GPTS are recognized, and, like the GPTS, they are grouped into two distinct intervals: the upper five mark an interval that is of dominantly normal polarity between 265 and 300 mcd. These events are correlated with the subchrons of Chrons C3Bn-C4n. The remaining four normal polarity intervals lie directly above C5n and are separated from the upper five normal polarity intervals by a moderately thick interval of reversed polarity (300-319 mcd) that contains the FO of the nannofossil D. quinqueramus (8.3 Ma) and the diatom Hemidiscus ovalis (7.9 Ma) and is correlated with Chron C4r of the GPTS. The lower four normal polarity intervals are correlated with Chron C4An and Subchrons C4Ar.1n, C4Ar.2n, and C5n.1n, respectively (Fig. F19C). The very short normal event, C4r.1n, is not observed in the record from Site 1123. These correlations show that the late Miocene record at Site 1123 is remarkably complete and possesses strikingly uniform sedimentation rates.
The middle Miocene polarity record is equally complex and comprises 12 normal and 12 reversed chrons and subchrons. Again, the earliest part of the epoch is marked by a long chron, this time of reversed polarity (Chron C5Br). The correlative polarity record from Site 1123 (380-505 mcd, Fig. F19C, F19D) is complicated by poor recovery in the lowermost interval of Hole 1123B (see summary log in "Lithostratigraphy"). However, a thick interval of reversed polarity between 482 and 499 mcd contains the foraminiferal evolutionary transition of G. miozea to G. praemenardii dated in New Zealand at ~15.8 Ma (Scott, 1979; Morgans et al., 1996; see "Biostratigraphy") and is correlated with C5Br. Between 380 and 482 mcd of Hole 1123B, no less than 13 individual normal polarity events were recognized in initial observations of the polarity record. This suggests that the polarity record is complete, despite the poor recovery. It is surmised that the loss of recovery (between 40% and 60%) must be distributed throughout each individual coring interval, most likely by abrasive loss of material at drilling-induced breaks in the core from the XCB drilling process. The retained core thus represents a compression of the 9.5-m-long cored interval into the reduced record actually recovered. In order to test this hypothesis, the whole-core magnetic susceptibility and inclination records from each core interval (see "Composite Depths") were stretched linearly from their recorded length to fill each 9.5-m coring interval (Fig. F22), and the stretched records compared with the downhole logging record of magnetic susceptibility. The stretched whole-core magnetic susceptibility is very comparable to the downhole log susceptibility record and was, therefore, used to characterize the magnetic polarity stratigraphy for Cores 181-1123B-41X through 50X (Fig. F19D). The resulting stretched polarity record is, on average, of reversed polarity and more detailed correlation with the GPTS is afforded by several foraminiferal and nannofossil events. Between 380 and 404 mcd polarity is mostly reversed and correlated with C5r of the GPTS. At ~405 mcd, the LO of the nannofossil Calcidiscus premacintyrei and the diatom N. denticuloides (12.2-12.6 Ma and 11.3-13.5 Ma, respectively) suggests that the N-R-N polarity pattern between 404 and 421 mcd (Fig. F19D) correlates with Chron C5A of the GPTS. The foraminiferal evolutionary transition of G. praemenardii to G. miotumida (~13 Ma) between ~435 and 455 mcd suggests that the normal polarity events in this interval correlate with Subchrons C5Ar.1n, C5Ar.2n, and Chron C5AAn, respectively. The LO of the nannofossil Sphenolithus heteromorphus (13.5-13.6 Ma) at ~463 mcd suggests that the normal polarity events between 457 and 482 mcd correlate with Chrons C5Abn, C5Acn, C5ADn and Subchrons C5Bn.1n and C5Bn.2n. Within the upper Miocene, two short reversals are recognized in the polarity record from Site 1123 that do not have equivalents in the GPTS (Fig. F19D). A short reversed polarity interval separates Subchron C5Ar.1n (432.6-435.3 mcd) into two normal polarity events, and another short reversed polarity interval separates Chron C5ADn (471.6-474.45 mcd) into two normal polarity events.
Foraminifer and nannofossil events (see "Biostratigraphy") define a long hiatus in the record from Site 1123 where lowermost Oligocene strata are separated from middle to lower Miocene strata by a distinct Cruziana ichnofacies horizon at 596.7 mcd (see "Lithostratigraphy"). The early Miocene polarity record in Hole 1123C is complete down to and including Chron C6n and the upper part of Chron C6r (Fig. F19D, F19E). The acme of the foraminifer G. zealandica (16.7-18.6 Ma), the FO of G. miozea (16.7 Ma), the FO of the nannofossil C. premacintyrei (17.4 Ma), and the LO of the foraminifer G. bella (16.3 Ma) constrain the four normal polarity events between 499 and 522 mcd to the three subchrons of Chron C5Cn and Chron C5Dn, respectively, and the reversed polarity interval between 522 and 534 mcd to Chron C5Dr (Fig. F19E). Continuing down, still constrained by the acme of the foraminifer G. zealandica (16.7-18.6 Ma) and also by the FO of the nannofossil S. heteromorphus, the R-N-R pattern between 522 and 546 mcd correlates with Chrons C5Dr, C5En, and C5Er (Fig. F19E). A debris flow between 552 and 562 mcd (see "Lithostratigraphy") complicates interpretation of the remaining early Miocene polarity record, but the complication is not severe as we consider it a single instantaneous event. Two biostratigraphic datum events occur within the material of the debris flow: the LO of the foraminifer G. incognita (18.5-19.4 Ma) and the FO of the nannofossil S. belemnas (19.2 Ma). Although the ~10-m sampling interval of core catchers prevents precise location of datums, and the fossils have been reworked from older strata, they still afford age constraints on the debris flow, as it must be younger than the fauna that it contains. By inference, the strata immediately overlying the debris flow must also be younger than the fauna contained within the debris flow; hence, the interval of normal polarity between 546 and 552 mcd must correlate with the upper part of Chron C6n of the GPTS. The inclination pattern between 552 and 562 mcd provides additional constraint on the thickness of strata affected by the debris flow. The signal is both noisy and of intermediate direction, rather than of truly normal or reversed polarity, and probably represents a fabric direction imparted by the debris-flow regime. It is assumed that the deposition of the debris flow was essentially instantaneous and also that because it was in a depositing phase it did not cause much erosion at its base. Using these constraints and assuming similar sedimentation rates beneath the debris flow as those above, and also constrained by the FO of G. incognita (21.6 Ma) within the hiatus at 597 mcd, the normal polarity between 562 and 579 mcd is correlated with the remainder of Chron C6n and the underlying reversed polarity interval (579-697 mcd) with Chron C6r (Fig. F19E).
The hiatus at 597 mcd is not constrained by magnetic polarity as it occurs within an interval of continuously reversed polarity. Beneath the hiatus, however, the nannofossil C. subdistichus acme (33.3 Ma) suggests an earliest Oligocene age for the strata immediately underlying the hiatus and that the normal-reversed polarity events represent C13n and the lowermost part of C12r, respectively. Three more datums confirm that the record beneath the hiatus extends into latest Eocene time; the FO of the foraminifer Paragloborotalia gemma (~35 Ma), the LO of G. index (34.3 Ma), and the LO of the nannofossil D. saipanensis (34.2-35.4 Ma). The nannofossil Isthmolithus recurvus occurs to the base of Hole 1123C and its first occurrence must be lower than the base of Hole 1123C. Its FO is ~36 Ma (Wei, 1992; Wilson et al., 1998) and the base of Hole 1123C must therefore be younger than 36 Ma. Despite poor recovery and a weak and noisy paleomagnetic signal, a normal polarity event is identified between 625 and 627 mcd and is correlated with C15n, and reversed polarity is identified between 609 and 625 mcd and is correlated with Chron C13r. It is possible that, with further work, the interval between 616 and 619 mcd may prove to be normal and thus correlate with C15n and move the polarity interpretation to be slightly older, but present analysis of the polarity of this interval is still ambiguous.
In the 597-m Neogene composite record of Site 1123, all but two polarity subchrons of the GPTS are present. Absent are the Reunion normal polarity subchron (C2r.1n) and Subchron C5r.2n. However, three new polarity events are recognized: one is a normal polarity event within Chron C3r (Fig. F19B), the other two are reversed polarity events within Subchrons C5Ar.1n and C5An, respectively. A paleomagnetic age model for Site 1123, listing depths and ages of magnetic polarity reversal events, is given in Table T11.
An age-depth summary for Site 1123, using the GPTS reversal polarity ages given in Berggren et al. (1995b), is given in Figure F23. The average sedimentation rate for lithostratigraphic Units I and II (alternating drift and pelagic sediments, see "Lithostratigraphy") is 35 m/m.y. and remarkably uniform. Small variations in average sedimentation rate are separated by well-defined inflection points. Between three inflection points at ~170, 265, and 400 mcd, respectively, slightly faster and slower average sedimentation rates are recorded (47 m/m.y., between 170 and 265 mcd, and 28 m/m.y., between 260 and 400 mcd). Possible small hiatuses or condensed intervals of less than 100 k.y. are noted at 270 and 340 mcd. A major inflection point at 465 mcd (~13.5 Ma) denotes a major change in sedimentation at Site 1123 and separates the relatively fast sedimentation of lithostratigraphic Units I and II from the much slower sedimentation of Unit III, which is a solely pelagic nannofossil mudstone and chalk. Average sedimentation rates are ~15 m/m.y between this inflection point and the hiatus at 597 mcd and also in the Eocene/Oligocene micritic limestone of Unit IV beneath the hiatus. The debris flow between 552 and 652 mcd is clearly a single depositional event that does not represent much time. It is denoted on Figure F23 as a vertical line. The hiatus between Units III and IV represents almost exactly 12 m.y. of missing time (21-33 m.y.) and is presumably correlative, in part, with the Marshall Paraconformity (Carter, 1985; Fulthorpe et al., 1996). In most onshore localities in New Zealand, the paraconformity separates lower Oligocene and upper Oligocene sediments, and the hiatus at the type locality is 3-4 m.y. (32.4-29.0 Ma) (Fulthorpe et al., 1996). In cases where the paraconformity spans longer intervals, it is generally inferred to be a result of erosion of older strata beneath, which extends the base of the unconformity in places to 36 Ma (Carter, 1985). Here, however, as for other offshore localities (Fulthorpe et al., 1996), the hiatus extends to younger strata, suggesting a prolonged phase of erosion/nondeposition.
Units I and II at Site 1123 contain more than 50 tephra horizons (see "Lithostratigraphy"). The age model presented in Figure F23 affords good constraints on the ages of these horizons.