PRINCIPAL RESULTS

Site 1166 is situated on the Prydz Bay continental shelf on the southwestern flank of Four Ladies Bank, ~40 km southwest of ODP Site 742 (Leg 119) (Fig. F10). Prydz Bay is at the downstream end of a drainage system that originates in the Gamburtsev Mountains of central East Antarctica. The early development and growth of the Cenozoic Antarctic Ice Sheet is believed to have started in middle Eocene to early Oligocene time, but to date, drilling on the continent and the continental margin has not sampled a stratigraphic section that clearly spans and includes the transition period from preglacial to glacial conditions. Site 1166 was chosen to recover core from the Cenozoic sediments below the horizon reached at Site 742. This was intended to provide an age for the arrival of glaciers in Prydz Bay and a record of changes in paleoenvironments and biota with the onset of glaciation.

When the JOIDES Resolution arrived in Prydz Bay, the drill site had to be moved to an alternate site because pack ice covered much of western Prydz Bay where the primary site was located. An additional 3-nmi displacement of the site was required because a tabular iceberg was floating directly over the alternate site. The target sedimentary section (Fig. F11) was thinner and likely a little younger near the base of the section at the final alternate site (i.e., Site 1166) than at the primary site. The drilling at Site 1166 was conducted safely and achieved the desired objective under the sometimes cold and stormy operating conditions.

Drilling at Site 1166 had to be halted temporarily about midway through the drilling because of a large storm with wind gusts exceeding 40 kt and swells >2 m, the maximum limit for shallow-water drilling. During the storm, an iceberg approached within 0.6 nmi and the drill pipe had to be pulled out of the seafloor without a reentry cone in place. The ship was moved away from the site. Following the storm, the ship returned to the site and successfully reentered the hole to continue uninterrupted drilling to a total depth of 381.3 mbsf (Table T1). Recovery at the site was 18.6%; low recovery was due partly to drilling through the upper section (135 mbsf) of diamictites, similar to those sampled at Site 742, and because of sandy fine-grain sediments in the lower part of the hole. Three full logging runs were obtained with excellent data from ~50 mbsf to the bottom of the hole.

The sedimentary section at Site 1166 comprises a diverse suite of strata that from the top down include glacial, early glacial, and preglacial rocks that consist of poorly sorted, sandy, and fine-grained sediments. The ages range from Holocene at the seafloor to Late Cretaceous at the bottom of the hole, with many disconformities throughout.

The sediments are divided into five lithostratigraphic units (Fig. F12). Lithostratigraphic Unit I is of late Pliocene to Holocene age and comprises four subunits. The uppermost subunit (Subunit IA; 0.0-2.74 mbsf) is a biogenic-rich clay interval with pebble-sized clasts. This unit is interpreted as an ice-keel turbate, based in part on iceberg furrows that occur in echo-sounder records in this area (O'Brien and Leitchenkov, 1997).

Below Subunit IA are two intervals of diamicton and one of poorly sorted clayey sandy silt. Subunits IB (2.79-106.36 mbsf) and ID (123.0-135.41 mbsf) are separated by a thin interval (Subunit IC; 113.30-117.22 mbsf) of biogenic-rich clayey silt. Subunit IB is predominantly clayey silt with rock pebbles and clasts that include common fibrous black organic matter. Minor (<2 cm thick) sand and granule beds are present. Subunit IC is interbedded dark gray sandy silt with lonestones and greenish gray diatom-bearing clayey silt with dispersed granules. Biogenic-rich intervals are slightly bioturbated. Contacts with the dark silt are sharp. Subunit ID has interbeds of dark gray clast-poor and clast-rich diamicton. Clast lithologies vary and include gneiss, granite, and diorite rocks. The diamictons suggest subglacial deposition; the sandy silt with IRD is a glaciomarine unit deposited during interglacial periods.

Unit I has lonestones throughout. For Subunits IB and ID, the diamictons suggest subglacial deposition or deposition of a proglacial morainal bank. Sandy silt intervals with shell fragments and microfossils (Subunit IC) suggest current reworking and glaciomarine sedimentation during times of significant glacial retreat. The lonestones signify deposition of IRD. In seismic reflection data, Unit I comprises the topset beds of the seaward-prograding sedimentary section of the outer continental shelf. The contact between lithostratigraphic Units I and II is an abrupt unconformity and was recovered (interval 188-1166A-15R, 8-11 cm [135.63 mbsf]). XRD data above and below the unconformity show shifts in relative abundances of mica, illite, kaolinite, hornblende, and plagioclase. The shifts suggest increased amounts of weathered terrigenous material within Unit I, just above the unconformity.

Unit II (135.63-156.62 mbsf) is an upper Eocene to lower Oligocene diatom-bearing claystone with thin interbedded sands and lonestones. Carbonate contents range from 0.4 to 3.3 wt%. Sands are poorly sorted and bioturbation is moderate. Rare fibrous black organic clasts are present within the sand beds. The bottom of Unit II has rhythmically interbedded centimeter-thick sand and dark claystone. XRD shows gibbsite and kaolinite in Unit II, suggesting erosion of chemically weathered material produced during soil formation. Unit II is a marine sequence that records ice rafting of pebbles during a marine transgression. At Site 742 (Leg 119), 40 km away, similar or younger-age sediments that were interpreted as proximal glacially influenced proglacial or subglacial deposits were sampled. The contact between lithostratigraphic Units II and III is abrupt and was also recovered (interval 188-1166A-17R, 77-78 cm).

Unit III (156.62-267.17 mbsf) consists of massive and deformed sands with a silty clay matrix. The unit contains late Eocene palynomorphs. Calcium carbonate content ranges from 0.4 to 8.4 wt%. The sands have a uniform fabric, are poorly sorted, and lack internal structure. Pebbles of quartzite and rare fibrous black organic fragments are widely dispersed throughout. Two calcite-cemented intervals are present in the sands. The lower part of Unit III is deformed and folded by soft-sediment deformation of sandy beds and includes black organic-rich material with pieces of wood. The coarse-grained sands record deposition on an alluvial plain or delta, and the deformed beds record some reworking of material from underlying organic-rich horizons. A similar-looking sequence of carbonaceous material was drilled in the bottom 2 m at Site 742 and was interpreted as fluvial or possibly lacustrine. At Site 742, however, a sequence comparable to the homogeneous coarse sands was not recovered. The coarse sands of Unit III may record a preglacial alluvial plain or a braided delta of a glacial outwash system. The contact between lithostratigraphic Units III and IV was not sampled.

Unit IV (276.44-314.91 mbsf) comprises black highly carbonaceous clay and fine sandy silt with organic-rich laminae and rare to moderate bioturbation. Palynomorphs suggest a Late Cretaceous (Turonian) age. The sandy silt contains abundant mica and some pyrite. The organic-rich laminae have organic carbon (OC) values as high as 9 wt% and contain common occurrences of authigenic sulfides. Calcium carbonate content ranges from 0.3 to 3.7 wt%. Unit IV records deposition in a restricted marine or lagoonal environment.

Unit V (342.80-342.96 mbsf) consists of a small sample of undated finely laminated gray claystone that was captured in the core catcher within a thick no-recovery zone at the bottom of the hole. From resistivity and velocity log data, the unit may have a relatively large clay content. The claystone is preglacial in origin. It is the same age as Unit IV.

Diatoms, radiolarians, foraminifers, and calcareous nannofossils were examined at Site 1166. Diatoms are present in limited intervals of the core and provide the primary biostratigraphic age control. Three distinct diatom assemblages were noted—Quaternary, Pliocene, and late Eocene-early Oligocene age. Extant Quaternary diatoms occur at 2.12-2.92 mbsf (age = <0.66 Ma). Quaternary to upper Pliocene diatoms occur in diamicts to at least 106.37 mbsf. Two layers of diatomaceous clay (~113.95-114.10 mbsf and ~114.50-115.15 mbsf) are present, containing the upper Pliocene diatom Thalassiosira kolbei (1.8-2.2 Ma) and the Thalassiosira vulnifica to Thalassiosira striata-T. vulnifica Zones (2.2-3.2 Ma), respectively. Radiolarians suggest an age of >2.4 Ma for the lower bed. The boundary between lithostratigraphic Units I and II is a major disconformity (~30 m.y.). Diatoms in Unit II between 135.73 and 153.48 mbsf are of early Oligocene-late Eocene age (~33-37 Ma). Diatoms were not recovered below Unit II, but several specimens of pollen, spores, dinoflagellates, and wood fragments were noted in lower intervals of the hole. Planktonic foraminifers are common above ~90 mbsf, and their ages generally agree with those for diatoms and radiolarians. Calcareous nannofossils are rare. Palynological samples give ages for Units II through V based on correlation with Australian palynozones (O'Brien et al., in press).

Paleomagnetic stratigraphy is difficult because of limited core recovery, but a clear pattern of magnetic polarity intervals is recorded where the recovery is relatively high. A correlation to the geomagnetic polarity time scale (GPTS) is in progress using key biostratigraphic datums (Fig. F13). Downhole variations in the concentration-dependent and magnetic mineralogy-dependent parameters show that the main lithostratigraphic units have alternating high and low magnetic mineral concentrations and distinct magnetic signatures. Iron sulfide minerals are present below ~140 mbsf.

Interstitial water profiles document downhole sediment diagenesis and diffusional exchange with bottom seawater. From 0 to 150 mbsf, the oxidation of organic matter reduces sulfate values from 28 to 8 mM and ammonium increases from 177 to 1277 mM. From 0 to 75 mbsf, alkalinity decreases from 4.5 to 1 mM, silica decreases from 800 to 200 mM, potassium decreases from 12 to 2 mM, and calcium increases from 10 to 22 mM. The profiles suggest diagenetic silicate reactions are occurring within the sulfate reduction zone. Between 150 and 300 mbsf, calcium and magnesium show minor changes in relative concentration (15 and 24 mM, respectively) suggesting diffusional processes are dominant.

OC contents of the sediments based on 14 samples (selected by dark color) vary according to lithostratigraphic unit. The diamictite (Subunit IB) has OC values of 0.4-1.4 wt%; the massive sand (Unit III) has OC values of 0.2-0.5 wt%, except for one bed near the base of the fluvial/deltaic sand section of Unit III that contains 9.2 wt% OC; and the carbonaceous claystone (Unit IV) has 1.5-5.2 wt% OC. Inorganic carbon was low (<0.1 wt%) throughout most of the recovered section. Gas analyses showed only background levels of methane (4-10 ppmv), and no other hydrocarbons were detected. Most samples are enriched in carbon relative to nitrogen, which suggests the input of land-plant organic matter, especially for samples with >1 wt% OC. Rock-Eval pyrolysis analysis shows that the pyrolyzable fraction of the OC is low (hydrogen index values of 50 mg of hydrocarbon per gram of carbon or less), consistent with degraded plant material as the source of the carbon in the more carbonaceous (>2 wt%) samples. Samples with lower carbon contents (<1.4 wt%) may contain a recycled higher thermal maturity component. This recycled organic component is suggested by Rock-Eval T_max values that approach 490°C as OC decreases toward values of 0.5 wt%. The diamictites (Unit I) have a greater proportion of recycled organic matter than the carbonaceous units (base of Units III and IV), which contain mostly first-cycle organic matter.

The majority of the sedimentary section has porosities between 20% and 40%, with the exception of Unit II, where the average porosity is 50%. P-wave velocities change abruptly at most lithostratigraphic boundaries. The measured shear strengths show that the sediments, especially in Unit I, are overconsolidated, with an overconsolidation ratio of ~2 below 70 mbsf. The overconsolidation record implies at least one or two periods when sediments were either compacted by a 250- to 300-m-thick sediment column, now eroded away, or were loaded by 330- to 420-m-thick nonbuoyant ice during prior glaciations.

Wireline logging was carried out in Hole 1166A with excellent results (Fig. F14). Three runs were made using the triple combination (triple combo), sonic/geological high-sensitivity magnetic tool (GHMT), and Formation MicroScanner (FMS) tool from 33 mbsf to the bottom of the hole at 385 mbsf. Six logging units are recognized, and each stratigraphic unit has a distinctive signature and appearance, especially in the resistivity, sonic velocity, and FMS data. The deeper parts of the hole (logging Units 4b, 5a, 5b, and 6, equivalent to lithostratigraphic Units III [lower part], IV, and V) consist of preglacial to early glacial sediments that have large gamma-ray fluctuations indicative of heavy mineral K, Th, and U contents associated in part with the high OC values here. These units also have lower velocity, density, and resistivity than the thick overlying deltaic sands of lithostratigraphic Unit III. All log traces show abrupt shifts at the logging Units 3/4 and 2/3 boundaries (equivalent to lithostratigraphic Units II/III and I/II boundaries) and are recognized as unconformities. The diamictons with interbedded glaciomarine clays and silts of lithostratigraphic Unit I have generally high and variable magnetic susceptibilities that suggest high variability in magnetite concentrations. FMS images clearly show the variability in the lithostratigraphic units, the presence of lonestones, and the deformation of the organic-rich silt-sand horizons (lithostratigraphic Unit III) (Fig. F15). The resistivity and velocity logs, along with seismic reflection profiles, suggest the alluvial sands (Unit III) to glaciomarine diatom-bearing claystones (Unit II) transition represents a marine transgression.

Logging while drilling (LWD) was done in Hole 1166B to test the Power Pulse and compensated dual resistivity tools and record spectral gamma-ray and resistivity data in the uppermost 42 m of the sediment. This interval could not covered by wireline logging because of the pipe position. Resistivity values increase in linear segments from near zero at the seafloor to the 3.5-m values measured by the wireline logs at the base of the pipe.

A primary objective of Leg 188 was achieved when a set of cores that record intervals in the history of Antarctic paleoenvironments for the Prydz Bay region extending back through the early stage of glaciation to preglacial times was recovered at Site 1166 (Fig. F16). Drilling during Leg 119 in Prydz Bay recovered a record of early proximal glaciation at Sites 739 and 742 but did not capture the transition to warmer climates as would be indicated by the presence of local vegetation. Correlation of Site 1166 to Site 742 (~40 km apart) by comparison of downhole logs and regional seismic stratigraphy shows that Units I and II at Site 1166 are equivalent to (or older than) similar units at Site 742. Below the level of Unit II, however, Site 1166 samples are stratigraphically lower and record a more temperate alluvial facies than seen at Site 742. The lower part of Unit III (i.e., the deformed organic-rich sands and silts) may have been sampled in the lowermost 2 m of core at Site 742, but confirmation of this awaits further comparison of the two drill sites. If the organic units are the same, then a thick section of sands (Unit III) is missing at Site 742. The deepest unit at Site 1166 (Unit V) lies below a regional seismic unconformity that can be traced to Site 741, ~110 km away, where gray claystones similar to those of Unit V were also sampled. The age of the claystone at Site 741 is Early Cretaceous, which is preglacial.

The paleoenvironmental record inferred from the cores at Site 1166 shows a systematic uphole change from preglacial warm to full-glacial cold climates, such as that envisioned for the Prydz Bay region in Figure F16. The rich carbonaceous strata (Cretaceous Unit IV), which overlie Unit V, record a time of more temperate climatic conditions when vegetation existed on Antarctica. The sands (late Eocene Unit III) of the alluvial plain environment may represent the transition into the progressively colder climates that are recorded in the proglacial (late Eocene to early Oligocene Unit II), glacial marine (Unit II and late Pliocene and younger Unit I), and subglacial (Unit I) sediments.

Site 1167 is located in the middle of the Prydz Channel Trough Mouth Fan (Figs. F10, F17). Construction of the fan started in late Miocene to mid-Pliocene time when the Lambert Glacier formed a fast-flowing ice stream on the western side of Prydz Bay. The fan has grown most during episodes when the Lambert Glacier grounded at the shelf edge, delivering basal debris to the fan apex. This material was then redistributed by sediment gravity flows and meltwater plumes. Models of trough mouth fan sedimentation suggest that thick siliciclastic units should correspond to peaks in Antarctic ice volume, whereas periods of reduced ice volume should be represented by hemipelagic sediments. Thus, the alternation of facies should reflect the number of times the East Antarctic Ice Sheet has expanded to the shelf edge in latest Neogene time.

Hole 1167A was cored with the advanced hydraulic piston corer (APC) system to refusal at 39.7 mbsf. Coring then proceeded with the extended core barrel (XCB) system to a total depth of 447.5 mbsf (Table T1). Planned drilling time at the site was shortened by 42 hr because of icebergs and a ship schedule change, and the target depth of 620 mbsf (base of the Prydz Trough Mouth Fan) was not achieved. Four icebergs approached to within 0.1 nmi of the drill site, causing a total of 27 hr delay.

The sedimentary section at Site 1167 comprises a 447.5-m-thick sequence of clayey silty sands with dispersed rock clasts with minor beds of coarse sands, clays, and sandy clays. Two lithostratigraphic units are identified (Fig. F18).

Unit I (0-5.17 mbsf) is composed of olive and reddish brown clay and sandy clay with minor admixtures of biogenic components (e.g., as much as 2% diatoms and 1% sponge spicules). There are isolated beds of fine sand and rare lonestones. Diffuse reddish brown color bands are present in several thin intervals. The transition to Unit II is gradational. Unit I records a period of hemipelagic deposition when fine particles, biogenic material, and IRD settled out of the water column.

Unit II (5.17-447.5 mbsf) makes up the majority of the section at Site 1167 and is composed of one major facies (II-1) and three minor facies. Facies II-1 is composed of interbedded, poorly sorted dark gray sandy silt, silty sand, clayey sand, and clast-poor diamicton. Numerous color alternations of dark gray and dark reddish gray with sharp contacts occur between 64 and 98 mbsf. Some decimeter- to meter-scale successions of clast-poor diamicton and gravel beds are noted. Lonestones are common, with variable lithologies including granite, granite gneiss, garnet-bearing gneiss, metaquartzite, and sandstone. Dolerite, schist, conglomerate, and rare carbonized wood are also present. Sandstone and granite components vary systematically in the hole, with sandstone lonestones common below 200 mbsf and granite lonestones common above 200 mbsf. Facies II-2 is composed of gray, moderately sorted coarse sand. Grains are subrounded and predominately quartz, K-feldspar, and mafic minerals. The first occurrence of Facies II-2 downcore is at 179 mbsf. Facies II-3 is composed of dark gray clay with silt laminations, rare sand grains, and no lonestones. Sharp contacts mark the top and base of this facies. Some silt laminae converge and indicate cross-bedding. Facies II-4 is composed of green gray clay with dispersed clasts, abundant foraminifers, and few nannofossils. The upper contact is sharp, and the lower contact is gradational to sharp.

Unit II (Facies II-1 and II-2) records deposition by mass transport, probably massive debris flows, as evidenced by poor sorting, abundant floating clasts, little visible grading, and a lack of biogenic components. The debris flows most likely represent deposition during glacial periods when ice extended to the shelf break and could deliver large volumes of sediment to the upper continental slope. Individual flows cannot be identified visually. The thin intervals of fine-grained sediment (Facies II-3 and II-4) are similar in appearance and composition to muddy contourites observed at Site 1165, and hence, may denote times when contour currents were active on the fan. The silt laminae and bioturbation in Facies II-3 are not consistent with turbidite deposition. Facies II-4 may record short intervals, possibly interglacials, when pelagic deposition dominated.

Sixteen lithologic varieties of lonestones were cataloged, and they generally vary randomly in size, with only a small size increase downhole to 200 mbsf. Below ~160 mbsf, the number of lonestones per meter remains fairly consistent, except for three intervals (160-210, 300-320, and 410-420 mbsf) where there are downward increases. Systematic variations in concentration of sandstone and granite clasts (noted above) suggest that two different source areas may have delivered material to Site 1167 at different times (Fig. F19).

XRD analyses show that the total clay mineral content is relatively constant throughout the hole. XRD analyses of clay types give mixed results for Units I and II, with smectite more common in Unit I and at depths below 382 mbsf than elsewhere and illite found in all samples. Further detailed analyses are likely to clarify whether changes in illite-smectite ratios relate to times of glacial advances.

Chronostratigraphy at Site 1167 is poorly controlled because of the unexpected paucity of siliceous microfossils; however, dates in Unit I are younger than 0.66 Ma, and a sample from ~215 mbsf seems to be of early or middle Pleistocene age. Foraminifers are present consistently throughout the section and include pelagic foraminifer shelf faunas in diamictons and in situ midbathyal faunas in a few samples. Changes in foraminifer faunas closely match changes detected in various lithological parameters. Age control at this time is not adequate to determine average sedimentation rates.

Magnetostratigraphic analyses identified the Matuyama/Brunhes boundary between 30 and 34 mbsf (Fig. F20). The magnetic polarity below 34 mbsf remains mainly reversed and possibly includes the Jaramillo and Olduvai Subchrons. The concentration-dependent magnetic parameters (susceptibility and anhysteretic and isothermal remanent magnetization) indicate that magnetite concentrations have large-scale cyclic (tens to hundreds of meters) variations, which are not commonly seen (Fig. F20). The values increase abruptly uphole at ~208 mbsf, between 113 and 151.2 mbsf, and between 55 and 78.5 mbsf followed by a nearly linear uphole decrease. Superimposed on the large-scale cycles are small-scale variations. The anhysteretic over isothermal remanent magnetization ratio indicates that the magnetic grain size changes uphole from finer to coarser above 217 mbsf. The origin of the large-scale cycles is not yet understood, but it is likely related to systematic changes in sediment provenance caused by changes in the volume of ice from different sources and the location of areas of maximum erosion during glacial periods.

Interstitial water profiles document downhole sediment diagenesis, mixing of chemically distinct subsurface interstitial waters, and diffusional exchange with modern bottom seawater. From 0 to 20 mbsf, chlorinity and sulfate increase by ~3% over seafloor values, suggesting that high-salinity LGM seawater is preserved (Fig. F21). Sulfate decreases downhole from the seafloor (30 mM) to 433 mbsf (24 mM) in a stepped profile, raising the possibility that a number of "fossil" sulfate reduction zones may also be preserved (Fig. F21). Dissolved manganese increases downhole from 15 to 20 mM between the seafloor and ~25 mbsf. Alkalinity decreases downhole from 3 to 1.3 mM between the seafloor and 40 mbsf before steadily increasing to 2 mM at 433 mbsf. From 0 to ~60 mbsf, dissolved downhole profiles of calcium (10-25 mM), magnesium (56-42 mM), potassium (12-2 mM), and lithium (30-5 mM) all suggest diagenetic silicate-clay reactions are occurring. Below 5 mbsf, dissolved silica concentrations are enriched slightly over modern bottom waters (~300 vs. ~220 mM), reflecting the absence of biogenic opal within the sediments. Calcium carbonate is a minor component in the matrix sediments throughout the hole and is slightly more abundant in lithostratigraphic Unit II than in Unit I.

The concentration of hydrocarbon gases was at background levels (4-10 ppmv) for methane, and ethane was present above detection limits only in a few cores deeper than 350 mbsf. The OC content averages ~0.4 wt% with no apparent trend with depth. Organic matter characterization by Rock-Eval pyrolysis indicates that all samples contain predominantly recycled and degraded thermally mature organic matter.

Sediment water content and void ratio decrease sharply with depth in lithostratigraphic Unit I, reflecting normal compaction. Within Unit II, these properties were relatively uniform, except for a downhole decrease at 210 mbsf, where grain density and magnetic susceptibility values also decrease abruptly. P-wave velocities increase at this depth. Undrained shear strength values increase uniformly throughout the hole at a lower than typical rate, possibly because of the clay mineralogy combined with the high proportions of silt and sand within the sediment. There is no evidence of sediment overcompaction.

Wireline logging operations in Hole 1167A were attempted with the triple combo tool string. The tool string was lowered to 151 mbsf, where an obstruction halted further progress. A conglomerate interval was noted in the cores at this depth. Log data were collected from this depth to the base of pipe at 86.9 mbsf, covering an interval of 66 m. Time constraints, poor hole conditions, and problems encountered with the lockable flapper valve resulted in a decision to switch to LWD in a new hole. Excellent spectral gamma-ray and resistivity data were recorded to 261.8 mbsf before time ran out. Resistivity data show several clay and gravel-rich beds, with high gamma-ray values for a red bed interval at 60-90 mbsf, and low values between 90 and 120 mbsf and 215 and 255 mbsf. The change to low values may be due to a reduced concentration of granitic clasts or a change from a clay-rich to a sandier matrix.

Site 1167 is the first drill site where the sedimentary fans that are common on the upper continental slope around Antarctica, seaward of glacially carved sections of the continental shelf, were directly sampled. The site reveals previously unknown large-scale (20- to >200-m-thick) cycles in magnetic susceptibility and other properties that are not yet fully explained but are likely due to systematic changes in the Lambert Glacier ice-drainage basin during Pleistocene and late Pliocene(?) time. Within the large cycles are likely many separate debris flows and interbedded hemipelagic muds that indicate times of individual advances and retreats of the ice front to, or near, the continental shelf edge. The debris flows are well represented in the cores, but the mud intervals are sparse and may either have not been recovered or have been removed by younger flows. Because there are few age control points, it is not yet possible to determine sedimentation rates at Site 1167. If the rates are high, as we suspect from the few available age dates, then almost all sediment during the latest Neogene glacial intervals sampled at Site 1167 were deposited as debris flows on the trough mouth fan and are not reaching Wild Drift (Site 1165), where sediment rates are low. Alternatively, some of the fine component of the latest Neogene glacial sediment is being carried away by deep ocean currents.

Site 1165 is situated on the continental rise offshore from Prydz Bay over mixed pelagic and hemipelagic sediments of the central Wild Drift (Figs. F10, F22A, F22B). The drift is an elongate sediment body formed by the interaction of sediment supplied from the shelf and westward-flowing currents on the continental rise. The site is in 3537 m of water and was selected to provide a record of sedimentation that extends back to the onset of contour current-influenced deposition on the rise. The main objective was to obtain a proximal continental-rise record of Antarctic glacial and interglacial periods for comparison with other sites around Antarctica and with those of Northern Hemisphere ice sheets.

Prior to drilling, a single seismic-reflection profile was recorded across the location of Site 1165, using the ship's water-gun, to verify the location of the site. Three holes were drilled at the site (Table T1). Hole 1165A consisted of a mudline core that was dedicated to high-resolution interstitial water sampling. Hole 1165B was cored with the APC to 147.9 mbsf (86.4% recovery) and deepened with the XCB to 682.2 mbsf (57.3% recovery). Hole 1165C was washed down to a depth of 54 mbsf, where a single core was taken at an interval that had been missed in Hole 1165B. Continuous RCB coring began at 673 mbsf and continued to a total depth of 999.1 mbsf, with 80% recovery. Coring operations were interrupted by five icebergs that came to within 200 m of the drill site and caused a total loss of 98 hr. Hole 1165C was successfully logged with the triple combo tool string from 176 to 991 mbsf and with the sonic tool from 176 to 580 mbsf.

Drilling at Site 1165 yielded a relatively continuous 999-m-thick sedimentary section of early Miocene- to Pleistocene-age terrigenous and hemipelagic deposits (Fig. F23) with only few minor (<2 m.y.) disconformities. Dispersed clasts (IRD) are present down to the bottom of the hole, but lonestones are infrequent below 500 mbsf (lower Miocene). Both dispersed clasts and lonestones are relatively abundant above 300 mbsf (middle Miocene and younger).

The sedimentary section is divided into three lithostratigraphic units that are characterized by cyclic variations between biogenic-bearing (lighter) and terrigenous-dominated (darker) intervals. The cyclic variations in lithology are also recorded as visual color alternations and cycles in spectrophotometer lightness factor, bulk density, magnetic susceptibility, and other laboratory and downhole log parameters to varying degrees. In general, cores get darker downhole as biogenic intervals become thinner relative to the thickness of terrigenous-bearing intervals. For the same reason, light-dark cyclicity is more prominent above ~400 mbsf. Darker units generally have higher bulk density, magnetic susceptibility, and OC values. The sediment consists mostly of quartz, calcite, plagioclase, K-feldspar, and a mixture of clay minerals, as well as minor hornblende and pyrite. Silt-sized components are mainly quartz, but plagioclase, biotite, amphibole, and other heavy minerals are common.

Unit I (0-63.8 mbsf) consists of structureless brown clay and diatom-bearing clay. There are beds with minor diatom-bearing greenish gray clay that have dispersed sand grains, granules, and lonestones. There is minor laminated silt and minor brown foraminifer-bearing clay. Foraminifers comprise 5%-15% of the sediment above 13 mbsf. One interval within Unit I (20-30 mbsf) is characterized by alternations between two facies like those of Unit II (i.e., Facies II-1 and II-2).

Unit II (63.8-307.8 mbsf) is characterized by alternations of two main facies that differ in color and composition. Facies II-1 consists of structureless, homogeneous greenish gray diatom clay, and Facies II-2 is mostly dark gray diatom-bearing clay with some intervals of scattered silt laminae. Many lower boundaries of Facies II-2 are sharp, and upper boundaries are transitional with bioturbation that increases upward into Facies II-1. Higher amounts of siliceous microfossils and IRD (floating sand grains and pebbles) are found in Facies II-1 than II-2. Facies II-1 is characterized by lower grain density because of the higher diatom content. A third facies (II-3) occurs rarely and consists of several 15- to 40-cm-thick nannofossil chalk beds that have a sharp base and pass gradually up into Facies II-1. Within Unit II, three subunits are identified based on the different proportions of Facies II-1 and II-2. Subunit boundaries are at 160 and 252 mbsf and partially denote amounts of IRD, with greater amounts of IRD in Subunits IIA and IIC than in Subunit IIB.

Unit III (307.8-999.1 mbsf) comprises a section of thinly bedded planar-laminated claystone that is divided, like Unit II, into two main facies that differ in color, composition, and bedding characteristics. Facies III-1 consists of greenish gray bioturbated structureless clay and claystone and diatom-bearing clay and claystone with dispersed coarse sand grains and rare granule to pebble-sized lonestones. Facies III-2 is composed of dark gray thinly bedded planar-laminated clay and claystone with abundant silt laminae. Lonestone (dolerite, diorite gneiss, and mudstone) abundance in Unit III is low and decreases downhole.

In Facies III-1, the thickness of the greenish gray intervals is generally <1 m and some intervals have higher concentrations of silt- and sand-sized material. The upper contacts are commonly sharp, laminae are rare, and bioturbation increases upcore. Angular mud clasts and benthic foraminifers are present in this facies below 800 mbsf. Siliceous microfossil content is low.

Facies III-2 becomes increasingly fissile with depth and changes to very dark gray or black below 894 mbsf. Bioturbation is rare in Facies III-2, and light-color silt laminae are a conspicuous feature (average = 150-200 laminae/m). Many cross-laminated silt ripples are present, and ripples are more common below 673 mbsf than above. Microfossils are rare and seem to disappear completely below ~600 mbsf. Below 842 mbsf, laminae with calcite cement are present and sections of the core become increasingly cemented with what are most likely authigenic carbonates. A change to darker-color claystones occurs at 894 mbsf, where fracture patterns also become curved.

In both Facies III-1 and III-2, individual 0.5-cm-sized horizontal Zoophycos burrows are evident along with clusters of unidentified millimeter-sized burrows.

An excellent record of siliceous microfossils is found at Site 1165 down to 600 mbsf, where biogenic opal disappears because of an opal-A/opal-CT diagenetic transition. Neogene high-latitude zonal schemes for both diatoms and radiolarians yielded identification of 21 diatom and 12 radiolarian biostratigraphic datums between 0 and 600 mbsf. Below 600 mbsf, age assignments are inferred from calcareous nannofossils, which are present in only a few discrete intervals with moderate to good preservation of low-diversity assemblages. Nannofossils yield Pleistocene to earliest Miocene ages. Benthic foraminifers are more common than planktonics, which are rare. The foraminifers indicate several intervals of redeposited material.

A magnetostratigraphy was determined for Site 1165 for the interval 0-94 mbsf and below 362 mbsf (Fig. F24). The magnetostratigraphic record and biostratigraphic ages, when combined, yield an age vs. depth model that shows relatively rapid deposition in early Miocene time (~120 m/m.y.), somewhat slower deposition in middle to late Miocene time (~50 m/m.y.), and even slower deposition since late Miocene time (~15 m/m.y.) (Fig. F25). From magnetostratigraphy, the bottom of Hole 1165C (999.1 mbsf) has an age of ~21.8 Ma. The uncertainty in ages is larger below 600 mbsf, where only few biostratigraphic ages exist to constrain the paleomagnetic reversal stratigraphy.

Measurements of rock magnetic properties in the interval 114-370 mbsf indicate that magnetic mineral concentrations (i.e., concentrations of magnetite) drop significantly, with nearly a complete loss of magnetic intensity. Shipboard analysis suggests that this unusual loss of magnetic signal is caused by diagenetic dissolution of magnetite in the presence of silica-rich pore-waters. A zone of low grain density is observed in cores from part of this zone ~140 mbsf and may also coincide with the disappearance of magnetite.

The interstitial water profiles document increasing downhole sediment diagenesis at Site 1165. Sulfate values decrease linearly from 30 to 2 mM in the interval from 0 to 150 mbsf, but increases occur in ammonium (20-384 µM), phosphate (3-10 µM), and alkalinity (3-8 mM) throughout the same interval because of destruction of organic matter. Ammonium increases (to 800 µM), phosphate decreases (to 0), and alkalinity decreases linearly (to 1 mM at 999 mbsf) throughout the interval from 150 to 400 mbsf. Concentrations of dissolved silica increase from 522 µM at the seafloor to a maximum of 1000 µM at 200 mbsf and reflect the dissolution of abundant siliceous microfossils. Silica values decrease from 400 to 999 mbsf. The theoretical opal-A/opal-CT transition is at ~600 mbsf, based on measured downhole temperatures. A strong seismic reflection is observed at this depth. Below 150 mbsf, calcium and magnesium values correlate inversely, which suggests diagenetic control and an unidentified Ca-rich lithology below the drilled section. A decrease in potassium values below 50 mbsf and an increase in fine-grained K-feldspar, identified by XRD measurements, suggests an authigenic origin for the K-feldspars.

Hydrocarbon gas contents of cores are low (<400 ppmv) within the sulfate reduction zone down to ~150 mbsf (Fig. F26). Below 150 mbsf, methane increases rapidly and reaches values of 20,000-40,000 ppmv between 270 and 700 mbsf, and 40,000-100,000 ppmv from 700 to 970 mbsf. These headspace gas measurements indicate only the residual gas in the pore water of cores after outgassing upon retrieval to the surface. The gas concentrations are equivalent to ~8-35 mM dissolved methane when adjusted for sample size variation, density, and porosity. Cores deeper than 700 mbsf may contain more gas because of increased lithification and retarded outgassing. Ethane is present in headspace gas samples deeper than 157 mbsf, and the C₁/C₂ value shows the expected decrease with depth for the observed geothermal gradient at this site (secant temperature gradient of 43.6°C/km). OC contents of Site 1165 sediments are low (0.1-0.8 wt%), except for high OC (1.8-2 and 1.3 wt%) beds at 122 and 828 mbsf (Fig. F26). Cores from within the gas hydrate stability zone (seafloor to 460 mbsf) were examined immediately upon recovery, but no hydrates were observed.

Acoustic velocities and shear-strength measurements increase as a result of normal compaction (in the upper 500 mbsf) and diagenesis (with a greater effect below 600 mbsf). Horizontal downhole P-wave velocities increase abruptly at 114 mbsf from ~1535 to 1575 m/s and then linearly from 114 to 603 mbsf. Other significant downhole P-wave velocity increases occur at 692 mbsf, at 800 mbsf (to 2660 m/s), and 960 mbsf (to 2840 m/s). These abrupt increases are observed as reflections on seismic data.

Hole 1165C was logged with a single pass of the triple combo tool from 176 to 994 mbsf and a single pass of the sonic tool from 176 to 580 mbsf. Total gamma-ray and bulk-density log traces covary, suggesting that lower-density diatom-bearing strata are present throughout the hole. The boundary between lithostratigraphic Units II and III at 305 mbsf is marked by a large fluctuation in total gamma-ray values, which suggests a shift in clay content at the boundary. The distinct silica- and calcite-cemented intervals observed in the cores are marked by inflections in most log traces.

Site 1165 provides paleontologic and sedimentologic evidence that numerous alternations or cycles between biogenic material and clay-sized terrigenous debris from Antarctica have occurred since earliest Miocene time (Fig. F27). The cyclicity observed in the lightness values for two intervals (83-100 mbsf and 107-123 mbsf; in lithostratigraphic Unit II) in the cores was analyzed for spectral content to evaluate the potential effect of Milankovitch periodicities on biogenic/terrigenous sedimentation cycles at this site (Fig. F27). For the shallower interval, significant spectral peaks are found at periods of 3.28, 1.45, 1.08, and 0.64 m thickness. A sedimentation rate of 3.5 cm/k.y. is reasonably well constrained for the interval and gives periods of 93.7, 41.5, 20.8, and 18.2 k.y., respectively. These periods are similar to the Milankovitch cycles of 100 k.y. (eccentricity), 41 k.y. (obliquity), 23 k.y., and 19 k.y. (precession), suggesting an orbitally forced origin for the light-dark cyclicity. For the deeper interval, significant spectral peaks are at 4.27, 1.55, 0.95, and 0.72 m, which have similar peak periodicity ratios to the upper interval, again suggesting an orbital origin for the deeper cycles. In the deeper interval, the sedimentation rate is less well constrained. But, by using the peak-to-peak ratios as a guide, an inferred sedimentation rate of 3.8 to 4.1 cm/k.y. would give the same periods as the shallow interval (i.e., 93.7, 41.5, 20.8, and 18.2 k.y.). The observed lightness changes correlate also to variations in bulk density and magnetic susceptibility, indicating that the lightness (color) data likely document orbitally driven changes in the Site 1165 depositional environment since at least early Miocene time. A similar approach has been used recently to detect orbital signals in upper Oligocene to lower Miocene sedimentary sequences drilled in the Ross Sea, Antarctica, by the Cape Roberts Project (Claps et al., in press). Milankovitch cyclicities have been reported for late Miocene-age and younger diatom-bearing hemipelagic and terrigenous sediments at ODP Site 1095 from a drift deposit, like the Wild Drift, adjacent to the Antarctic Peninsula (Shipboard Scientific Party, 1999).

In addition to the cyclic variations, significant uphole changes occurred at Site 1165 from early to late Miocene times (Units III and II), and include an eightfold decrease in average sedimentation rates from 12 to 1.5 cm/k.y., an increase in the total clay content, an increase in amount of sand-sized and lonestone IRD, and other changes (e.g., first appearance of glauconite) (Figs. F22, F25). The changes likely reflect shifts in the onshore Prydz Bay region (sediment source area) to paleoenvironments that produced less terrigenous sedimentation, a lower-energy current regime, and more floating ice at Site 1165 starting in middle Miocene time. The uphole shift to increased clay (~305 mbsf) and first appearance of glauconite (~220 mbsf) heralds (1) a mid-Miocene change to erosion of sedimentary basins on the shelf and (2) subsequent inferred overdeepening of the shelf to the large water depths of today. Such deep-cut erosion would be by grounded glaciers crossing the continental shelf and dispersing icebergs with entrained sediment. The times of lowest sediment supply, in the latest Neogene (i.e., above 60 mbsf), have upward-decreasing silica contents (i.e., fewer diatoms and sponge spicules) and varied (cyclic?) IRD concentrations, which may reflect increasing extent of sea ice cover and ice sheet fluctuations. During latest Neogene time, a widely recognized period of intense Antarctic (and Arctic) glacier fluctuations, thick debris flows blanketed the adjacent continental slope (Site 1167), but little sediment was deposited on Wild Drift (Site 1165).