Physical properties and sedimentology on core GeoB1510-1

Detailed information about the sediment properties and microstructure can be provided through the analysis of digital ultrasonic P wave seismograms recorded automatically during full waveform core logging. The physical parameter which predominantly affects the elastic wave propagation in water-saturated sediments is the P wave attenuation coefficient. The related sedimentological parameter is the grain size distribution. A set of high-resolution ultrasonic transmission seismograms (ca. 50-500 kHz), which indicate downcore variations in the grain size by their signal shape and frequency content, are presented. Layers of coarse-grained foraminiferal ooze can be identified by highly attenuated P waves, whereas almost unattenuated waves are recorded in fine-grained areas of nannofossil ooze. Color-encoded pixel graphics of the seismograms and instantaneous frequencies present full waveform images of the lithology and attenuation. A modified spectral difference method is introduced to determine the attenuation coefficient and its power law a = kfn. Applied to synthetic seismograms derived using a "constant Q" model, even low attenuation coefficients can be quantified. A downcore analysis gives an attenuation log which ranges from ca. 700 dB/m at 400 kHz and a power of n = 1-2 in coarse-grained sands to few decibels per meter and n ? 0.5 in fine-grained clays. A least squares fit of a second degree polynomial describes the mutual relationship between the mean grain size and the attenuation coefficient. When it is used to predict the mean grain size, an almost perfect coincidence with the values derived from sedimentological measurements is achieved.

Geochemistry of Pleistocene volcanic rocks in the Mariana Forearc

Geophysical surveys of the Mariana forearc, in an area equidistant from the Mariana Trench and the active Mariana Island Arc, revealed a 40-m-deep graben about 13 km northwest of Conical Seamount, a serpentine mud volcano. The graben and its bounding horst blocks are part of a fault zone that strikes northwest-southeast beneath Conical Seamount. One horst block was drilled during Leg 125 of the Ocean Drilling Program (Site 781). Three lithologic units were recovered at Site 781: an upper sedimentary unit, a middle basalt unit, and a lower sedimentary unit. The upper unit, between 0 and 72 mbsf, consists of upper Pliocene to Holocene diatomaceous and radiolarian-bearing silty clay that grades down into vitric silty clay and vitric clayey silt. The middle unit is a Pleistocene vesicular, porphyritic basalt, the top of which corresponds to a high-amplitude reflection on the reflection profiles. The lower unit is a middle to upper (and possibly some lower) Pliocene vitric silty clay and vitric clayey silt similar to the lower part of the upper unit. The thickness of the basalt unit can only be estimated to be between 13 and 25 m because of poor core recovery (28% to 55%). The absence of internal flow structures and the presence of an upper glassy chilled zone and a lower, fine-grained margin suggest that the basalt unit is either a single lava flow or a near-surface sill. The basalt consists of plagioclase phenocrysts with subordinate augite and olivine phenocrysts and of plagioclase-augite-olivine glomerocrysts in a groundmass of plagioclase, augite, olivine, and glass. The basalt is an island arc tholeiite enriched in large-ion-lithophile elements relative to high-field-strength elements, similar to the submarine lavas of the southern arc seamounts. In contrast, volcanic rocks from the active volcanoes on Pagan and Agrigan islands, 100 km to the west of the drill site, are calc-alkaline. The basalt layer, the youngest in-situ igneous layer reported from the Izu-Bonin and Mariana forearcs, is enigmatic because of its location more than 100 km from the active volcanic arc. The sediment layers above and below the basalt unit are late Pliocene in age (about 2.5 Ma) and normally magnetized. The basalt has schlierenlike structures, reverse magnetization, and a K-Ar age of 1.68±0.37 Ma. Thus, the basalt layer is probably a sill fed by magma intruded along a fault zone bounding the horst and graben in the forearc. The geochemistry of the basalt is consistent with a magma source similar to that of the active island arc and from a mantle source above the subducting Pacific plate.

Sediments in Arctic sea ice

Despite the Arctic sea ice cover's recognized sensitivity to environmental change, the role of sediment inclusions in lowering ice albedo and affecting ice ablation is poorly understood. Sea ice sediment inclusions were studied in the central Arctic Ocean during the Arctic 91 expedition and in the Laptev Sea (East Siberian Arctic Region Expedition 1992). Results from these investigations are here combined with previous studies performed in major areas of ice ablation and the southern central Arctic Ocean. This study documents the regional distribution and composition of particle-laden ice, investigates and evaluates processes by which sediment is incorporated into the ice cover, and identifies transport paths and probable depositional centers for the released sediment. In April 1992, sea ice in the Laptev Sea was relatively clean. The sediment occasionally observed was distributed diffusely over the entire ice column, forming turbid ice. Observations indicate that frazil and anchor ice formation occurring in a large coastal polynya provide a main mechanism for sediment entrainment. In the central Arctic Ocean sediments are concentrated in layers within or at the surface of ice floes due to melting and refreezing processes. The surface sediment accumulation in central Arctic multi-year sea ice exceeds by far the amounts observed in first-year ice from the Laptev Sea in April 1992. Sea ice sediments are generally fine grained, although coarse sediments and stones up to 5 cm in diameter are observed. Component analysis indicates that quartz and clay minerals are the main terrigenous sediment particles. The biogenous components, namely shells of pelecypods and benthic foraminiferal tests, point to a shallow, benthic, marine source area. Apparently, sediment inclusions were resuspended from shelf areas before and incorporated into the sea ice by suspension freezing. Clay mineralogy of ice-rafted sediments provides information on potential source areas. A smectite maximum in sea ice sediment samples repeatedly occurred between 81°N and 83°N along the Arctic 91 transect, indicating a rather stable and narrow smectite rich ice drift stream of the Transpolar Drift. The smectite concentrations are comparable to those found in both Laptev Sea shelf sediments and anchor ice sediments, pointing to this sea as a potential source area for sea ice sediments. In the central Arctic Ocean sea ice clay mineralogy is significantly different from deep-sea clay mineral distribution patterns. The contribution of sea ice sediments to the deep sea is apparently diluted by sedimentary material provided by other transport mechanisms.

Petrographic description and mineral composition of ODP Site 161-975 sediments

Ocean Drilling Program Site 975 is located near the base of the Menorca Rise in the South Balearic Basin of the western Mediterranean Sea. Coring at this site penetrated the Pliocene/Miocene boundary and recovered a sequence of sediments that represent the final stages of salt deposition and the transition from evaporitic to open marine conditions at the end of the Miocene (Messinian). Detailed petrographic observations and bulk mineralogical analyses by X-ray diffraction form the basis for preliminary interpretations of depositional environments for this section. Gypsum is thought to have been deposited in an evaporating basin below wave base. Cycles consisting of a clay layer overlain by gypsiferous chalk, laminated gypsum, and finally pinch-and-swell gypsum suggest upsection increases in salinity. The gypsum section is overlain by two exotic sand layers thought to mark events of fresher water (marine or meteoric) inflow to the basin. Gypsum deposition terminated and was replaced by inorganic precipitation of micritic calcite with periodic, variable dilution by fine-grained terrigenous sediment. The micritic sediments have fine, slightly wavy, laminations indicating either an algal/microbial mat origin, or varve-like fluctuations in deposition, perhaps in a deep basin. The Pliocene/Miocene boundary falls within an interval of banded micritic silty clays that reflect the final environmental fluctuations during the transition to the open marine conditions of the Pliocene.

Bulk geochemistry and grain-size composition of surface sediments from the SW Atlantic

Surface sediments from the South American continental margin surrounding tbe Argentine Basin were studied with respect to bulk geochemistry (Caeo) and C ) and grain-size composition (sand/silt/clay relation and terrigenous silt grain-size distribution). The grain-size distributions of the terrigenous silt fraction were unmixed into three end members (EMs), using an end-member modelling algorithm. Three unimodal EMs appear to satisfactorily explain the variations in the data set of the grain-size distributions ofterrigenous silt. The EMs are related to sediment supply by rivers, downslope transport, winnowing, dispersal and re-deposition by currents. The bulk geochemical composition was used to trace the distribution of prominent water masses within the vertical profile. The sediments of the eastern South American continental margin are generally divided into a coarse-grained and carbonate-depleted southwestern part, and a finer-grained and carbonate-rich northeastern part. The transition of both environments is located at the position of the Brazil Malvinas Confluence (BMC). The sediments below the confluence mixing zone of the Malvinas and Brazil Currents and its extensions are characterised by high concentrations of organic carbon, low carbonate contents and high proportions of the intennediate grain-size end member. Tracing these properties, the BMC emerges as a distinct north-south striking feature centered at 52-54°W crossing the continental margin diagonally. Adjacent to this prominent feature in the southwest, the direct detrital sediment discharge of the Rio de la Plata is clearly recognised by a downslope tongue of sand and high proportions of the coarsest EM. A similar coarse grain-size composition extends further south along the continental slope. However, it displays bener sorting due to intense winnowing by the vigorous Malvinas Current. Fine-grained sedimentary deposition zones are located at the southwestern deeper part of the Rio Grande Rise and the southern abyssal Brazil Basin, both within the AABW domain. Less conspicuous winnowing/accumulation panerns are indicated north of the La Plata within the NADW level according to the continental margin topography. We demonstrate that combined bulk geochemical and grain-size properties of surface sediments, unmixed with an end-member algorithm, provide a powerful tool to reconstruct the complex interplay of sedimentology and oceanography along a time slice.

Mineral composition, major element oxides and trace element concentration of sediments from ODP Leg 180 sites

Middle Miocene to Holocene fine-grained argillaceous sediments (clays, claystones/muds, and mudstones), which volumetrically dominated the sediment recovery in the Woodlark Basin during Leg 180, were chemically analyzed for major elements, trace elements, and some rare earth elements by X-ray fluorescence. Selected samples also underwent X-ray diffraction (XRD) analysis for mineral determination. The results shed light on sediment provenance when combined with shipboard sediment descriptions, smear slide study, and XRD. The oldest sediments recovered (Site 1108) of middle-late Miocene age include volcanogenic muds with distinctive high MgO and K2O, indicative of a relatively basic calc-alkaline source related to an inferred Miocene forearc succession. The forearc basement, composed of diabase and basalt, was locally exposed (Site 1109) and eroded in the late Miocene (<5.4-9.93 Ma), giving rise to fluvial conglomerates (Sites 1109, 1115, and 1118). Chemically distinctive fine-grained claystones and siltstones (with relatively high Ti, low K) are compatible with derivation from tropically weathered basic igneous rocks, correlated with the Paleogene Papuan ophiolite. Overlying latest Miocene-Pleistocene fine-grained sediments throughout the Woodlark Basin were partly derived from calc-alkaline volcanic sources. However, relatively high abundances of Al2O3 and related element oxides (K2O and Na2O) and trace elements (e.g., Rb and Y) reflect an additional terrigenous input throughout the basin, correlated with pelitic metamorphic rocks exposed on Papua New Guinea and adjacent areas. In addition, sporadic high abundances of Cr and Ni, some other trace metals, and related minerals (talc, crysotile, and chlorite) reflect input from an ophiolitic terrain dominated by ultramafic rocks, correlated with the Paleogene Papuan ophiolite. The source areas possibly included serpentinized ultramafic ophiolitic rocks exposed in the Papua New Guinea interior highlands. Chemical evidence further indicates that fine-grained terrigenous sediment reached the Woodlark Basin throughout its entire late Miocene-Holocene history. Distinctive high-K volcanogenic muds rich in tephra and volcanic ash layers that appear at <2.3 Ma (Sites 1109 and 1115) are indicative of high-K calc-alkaline volcanic centers, possibly located in the Dawson Strait, Moresby Strait, or Dobu Seamount area. Chemical diagenesis of fine-grained sediments within the Woodlark Basin is reflected in clay neomorphism and localized formation of minerals including dolomite, ankerite, and zeolite but has had little effect on the bulk chemical composition of most samples.

Whole-rock geochemistry of amphibolites and metagabbros from ODP Holes 173-1067A and 173-1068A

The Leg 173 Site 1067 and 1068 amphibolites and metagabbros from the west Iberia margin exhibit variable whole-rock compositions from primitive to more evolved (Mg numbers = 49-71) that are generally incompatible trace and rare earth element enriched (light rare earth element [LREE] = 11-89 x chondrite). The Site 1067 amphibolites are compositionally similar to the basalts reported at Site 899 from this same region, based on trace and rare earth element contents. The Site 1068 amphibolites and metagabbros are similar to the Site 899 diabases but are more LREE enriched. However, the Sites 1067 and 1068 amphibolites and metagabbros are not compositionally similar to the Site 900 metagabbros, which are from the same structural high as the Leg 173 samples. The Leg 173 protoliths may be represented by basalts, diabases, and/or fine-grained gabbros that formed from incompatible trace element-enriched liquids.

Mineralogical, sedimentological and paleontological analyses of ODP Hole 127-797B upper Quarternary sediments

Nearly continuous cores of Quaternary fine-grained sediments with distinct dark-light colored cycles were recovered from Sites 794, 795, and 797 in the basinal parts of the Japan Sea during Leg 127. A comparison of gray value (darkness) profiles supplemented by visual inspection of core photographs between sites indicated that most of the dark and light layers were correlatable between sites, and that two of the dark layers lie close to adjacent marker ash layers. These observations indicate that deposition of dark and light layers resulted from basin-wide synchronous events. In order to understand the origin of these dark-light cycles, petrographical, mineralogical, compositional, and paleontological studies were carried out on closely spaced samples from the upper Quaternary sediments recovered from Site 797. Age model was constructed based on comparison between variation in diatom abundance and the standard oxygen isotope curve of Imbrie et al. (1984), the latter was interpolated between the five age controlled levels established at Site 797. The two curves show similar patterns which enabled us to "tune" the sediment ages to the oxygen isotope stages. We have to use variation in diatom abundance as a substitute for oxygen isotope curve since oxygen isotopic data are not available at the studied sites. Bottom water oxygenation conditions were estimated based on two criteria: (1) the degree of lamina preservation and (2) the ratio of Corg to Stot. The surface water productivity was deduced from the Corg and biogenic silica content. Results suggest that the bottom water oxygenation level and the surface water productivity varied significantly in response to the glacial-interglacial cycles. Glacio-eustatic sea-level changes and subsequent changes in water circulation in the Japan Sea appear to have been responsible for these variations and consequent changes in sediment composition throughout the Quaternary.

Nd and Sr isotopic compositions and isotope dilution trace element concentrations from 7 sediment cores along the southeast Greenland margin

The Nd and Sr isotopic compositions of Quaternary glacial and glacimarine siliciclastic sediments deposited along the margin of southeast Greenland were determined to assess the roles of the Greenland, Iceland, and more distal ice sheets in delivering detritus to this portion of the northern North Atlantic. The isotopic compositions of detritus generated by portions of the southern Greenland Ice Sheet were defined through measurements of till and trough mouth fan sediments. Massive diamicts from the Scoresby Sund trough mouth fan show a restricted range of e-Nd (-11.8 to -16.6) and 87Sr/86Sr (0.7192-0.7246) consistent with their derivation from mixtures of sediments derived from Paleoproterozoic and/or Caledonian basement and Tertiary Greenland basalts. Further south at Kangerlussuaq, till isotopic compositions covary with the underlying basement type, with low e-Nd values in the inner fiord (-18.1) reflecting the erosion of the local Precambrian gneisses, but with higher e-Nd values (-2.3 to 2.5) found where the trough crosses East Greenland Tertiary basalts. Fine-grained (< 63 µm) sediments deposited along the southeast Greenland margin also show regular spatial isotopic variations. Ambient sediments and ice-rafted detritus in the southern Irminger Basin trend towards low e-Nd values (to ~ -28) and 87Sr/86Sr ratios (~ 0.711 to ~ 0.715) and are likely derived from proximal Archean gneisses of SE Greenland. Further north in the northern Irminger and Blosseville Basins, sediments trend toward much higher e-Nd (> -4) and low 87Sr/86Sr (< 0.709) reflecting a component derived from the local Iceland volcanic rocks and/or the East Greenland Tertiary basalts. In all three regions, the locally-derived detritus is intermixed with sediment with an intermediate e-Nd value (~ -10) and 87Sr/86Sr (~ 0.718) that was likely delivered by icebergs emanating from the Eurasian Ice Sheets and not from eastern Greenland. Deposition of glacial sediments from both proximal and distal (Eurasian) sources occurred adjacent to SE Greenland throughout the past 50 Ka, with periodic increases in IRD deposition at various times including those of Heinrich events 1, 2 and 4. These results suggest that at least the southern portions of the Greenland Ice Sheet experienced periodic instabilities during the Last Glacial period.

Mineralogy and isotope chemistry of GISP2 and potential source areas

Samples of dust from the Greenland Ice Sheet Project 2 (GISP2) ice core, Summit, Greenland, dated within marine isotope stage 2 (between 23,340 and 26,180 calendar years B.P.) around the time of the coldest, local, last glacial temperatures, have been analyzed to determine their provenance. To accomplish this, we have compared them with approximately Coeval aeolian sediments (mostly loesses) sampled in possible source areas (PSAs) from around the northern hemisphere. The <5-µm grain-size fraction of these samples was analyzed on the basis that it corresponds to the atmospheric dust component of that time and locale, which was sufficiently fine grained to be transported over long distances. On the basis of comparison of the clay mineralogy and Sr, Nd and Pb isotope composition with ice dust and PSAs and assuming that we have sampled the most important PSAs, we have determined that the probable source area of these GISP2 dusts was in eastern Asia. The dust was not derived from either the midcontinental United States or the Sahara, two more proximal areas that have been suggested as potential sources based on atmospheric circulation modeling. Except for a brief period during an interstadial, when dust transport was exceptionally low (for glacial times) and had a mineralogical composition indicative of a slightly more southern provenance, the source area of the dust did not change significantly during times of variably higher fluxes of dust with larger mean grain size or lower fluxes of dust with smaller mean grain size. This includes the high-dust period that correlates with the Heinrich 2 period of major iceberg discharge into the North Atlantic. Variable wind strengths must therefore be invoked to account for these abrupt and significant changes in dust flux and grain size.

Grain-size distribution, sedimentation rate and x-ray fluorescence data of four sediment cores off the Senegal River estuary

Fine-grained sediment depocenters on continental shelves are of increased scientific interest since they record environmental changes sensitively. A north-south elongated mud depocenter extends along the Senegalese coast in mid-shelf position. Shallow-acoustic profiling was carried out to determine extent, geometry and internal structures of this sedimentary body. In addition, four sediment cores were retrieved with the main aim to identify how paleoclimatic signals and coastal changes have controlled the formation of this mud depocenter. A general paleoclimatic pattern in terms of fluvial input appears to be recorded in this depositional archive. Intervals characterized by high terrigenous input, high sedimentation rates and fine grain sizes occur roughly contemporaneously in all cores and are interpreted as corresponding to intensified river discharge related to more humid conditions in the hinterland. From 2750 to 1900 and from 1000 to 700 cal a BP, wetter conditions are recorded off Senegal, an observation which is in accordance with other records from NW-Africa. Nevertheless, the three employed proxies (sedimentation rate, grain size and elemental distribution) do not always display consistent inter-core patterns. Major differences between the individual core records are attributed to sediment remobilization which was linked to local hydrographic variations as well as reorganizations of the coastal system. The Senegal mud belt is a layered inhomogeneous sedimentary body deposited on an irregular erosive surface. Early Holocene deceleration in the rate of the sea-level rise could have enabled initial mud deposition on the shelf. These favorable conditions for mud deposition occur coevally with a humid period over NW-Africa, thus, high river discharge. Sedimentation started preferentially in the northern areas of the mud belt. During mid-Holocene, a marine incursion led to the formation of an embayment. Afterwards, sedimentation in the north was interrupted in association with a remarkable southward shift in the location of the active depocenter as it is reflected by the sedimentary architecture and confirmed by radiocarbon dates. These sub-recent shifts in depocenters location are caused by migrations of the Senegal River mouth. During late Holocene times, the weakening of river discharge allowed the longshore currents to build up a chain of beach barriers which have forced the river mouth to shift southwards.

High-resolution sediment record from a submarine meandering canyon system in the eastern Atlantic

Based on a high-resolution sediment record from a submarine meandering canyon system offshore the present-day hyperarid Saharan Africa, two phases of turbidity-current activity can be distinguished during the past 13,000 years. Frequent, siliciclastic turbidity currents can be related to deglacial sea-level history, whereas rhythmically recurring fine-grained and carbonate-rich turbidity currents with recurrence times of roughly 900 years are inferred for the Holocene. Various trigger mechanisms can be considered to initiate turbidity currents, but only a few can explain a periodic turbidite activity. A comparison of Holocene turbidite recurrence times and basic cycles of 900 and 1,800 years found in various Holocene paleoclimate studies suggests that a previously unrecognized climate-related coupling may be active.

(Table T1) Mineral composition of turbidite sands and sandstones from ODP Leg 190 sites

During Ocean Drilling Program Leg 190 several turbidite successions in the Nankai Trough were drilled through including Pleistocene trench fill (Sites 1173 and 1174), Pleistocene-Pliocene slope basin deposits and underlying trench fill (Sites 1175 and 1176), Miocene Shikoku Basin deposits (Site 1177), and upper Miocene trench fill (Site 1178). Sands from the Pleistocene trench-fill succession of the Nankai Trough are of mixed derivation with significant monomineralic components (quartz and feldspar) and mafic to intermediate volcanic rock fragments, in addition to sedimentary and less abundant metamorphic detritus. They have a source in the Izu collision zone in central Honshu. Sands from the slope and accreted trench fill at Sites 1175 and 1176 are dominated by quartz with less abundant feldspar, sedimentary rock fragments, and only minor volcanic and metamorphic rock fragments. In contrast to the trench turbidites of Sites 1173 and 1174, these sands are very quartzose with characteristic radiolarian chert fragments. Volcanic rock fragments are mainly of silicic composition. Potential sources of these sands are uplifted subduction complexes of southwest Japan. Sands from the accreted trench turbidites at Site 1178 have clast types similar to those at Sites 1175 and 1176. In contrast, however, framework detrital modes are distinctive, with Site 1178 sands having substantially lower total quartz contents and more abundant fine-grained sedimentary rock fragments. These sands were also probably derived from the island of Shikoku, but their composition indicates that sedimentary rocks were abundant in the source area and these may have been Miocene forearc basin successions that were largely removed by erosion. Erosional remnants of Miocene forearc basin deposits are present on the Kii Peninsula east-northeast of Shikoku. Erosion followed a phase of exhumation of the Shimanto Belt indicated by apatite fission track ages at ~10 Ma. Sand in the lower-upper Miocene turbidites of the lower Shikoku Basin section at Site 1177 is more varied in composition, with the upper part of the unit similar to Site 1178 (i.e., rich in sedimentary rock fragments) and the lower part similar to those at Sites 1175 and 1176 (i.e., rich in quartz with some silicic volcanic rock fragments). Sands from the lower part of the Miocene turbidite unit were derived from a continental source with plutonic and volcanic rocks, possibly the inner zone of southwest Japan.

Environmental magnetism, geochemical measurements and colour reflectance of two sediment cores: GeoB4905-4 off Cameroon and GeoB4906-3 off Gabon

We combine environmental magnetism, geochemical measurements and colour reflectance to study two late Quaternary sediment cores: GeoB 4905-4 at 2° 30 N off Cameroon and GeoB 4906-3 at 0° 44 N off Gabon. This area is suitable for investigating precipitation changes over Central and West Africa because of its potential to record input of aeolian and fluvial sediments. Three magnetozones representing low and high degree of alteration of the primary rock magnetic signals were identified. The magnetic signature is dominated by fine-grained magnetite, while residual haematite prevails in the reduced intervals, showing increase in concentration and fine grain size at wet intervals. Our records also show millennial-scale changes in climate during the last glacial and interglacial cycles. At the northern location, the past 5.5 ka are marked by high-frequency oscillations of Ti and colour reflectance, which suggests aeolian input and hence aridity. The southern location remains under the influence of the Intertropical Convergence Zone and thus did not register aeolian signals. The millennial-scale climatic signals indicate that drier and/or colder conditions persisted during the late Holocene and are synchronous with the 900 a climatic cycles observed in Northern Hemisphere ice core records.

Ice rafte debris of ODP Hole 178-1101A

Sediment drifts on the continental rise west of the Antarctic Peninsula received fine-grained sediment and ice-rafted debris (IRD) directly from the continental shelf and thus indirectly record the history of West Antarctic glaciation. Site 1101 contains a 218-m-thick, nearly continuous section extending from the late Pliocene to the Holocene. To assess the presence of calving glaciers at sea level in the Antarctic Peninsula region, the mass accumulation rate (MAR) of IRD was calculated using the weight percent terrigenous sand fraction (250 µm to 2 mm). IRD MAR is cyclic throughout, with small peaks alternating with periods of low or no IRD. Many cycles have a sawtooth pattern that increases gradually to the peak then abruptly decreases to zero. This pattern is consistent with rapid disintegration of ice streams and release of icebergs from the continental shelf. Three unusually large peaks (three to five times the size of other peaks) occurred at approximately 2.8, 1.9, and 0.88 Ma and indicate periods of intense ice rafting. Lithofacies were described in detail using X-radiographs and core descriptions for the interval from 1.34 to 0.54 Ma. Glacial units are represented by thickly laminated mud deposited by distal turbidites and meltwater plumes. Less commonly, thinly laminated sediment formed by contour currents and diamicton by intense ice rafting. Interglacials are represented by foraminifer-bearing mud with IRD. Ice rafting appears to have increased in the later part of the glacial period and remained high in the interglacial period.