Archaeological sites in China during the Neolithic and Bronze Age

A total of 51,074 archaeological sites from the early Neolithic to the early Iron Age (c. 8000-500 BC), with a spatial extent covering most regions of China (c. 73-131°E and c. 20-53°N), were analysed over space and time in this study. Site maps of 25 Chinese provinces, autonomous regions and municipalities, published in the series 'Atlas of Chinese Cultural Relics', were used to extract, digitalise and correlate its archaeological data. The data were, in turn, entered into a database using a self-developed mapping software that makes the data, in a dynamic way, analysable as a contribution to various scientific questions, such as population growth and migrations, spread of agriculture and changes in subsistence strategies. The results clearly show asynchronous patterns of changes between the northern and southern parts of China (i.e. north and south of the Yangtze River, respectively) but also within these macro-regions. In the northern part of China (i.e. along the Yellow River and its tributaries and in the Xiliao River basin), the first noticeable increase in the concentration of Neolithic sites occurred between c. 5000 and 4000 BC; however, highest site concentrations were reached between c. 2000 and 500 BC. Our analysis shows a radical north-eastern shift of high site-density clusters (over 50 sites per 100 * 100 km grid cell) from the Wei and middle/lower Yellow Rivers to the Liao River system sometime between 2350 BC and 1750 BC. This shift is hypothetically discussed in the context of the incorporation of West Asian domesticated animals and plants into the existing northern Chinese agricultural system. In the southern part of China, archaeological sites do not show a noticeable increase in the absolute number of sites until after c. 1500 BC, reaching a maximum around 1000 BC.

Sedimentology of cores from the Bellingshausen Sea, Antarctic

On the continental rise west of the Antarctic Peninsula there are nine large mounds interpreted as sediment drifts, separated by turbidity current channels. Drift 7 is 150 km long, 70 km wide and up to 700 m high and is asymmetric, with steep sides on the south-east (towards the continent) and south-west, and gentle slopes to north-west and north-east. Cores on the gentle sides of the drift show a cyclicity between brown, bioturbated, diatom-bearing mud with foraminifera and radiolarians, and grey, laminated, barren mud. Biostratigraphic evidence is consistent with a Late Quaternary age. Detailed lithostratigraphy and magnetic susceptibility data allow precise correlation over distances of tens of kilometres. On the basis of chemostratigraphy, the brown sediment is interpreted as interglacial (isotope stages 1 and 5) and the grey as glacial (stages 2-4 and 6). Sedimentation rates are 3.0-5.5 cm/ka. Cores on the steep sides of the drift recovered a condensed section with thinner cycles and hiatuses. Fine grain size, very poor sorting and the absence of a mode in the silt size range indicate deposition from suspension with only weak current activity, There is little evidence for cyclic changes in bottom current strength. Supply of sediment to the benthic nepheloid layer was by entrainment ofmud from turbidity currents, and by settling ofpelagic material (biogenic grains, IRD, sediment suspended in meltwater plumes). Cyclic changes in sediment supply include more biogenic supply in interglacials with less sea ice cover, more terrigenous supply from turbidites in glacials with ice sheets grounded to the shelf edge, and changes in IRD content.

Granulometry, geochemistry, amino acids, radiocarbon ages and paleomagnetics of samples from Heidemann Valley

A Pliocene (2.6-3.5 Ma) age is determined from glacial sediments studied in a 20m long, 4 m deep trench excavated in Heidemann Valley, Vestfold Hills, East Antarctica. The age determination is based on a combined study of amino acid racemization, diatoms, foraminifera, and magnetic polarity, and supports earlier estimates of the age of the sedimentary section; all are beyond 14C range. Four till units are recognized and documented, and 16 subunits are identified. All are ascribed to deposition during a Late Pliocene glaciation that was probably the last time the entire Vestfold Hills was covered by an enlarged East Antarctic Ice Sheet (EAIS). Evidence for other more recent glacial events of the 'Vestfold Glaciation' may have been due to lateral expansion of the Sorsdal Glacier and limited expansion of the icesheet margin during the Last Glacial Maximum rather than a major expansion of the EAIS. The deposit appears to correlate with a marine deposition event recorded in Ocean Drilling Program Site 1166 in Prydz Bay, possibly with the Bardin Bluffs Formation of the Prince Charles Mountains and with part of the time represented in the ANDRILL AND-1B core in the Ross Sea.

Age determinations in Lena Delta

The age correlation between the three main geomorphological terraces in the Lena Delta, especially that of the second sandy terrace (Arga Island) and the third terrace (Ice Complex and underlying sands) is still being discussed, Knowledge about the age of the lee Complex and its underlying sands, and the Arga sands is necessary for understanding the past and modern structure of the delta. Geochronometrie data have been acguired for three sediment seguences from the Lena Delta by lumineseence dating using the potassium feldspar IR-OSL technique. Additionally, 14C dates are available for geochronological discussion. Typical sediments of the upper part of Arga Island as found in the area of Lake Nikolay are of Late Pleistoeene age (14.5-10.9 ka), Typical third terrace sediments from two seguenees located at the Olenyokskaya branch are older. At the profile "Nagym" sandy seguences were most probably deposited between about 65 ka and 50 ka before present. The lower part of the sandy seguence at "Kurungnakh Island" is possibly older than the sediments of the section at Nagym. However, methodological difficulties in luminescence dating (insufficient bleaching at the time of deposition) and younger 14C dates make the discussion of the results difficult.

(Table 1) Concentrations of amino acids in sediments from DSDP Site 68-502

In this chapter, we will report on the amino acids in the total acid hydrolysate of eight sediment samples from Leg 68 Site 502. This site was located on a topographic high at a depth of 3051 meters in the Colombian Basin of the western Caribbean Sea. Four holes were cored at the site by means of the hydraulic piston corer to a maximum sediment depth of 218 meters. The composite section is a virtually continuous, undisturbed sediment record covering almost 8 million years from the Holocene to late Miocene. Age estimates for the section are based on excellent magnetostratigraphic and biostratigraphic records. Four lithostratigraphic units (A, B, C, and D) were recognized, based on differences in color and content of clay, ash, foraminifers, and siliceous microfossils (Prell, Gardner, et al., 1980): A, yellowish brown to light brownish gray foraminifer-bearing (> 10%) nannofossil marl; B, gray to olive gray foraminifer-bearing nannofossil marl with occasional ash beds; C, light gray to dark greenish gray calcareous clay and foraminifer-bearing (< 10%) nannofossil marl; D, pale green to grayish green calcareous, ash-bearing clay with siliceous microfossils. The calcium carbonate content of these sediments increases from about 27 to about 49% from late Miocene to middle Pliocene (about 3.6 Ma) and remains uniform at about 48 to 50% from that time throughout the Quaternary. The eight sediment samples for amino acid analyses came from the third (502B) and fourth (502C) holes at Site 502. Samples ranged in sub-bottom depth from 4.3 to 225 meters spanning time from 0.3 to 7.7 Ma.

Composition of bottom sediments from the Mediterranean Sea off Syria

An investigation of recent bottom sediments between the Cyprus Island and the Syrian seacoast during Cruise 27 of R/V Vityaz-2 (1993) gave comprehensive field data significantly complementing our understanding of the sedimentation process in this part of the Mediterranean Sea. Mineralogical and geochemical indicators testify to different input into sedimentation of the Syrian and Nile River sources. The Nile River plays a leading role in terrigenous sedimentation in the southeastern Mediterranean Sea, especially in deep-sea areas. In contrast, contribution of weathering products of basalts and ophiolites from the Syrian drainage area (hornblende, monoclinic and rhombic pyroxenes, olivine, spinel, palagonite, and epidote) are particularly detectable in sediments of the near-coast zone. During Late Quaternary contribution of terrigenous material both from the Syrian and Nile sources was irregular in time.

Late Cretaceous planktonic foraminiferal assemblages from the Exmouth Plateau, ODP Sites 122-762 and 122-763, eastern Indian Ocean

The evolution of planktonic foraminifera during the Late Cretaceous is marked in the Santonian by the disappearance of complex morphotypes (the marginotruncanids), and the contemporary increasing importance and diversification of another group of complex taxa, the globotruncanids. Upper Turonian to lower Campanian planktonic foraminiferal assemblages from Holes 762C and 763B (Ocean Drilling Program, Leg 122, Exmouth Plateau, 47°S palaeolatitude) were studied in detail to evaluate the compositional variations at the genus and species level based on the assumption that, in the Cretaceous oceans as in the modern, any faunal change was associated with changes in the characteristics and the degree of stability of the oceanic surface waters. Three major groups were recognised based on gross morphology, and following the assumption that Cretaceous planktonic foraminifera, although extinct, had life-history strategies comparable to those of modern planktonics: 1 - r-selected opportunists; 2 - k-selected specialists; 3 - r/k intermediate morphotypes which include all genera that display a range of trophic strategies in-between opportunist and specialist taxa. Although planktonic foraminiferal assemblages are characterised by a progressive appearance of complex taxa, this trend is discontinuous. Variation in number of species and specimens within genera has allowed recognition of five discrete intervals each of them reflecting different oceanic conditions based on fluctuations in diversity and abundance of the major morphotypes. Planktonic forms show cyclical fluctuations in diversity and abundance of cold (r-strategists) and warm taxa (k-strategists), perhaps representing alternating phases of unstable conditions (suggesting a weakly stratified upper water column in a mesotrophic environment), and well-stratified surface and near-surface waters (indicating a more oligotrophic environment). Interval 1, middle Turonian to early Coniacian in age, is dominated by the r/k intermediate morphotypes which alternate with r-strategists. These cyclical alternations are used to identify three additional subintervals. Interval 2, aged middle to late Coniacian, is characterised by the increasing number of species and relative abundance of k-strategists. After this maximum diversification the k-strategists show a progressive decrease reaching a minimum value in Interval 3 (early to late Santonian), which corresponds to the extinction of the genus Marginotruncana. In the Interval 4, latest Santonian in age, the k-strategists, represented mainly by the genera Globotruncana, increase again in diversity and abundance. The last Interval 5 (early Campanian) is dominated by juvenile globotruncanids and r-strategists which fluctuate in opposite phase. The positive peak (Interval 2) related to the maximum diversification of warm taxa (k-strategists) in the Coniacian seems to correspond to a warmer episode. It is followed by a marked decrease in the relative abundance of warm taxa (k-strategists crisis) with a minimum in the late Santonian (Interval 3), reflecting a decrease in temperature. Detailed analysis of faunal variations allows the Santonian faunal turnover to be ascribed to a cooling event strong enough to cause the extinction of the marginotruncanids.

Late pleistocene and holocene sedimentation in the central and eastern Persian Gulf

The sandfraction of the sediment was analysed in five cores, taken from 65 m water depth in the central and eastern part of the Persian Gulf. The holocene marls are underlayn by aragonite muds, which are probably 10-11,000 years old. 1. The cores could be subdivided into coarse grained and fine grained layers. Sorting is demonstrated by the following criteria: With increasing median values of the sandfraction - the fine grained fraction decreases within each core; - the median of each biogenic component, benthonic as well as planktonic, increases; - the median of the relict sediment, which in core 1179 was carried upward into the marl by bioturbation, increases; - the percentages of pelecypods, gastropods, decapods and serpulid worms in the sandfraction increase, the percentages of foraminifera and ostracods decrease; - the ratios of pelecypods to foraminifera and of decapods to ostracods increase; - the ratios of benthonic molluscs to planktonic molluscs (pteropods) and of benthonic foraminifera to planktonic foraminifera increase (except in core 1056 and 1179); - the ratio of planktonic molluscs (pteropods) to planktonic foraminifera increases; - the globigerinas without orbulinas increase, the orbulinas decrease in core 1056. Different settling velocities of these biogenic particles help in better understanding the results : the settling velocities, hence the equivalent hydrodynamic diameters, of orbulinas are smaller than those of other globigerinas, those of planktonic foraminifera are smaller than those of planktonic molluscs, those of planktonic molluscs are smaller than those of benthonic molluscs, those of pelecypods are smaller than those of gastropods. Bioturbation could not entirely distroy this "grain-size-stratification". Sorting has been stronger in the coarse layers than in the finer ones. As a cause variations in the supply of terrigenous material at constant strength of tidal currents is suggested. When much terrigenous material is supplied (large contents of fine grained fraction) the sedimentation rates are high: the respective sediment surface is soon covered and removed from the influence of tidal currents. When, however, the supply of terrigenous material is small, more sandy material is taken away in all locations within the influence of terrigenous supply. Thus the biogenic particles in the sediment do not only reflect the organic production, but also the influence of currents. 2. There is no parameter present in all cores that is independently variable from grain size and can be used for stratigraphic correlation. The two cores from the Strait of Hormus were correlated by their sequences of coarse and fine grained layers. 3. The sedimentation rates of terrigenous material, of total planktonic and benthonic organisms and of molluscs, foraminifera, echinoids and ophiuroids are shown in table 1 (total sediment 6.3-75.5 cm/1000 yr, biogenic carbonate 1.9-3.6 cm/1000 yr). The sedimentation rates of benthonic organisms are nearly the same in the cores of the Strait of Hormus, whereas near the Central Swell they are smaller. In the upper parts of the two cores of the Strait of Hormus sedimentation rates are higher than in the deeper parts, where higher median values point to stronger reworking. 4. The sequence of coarse and fine grained intervals in the two cores of the Hormus Strait, attributed to variations in climate, as well as the increase of terrigenous supply from the deeper to the upper parts of the cores, agrees with the descriptions in the literature of the post Pleistocene climate as becoming more humid. The rise of sea level is sedimentologically not measurable in the marly sediments - except perhaps for the higher content of echinoids in the lower part of core 1056. These may be attributed to the influence of a migrating wave-base. 5. The late Pleistocene aragonite mud is very fine grained (> 50%< 2 p) and poor in fossils (0.5-1.8%) biogenic particles of total sediment. The sand fraction consists almost entirely of white clumps, c. 0.1 mm in diameter (1177), composed of aragonite needles and of detrital minerals with the same size (1201). The argonite mud was probably not formed in situ, because the water depth at time of formation was at most 35 m at least 12 m. The sorting of the sediment (predominance of the fine grained sand), the absence of larger biogenic components and of pellets, c. 0.2-0.5 mm in diameter, which are typical for Recent and Pleistocene locations of aragonite formation, as well as the sedimentological conditions near the sampling points, indicate rather a transport of aragonite mud from an area of formation in very shallow waters. Sorting as well as lenticular fabric in core 1201 point to sedimentation within the influence of currents. During alternating sedimentation - and reworking processes the aragonitic matrix was separated from the silt - and sand-sized minerals. The lenses grade into touches because of bioturbation. 6. In core 1056 D2 from Hormus Bay the percentages of organic carbon, total nitrogen and total carbonate were determined. With increasing amounts of smaller grain sizes the content of organic matter increases, whereas the amount of carbonate decreases. The amounts of organic carbon and of nitrogen decrease with increasing depth, probably due to early-diagenetic decomposition processes. Most of the total nitrogen is of organic origin, only about 10% may well be inorganically fixed as ammonium-nitrogen. In the upper part of the core the C/N-ratio increases with increasing depth. This may be connected with a stronger decomposition of nitrogen-containing organic compounds. The general decrease of the C/N-ratios in the lower part of the core may be explained by the relative increase of inorganically fixed ammonium-nitrogen with decreasing content of organic matter.

Dinoflagellate cyst assemblages, oxygen isotope records of foraminifera, and alkenone and Mg/Ca-based SST estimates for the time interval 3.40-3.18 Ma (Late Pliocene) from five North Atlantic and Caribbean sediment cores

The early Late Pliocene (3.6 to ~3.0 million years ago) is the last extended interval in Earth's history when atmospheric CO2 concentrations were comparable to today's and global climate was warmer. Yet a severe global glaciation during marine isotope stage (MIS) M2 interrupted this phase of global warmth ~3.30 million years ago, and is seen as a premature attempt of the climate system to establish an ice-age world. Our geochemical and palynological records from five marine sediment cores along a Caribbean to eastern North Atlantic transect show that increased Pacific-to-Atlantic flow via the Central American Seaway weakened the North Atlantic Current (NAC) and attendant northward heat transport prior to MIS M2. The consequent cooling of the northern high latitude oceans permitted expansion of the Greenland ice sheet during MIS M2, despite near-modern atmospheric CO2 concentrations. Before and after MIS M2, heat transport via the NAC was crucial in maintaining warm climates comparable to those predicted for the end of this century.

Late Cretaceous to Eocene calcareous nannofossil distribution near Gubbio, Italy

Using a modified sample preparation technique, we have been able to establish a detailed lower Campanian to upper Eocene nannofossil stratigraphy in the Bottaccione and Contessa Highway sections near Gubbio. Appearance and extinction levels of virtually all the commonly used calcareous nannofossil zonal markers have been recognized and can now be closely correlated with the planktonic foraminifera zonation and the magnetic reversal stratigraphy previously established in these sections. Comparisons with the nannofossil calibrations of the oceanic magnetic anomaly sequence in Deep Sea Drilling Project (DSDP) sites suggest that magmetic Subchrons C17N and C25N are missing in the Bottaccione section. The observed variability of the relative stratigraphic position of most plankton events is confirmed to less than one magnetic subchron. Absolute abundance, paleobiogeographic restriction, and differential preservation render some of the traditionally used biostratigraphic events less reliable than others.

Radioisotope stratigraphy, sedimentology and geochemistry of Late Quaternary sediments from the Eastern Arctic Ocean

To reconstruct Recent and past sedimentary environments, marine sediments of Upper Pleistocene and Holocene ages from the eastern Arctic Ocean and especially from the Nansen-Gakkel Ridge (NGR) were investigated by means of radioisotopic, geochemical and sedimentological methods. In combination with mass physical property data and lithological analysis these investigations allow clearly to characterize the depositional environments. Age dating by using the radioisotope 230Th gives evidence that the investigated sediments from the NGR are younger than 250,000 years. Identical lithological sediment sequences within and between sediment cores from the NGR can be related to sedimentary processes which are clearly controlled by palaeoclimate. The sediments consist predominantly of siliciclastic, terrigenous ice-rafted detritus (IRD) deriving from assorted and redeposited sediments from the Siberian shelfs. By their geochemical composition the sediments are similar to mudstone, graywacke and arcose. Sea-ice as well as icebergs play a major roll in marine arctic sedimentation. In the NGR area rapid change in sedimentary conditions can be detected 128,000 years ago. This was due to drastic change in the kind of ice cover, resulting from rapid climatic change within only hundreds of years. So icebergs, deriving mostly from Siberian shelfs, vanished and sea-ice became dominant in the eastern Arctic Ocean. At least three short-period retreats of the shelf ice between 186,000 and 128,000 years are responsible for the change of coarse to fine-grained sediments in the NGR area. These warmer stages lasted between 1,000 and 3,000 years. By monitoring and comparing the distribution patterns of sedimentologic, mass physical and geochemical properties with 230Th ex activity distribution patterns in the sediment cores from the NGR, there is clear evidence that sediment dilution is responsible for high 230Th ex activity variations. Thus sedimentation rate is the controlling factor of 230Th ex activity variations. The 230Th flux density in sediments from the NGR seems to be highly dependent On topographic Position. The distribution patterns of chemical elements in sediment cores are in general governed by lithology. The derivation of a method for dry bulk density determination gave the opportunity to establish a high resolution stratigraphy on sediment cores from the eastern Arctic Ocean, based on 230Thex activity analyses. For the first time sedimentation and accumulation rates were determined for recent sediments in the eastern Arctic Ocean by 230Th ex analyses. Bulk accumulation rates are highly variable in space and time, ranging between 0.2 and 30 g/cm**2/ka. In the sediments from the NGR highly variable accumulation rates are related to the kind of ice cover. There is evidence for hydrothermal input into the sediments of the NGR. Hydrothermal activity probably also influences surficial sediments in the Sofia Basin. High contents of As are typical for surficial sediments from the NGR. In particular SL 370-20 from the bottom of the rift valley has As contents exceeding in parts 300 ppm. Hydrothermal activity can be traced back to at least 130,000 years. Recent to subrecent tectonic activity is documented by the rock debris in KAL 370 from the NGR. In four other sediment cores from the NGR rift valley area tectonically induced movements can be dated to about 130,000 years ago, related most probably to the rapid climate change. Processes of early diagenesis in sediments from the NGR caused the aobilization and redeposition of Fe, Mn and Mo. These diagenetic processes probably took place during the last 130,000 years. In sediment cores from the NGR high amounts of kaolinite are related to coarse grained siliciclastic material, probably indicating reworking and redeposition of siberian sandstones with kaolinitic binding material. In contrast to kaolinite, illite is correlated to total clay and 232Th contents. Aragonite, associated with serpentinites in the rift valley area of the NGR, was precipitated under cold bottom-water conditions. Preliminary data result in a time of formation about 60 - 80 ka ago. Manganese precipitates with high Ni contents, which can be related to the ultrabasic rocks, are of similar age.

Sediments and tephra layers from the southern Tyrrhenian abyssal plain near Marsili Seamount

From the south-eastern Tyrrhenian deep-sea floor, four sediment cores of "Meteor" cruise 22 (1971) are described. These cores were taken in the basin between the Aeolian Islands and the Marsili Seamount, an elevation of more tha 3000 m above the sea floor. The sedimentation of the deep-sea basin is distinguished by a sequence of turbidites with a high sedimentation rate. The composition of the clastic material and the position of the cores in the mouth area of the morphologically very pronounced Stromboli Canyon suggest an interpretation of the turbidite sequence as fan of this canyon onto the deep-sea floor. A white rhyolitic pumice-tephra at the base of the 4 m thick sequence of turbidites in core M22-102 has been correlated with the Pelato eruption of the island of Liparo in the 6th century A.D. At the foot of the Marsili Seamount - apparently in morphologically elevated positions - the influence of the turbidite sedimentation increases, the rate of sedimentation is lower and stratigraphic omissions are probable. Here, rather compacted globigerina marls have been found in only 15 -25 cm depth. In addition, volcanic material in the form of lapilli layers, palagonitized ashes and detrital volcanic sands of the Marsili Seamount have been encountered in this area. An up to 3 cm thick layer of completely palagonitized basaltic ash intercalates with the marls at the base of two cores. Layers of very fresh olivine basaltic lapilli in core 103 and palagonitized lapilli of latitic composition in core 104 testify to an explosive submarine volcanism of the Marsili Seamount. According to the stratigraphy of core 103, the latest manifestations of this basaltic volcanism belong to the late Pleistocene (Emiliana huxleyi-zone of Nannoplankton stratigraphy) The basaltic lapilli are glassy to perhyaline with phenocrysts or microphenocrysts predominantely of olivine. The petrological character of the basaltic volcanites with high MgO, Ni, Cr and high MgO/FeO- and Ni/Co-ratios exhibits primitive basaltic features. These basalts clearly differ from basalts of the ocean floors, mid-ocean ridges and marginal basins. Prominent features are a missing iron-enrichment trend and low TiO2. Al2O3 tends to be high, as well as K2O and related trace elements (Ba, Sr). In spite of silica undrsaturation and high color index, the Marsili basalt exhibit some analogies with the calcalkaline basalts of the Aeolian arc, as well as the undersaturated basalts of some other circumoceanic areas.

Late Cenozoic eolian dust accumulation in the eastern equatorial Pacific Ocean

Sediments recovered during Ocean Drilling Program (ODP) Leg 138 in the eastern equatorial Pacific Ocean were analyzed for variations in eolian accumulation rate and mean grain-size. Latitudinal and temporal patterns of these parameters showed important changes in the intensity of atmospheric circulation and eolian flux associated with the intertropical convergence zone (ITCZ) and suggested that eolian input parameters could be used to define its paleoposition through time. Modern atmospheric circulation in the equatorial region is weakest in the intertropical convergence zone and increases as the trade winds are approached to the north and south. Thus, the expected spatial pattern of eolian grain size would have the finest material deposited beneath the ITCZ and a coarsening of material in both directions away from this zone. Sediments from ODP Leg 138 show this pattern for much of the Pleistocene and Pliocene but, prior to about 4 Ma, begin to lose the northern coarse component suggesting that the ITCZ was located north of its present position during the late Miocene. Eolian flux records also show a latitudinal pattern of deposition associated with the position of the ITCZ that, similar to eolian grain-size variability, suggests a more northerly position of the ITCZ during the late Miocene. Overall, the regional input of eolian material to the equatorial Pacific has decreased throughout the late Neogene. This reduction in eolian input reflects climatic changes to relatively wetter conditions in the continental eolian source regions beginning during the late Pliocene.

Organic geochemistry of late Cenozoic sediments from the eastern subtropical Atlantic

Detailed organic geochemical investigations have been performed on sediment samples from upwelling Site 658 and nonupwelling Sites 657 and 659. The major objective of this study has been the relationship between organic carbon accumulation and paleoclimatic and paleoceanographic conditions in the upwelling area off northwest Africa during late Cenozoic times. The study is based on results from organic carbon, nitrogen, and hydrogen analyses, Rock-Eval pyrolysis, kerogen microscopy, gas chromatography, and gas chromatography/mass spectrometry. In general, nonupwelling Sites 657 and 659 are characterized by low organic carbon values of less than 0.5%. At Site 657, four events of high organic carbon deposition (total organic carbon of 1%-3%) occur and represent turbidites and a slump interval. The upper Pliocene to Pleistocene sediments of upwelling Site 658 display high organic carbon contents of 0.5%-4%, with higher contents concentrated in the upper Pliocene. Accumulation rates of organic carbon vary between 0.1 and 0.5 gC/cm-**2/1000 yr, with maximum values between 3.5 and 3.1 Ma. Short-term cyclic ("Milankovitch-type") variations in organic carbon accumulation suggest climate-controlled mechanisms causing these fluctuations. The quality of organic matter at Site 658 is a mixture of kerogen type II and HI, with a dominance of the marine type. This is indicated by high hydrogen-index values of 200-400 mgHC/gC, low C/N ratios of 5-15, atomic H/C ratios of 1.0-1.5, and high amounts of marine macerals (alginite and liptodetrinite). We have estimated paleoproductivity for Sites 658 and 659 based on the amount of marine organic carbon. At open-marine Site 659, mean paleoproductivity varies between 20 and 50 gC/m**2/yr. At Site 658, mean paleoproductivity reaches high values of 160 to 320 gC/m**2/yr, very similar to those recorded in modern upwelling areas. The changes in productivity off northwest Africa are linked to changes in nutrient supply caused by both upwelling and fluvial input. The change from a dominantly humid climate to one characterized by fluctuations between humid and fully arid climates in northwest Africa occurs between 3.1 and 2.45 Ma.