Samples were obtained from either those taken aboard JOIDES Resolution during Leg 160 or in the Bremen Ocean Drilling Program (ODP) Core Repository. On the ship, approximately one-quarter of a core in sapropel intervals and the surrounding sediments was sliced off the working halves, wrapped in plastic, and stored frozen until shipping. In the shore-based laboratory, the quarter cores were cut in 1-cm subsamples and freeze dried. Samples taken in the Core Repository consisted of 1-cm sediment intervals across sapropels and adjoining sediments collected with plastic scoops, which were also freeze dried upon arrival in Warnemünde. For each sample, we calculated the revised meters composite depth (rmcd) below seafloor based on the tables provided by Sakamoto et al. (1998).The freeze-dried samples (~2-4 g dry weight) were subsequently split into subsamples. Two grams were wet sieved to obtain the >63-μm fraction, and the residue was used to pick up to 20 tests (sometimes samples were barren or contained less than 20 tests) of the planktonic foraminifer Globigerinoides ruber (white) for isotope analyses. This species is a shallow-dwelling planktonic foraminifer and the isotope signal of the calcite tests records global ice-volume variations, salinity, and temperature near the sea surface. The isotopic composition of oxygen and carbon in the calcite was analyzed with a Finigan MAT 251 mass spectrometer, and the measurement was calibrated against the Peedee Belemnite (PDB) reference scale through international isotope standards. The isotopic results are expressed in the usual notation delta = (Rs/Rst -1) * 1000, with Rs and Rst denoting the 13C/12C or 18O/16O ratios in the sample CO2 and the international standard, respectively. The standard deviation of the isotope analyses is ±0.02‰.The second subsample (~1-2 g dry weight) was used to extract lipids and to estimate the alkenone unsaturation index , a SST indicator (Brassell et al., 1986). Weighted splits of freeze-dried and homogenized sediments were extracted by ultrasonic agitation with 35-mL CH2Cl2 (2 ' 10 min). Elemental sulfur was removed by the addition of copper foil during the extraction. After each extraction step, samples were centrifuged, and the solvent was collected by pipette. The two lipid extracts were combined and dried in a rotary evaporator. The extracts were redissolved in CH2Cl2 and precleaned over a silica gel column (conditioned with 30-mL CH2Cl2) by elution with 14-mL CH2Cl2 and dried again in a rotary evaporator. The extracts were redissolved in n-hexane and subsequently fractionated using high-pressure liquid chromatography (HPLC). The HPLC column, packed with silica gel was MERCK LiChrospher Si 100-5. Four fractions were obtained using 5.5-mL n-hexane, 14-mL n-hexane/dichloromethane, 9-mL dichloromethane, and 9-mL acetone as eluents. The ketones (here: alkenones) were isolated in the second fraction. The first fraction contained the aliphatic hydrocarbons, and the third and fourth fractions yielded the more polar compounds.Gas chromatographic (GC) analyses were carried out on a HR 8000 Fisons gas chromatograph (FID, cold on-column injection) equipped with a 30-m (0.32 mmÆ) glass capillary column (DB5HT). Helium was used as a carrier gas (column head pressure 110 kPa). Oven temperature programming conditions were from 35°C to 300°C at 15°C min-1 followed by an isothermal period of 15 min at 300°C and from 300°C to 330°C at 15°C min-1 followed by an isothermal period of 10 min at 330°C. Alkenones were identified by comparing retention times with those of synthetic standards. Peak areas were converted to the UK'37 index, and SSTs calculated according to a global coretop calibration (Müller et al., 1998): SST (°C) = (C37:2 / (C37:2 + C37:3) - 0.044) / 0.033. Alkenone concentrations were determined from GC analyses of the ketone fractions based on peak responses relative to 5alpha-cholestane as an external standard. Results of triplicate analyses (n = 4) indicate that the standard deviation of temperatures calculated from UK'37 is 0.2°C (range 0°C–0.3°C).Age for each sample was calculated by linear interpolation between stratigraphic fix points (sapropel bases) as described in Emeis et al., (2003).Because of limited sample material, we were not able to perform all analyses on all sapropels.