We investigated samples from Site 1077. Sampling was mainly performed on Hole 1077A, although for some intervals additional samples were collected from Holes 1077B and 1077C. Meters below seafloor were transformed to meters composite depth (mcd) according to Shipboard Scientific Party (1998c).
High-resolution (20-cm sampling intervals) stable isotope analysis was performed on handpicked samples of the planktonic foraminifer Globigerinoides ruber (size >150 µm). A Finnigan MAT 251 micromass spectrometer equipped with a Kiel automated carbonate preparation device was used.
Samples (5 cm3) were also collected for siliceous microfossil analysis. Sampling intervals varied from every 45 cm for the upper 44 mcd (representing an average resolution of ~2 k.y.) to every 110-120 cm downcore to 70 mcd; resolution for the lower part ranges from ~5 to 10 k.y. Samples were freeze-dried, and 0.4 g of dry sediment for each sample was treated with hydrochloric acid and hydrogen peroxide to dissolve carbonates and organic matter following the method of Schrader and Gersonde (1978); sodium pyrophosphate was added to remove clay-sized particles in suspension. Acid and salt remains were removed by repeated steps of rising with distilled water and settling. Preparation of slides for qualitative and quantitative analyses was performed according to Lange et al. (1994).
Identification and counting of taxa was done with a Zeiss-axioscope with phase-contrast illumination at 1000× magnification. Because of the overwhelming dominance of a few diatom taxa, the counting procedure was divided into two steps: (1) for the most abundant diatom species (e.g., Thalassionema nitzschioides var. nitzschioides, Cyclotella litoralis, and Chaetoceros spp.) just one transect across the slide was quantified; (2) for the rest of the diatom species, radiolarians, silicoflagellates, ebridians, phytoliths, chrysophycean cysts, and the siliceous skeleton of the dinoflagellate Actiniscus pentasterias, a fraction of the slide (one-third, one-fifth, or one-tenth, depending on abundance) was counted. Sponge spicules were not included in our analysis. Definition of counting units followed that of Schrader and Gersonde (1978). Diatoms and silicoflagellates were identified to the lowest taxonomic level possible, whereas all other siliceous microfossils were counted as groups. Abundances of taxa and/or microfossil groups were calculated as concentrations per gram and as accumulation rates (AR). Relative abundances of individual species or group of species were calculated as percent of total assemblage. Accumulation rates for the different microfossils studied was calculated according to van Andel et al. (1975), as follows:
All data are available through the PANGAEA server (www.pangaea.de/Projects/SFB261/EUliana_et_al_2000/).
We used the PhFD Index introduced by Jansen et al. (1989) as an index of paleoaridity over equatorial Africa. This index is a ratio between the concentration of phytoliths and freshwater diatoms (Ph / [Ph + FD]). High values are related to arid conditions over the continent (higher abundance of phytoliths), whereas low values (high freshwater diatom concentrations) may reflect stronger river influence in the Congo area.
Biogenic opal content was measured on Site 1075 samples by Hui-Ling Lin at the National Sun Yat-Sen University of Taiwan. Opal contents were determined by the basic leaching method of Mortlock and Froelich (1989), modified by using different acid and base reagents (Lange et al., 1999; Lin et al., in press). Values reported here as opal contents are calculated (based on Mortlock and Froelich ) as follows:
Total carbon and organic carbon (TOC) concentrations were measured on Site 1075 samples using a LECO CS-244 carbon/sulfur analyzer. Analytical details are given in Lin et al. (in press).