Samples were taken aboard the JOIDES Resolution and in the ODP core repository at Bremen, Germany. In general, average sample spacing was 1.5 m for high-resolution sediments and 45 or 90 cm for sediments deposited at lower sedimentation rates. Additional samples were studied when needed. Raw samples were prepared as smear slides when calcium carbonate contents were below ~50 wt% (Sites 1091, 1093, and 1094). The other sediment samples containing carbonate above this value (Sites 1089, 1090, and 1092) were chemically treated after the method described by (Gersonde and Zielinski, 2000) in order to remove calcareous components. After the washing process, strewn slides were prepared for diatom analysis. Light microscopic studies were achieved using a Zeiss Axioskop with a magnification of 1000x (oil-immersion objective). At least three transects per slide were studied in sediments consisting of pure diatom ooze; double that number or more was necessary in samples with lower diatom content. The preservation of diatom valves was estimated qualitatively as good (no signs of dissolution, lightly silicified species present), moderate (lightly silicified species present but with alterations, diatom valves are slightly affected by dissolution) or poor (no fine silicified forms present, fragmented valves, dominance of robust silicified species). Estimation of species relative abundance follows the ODP style (Shipboard Scientific Party, 1999a) as follows:
The determination of a biostratigraphic datum was defined by the midpoint between two adjacent samples. Trace occurrences of species were not considered in defining datums. For all depth information of the studied samples the meters composite depth (mcd) scale was used (see individual site chapters in Gersonde, Hodell, Blum, et al., 1999).
The presented range charts were also archived in the PANGAEA information system at the Alfred Wegener Institute for Polar and Marine Research, Bremerhaven (AWI) (www.pangaea.de).