METHODS

Samples were oven-dried at 60°C and washed through a 63-µm sieve. Well-preserved specimens were chosen from the coarse-fraction residues and cleaned with analytical grade ethanol (99.7%) in an ultrasonic bath. They were dried again in an oven at 60°C and reacted with orthophosphoric acid in an automatic carbonate device (Kiel III) at 70°C to generate CO₂. The escaping CO₂ was then transferred to a Finnigan MAT252 mass spectrometer for measurement of stable isotopes at the Laboratory of Marine Geology in Tongji University, People's Republic of China. The epibenthic species Cibicidoides wuellerstorfi was preferred for isotope analysis because various studies have demonstrated that the ¹³C value of this species is approximately equal to bottom water ¹³C and is therefore a good proxy for reconstructions of deepwater ventilation (McCorkle and Keigwin, 1994). In intervals where C. wuellerstorfi was absent in the sediments, other benthic foraminifers were analyzed (Table T5) to obtain a complete oxygen isotope record. Globigerinoides ruber was mainly used for the upper 165 mcd, and Globigerinoides sacculifer was used for the lower part. Because of isotopic differences between species, adjustments were made to account for taxon-dependent deviations from seawater isotopic equilibrium using the adjustment factors given in Table T5, according to Shackleton et al. (1995), Shackleton and Hall (1983, 1990), and Z. Jian et al. (2003, unpubl. data). When several measurements were made on different planktonic or benthic foraminifers from the same sample, the adjusted average value was used (Tables T1, T2, T3, T4). Precision was regularly checked with the Chinese national carbonate standard GBW04405. Conversion to the international Peedee belemnite (PDB) scale was performed using NBS19 and NBS18 standards: ¹³C = 0.57, ¹⁸O = –8.49); the standard deviation was 0.07 for ¹⁸O and 0.04 for ¹³C for measurements carried out during the years 1999–2000.