METHODS

Sediment samples collected from pore-water squeeze cakes and selected carbonate-rich horizons were washed with pH-10 distilled water (buffered with NH₄OH) to remove salts, freeze-dried, and powdered with an agate mortar and pestle for X-ray diffraction (XRD), percent CaCO₃, and stable isotope (¹⁸O and ¹³C) measurements. Bulk mineralogy was measured on randomly oriented samples by XRD, using monochromatic CuK radiation on a Phillips Norelco 1720 at the University of North Carolina at Chapel Hill (UNC-CH). Scans were run from 5º to 70º 2, with a scan rate of 1.0º 2/min at 40kV/30mA. Each sample was spiked with an internal standard (20% potassium chloride [KCl]) to estimate the abundance of calcite, siderite, and dolomite. Weight percent of each mineral was determined by comparing the ratio of the reference peak (104) area (at 3.03Å, 2.89Å, and 2.79Å for calcite, dolomite, and siderite, respectively) to the reference peak (200) area of KCl (3.15Å) and comparing that ratio to calibration curves made for each mineral (Klug and Alexander, 1974). Total carbonate was measured by acid digestion and titration of evolved CO₂ using a Coulometrics Model 5011 CO₂ Coulometer at UNC-CH. To insure that all carbonate phases were digested, samples were acidified with 4-N HCL at 95ºC for 10 min. CO₂-yield tests confirmed that this procedure removed all carbonate phases (aragonite, calcite, dolomite, and siderite). Selected subsamples were examined with a Leica S440 Scanning Electron Microscope (SEM) and Energy Dispersive Spectrometer (EDS) at UNC-CH for qualitative evaluation of crystal habit and chemistry. Thin sections of epoxy-impregnated sediments were made and examined petrographically to evaluate the influence of diagenetic changes on sediment fabric.

Stable carbon and oxygen isotope measurements of the powdered samples involved digestion of carbonates under vacuum by selective extraction (Al-Aasm et al., 1990) in 100% phosphoric acid. All samples were placed in borosilicate vials and roasted under vacuum at 325ºC for 1 hr to remove volatile organics. The samples were then placed in side-arm reaction vessels sitting in a heated water bath. Liberated CO₂ was cryogenically separated and then measured on a Finnigan Mat 251 Isotope Ratio Mass Spectrometer at North Carolina State University (NCSU), Raleigh, North Carolina, and on a VG-903 Isogas Source Stable Isotope Ratio Mass Spectrometer at Geochron Laboratories (Geochron) in Cambridge, Massachusetts. Evolved CO₂ was cryogenically separated at ~-130ºC to prevent contamination with SO₂ produced during digestion of sulfide-bearing sediments (Des Marais, 1978).

The ¹³C and ¹⁸O values of calcite were measured at both NCSU and at Geochron Laboratories. At NCSU, calcite-digestion times were 10 min at 75ºC. At Geochron, calcite digestion took place for 20 min at 50ºC. Duplicate samples measured at both labs indicate that the data are internally consistent (± 0.2 for ¹³C and ± 0.4 for ¹⁸O). Although there may be some dissolution of dolomite or siderite within the first 10 (at 75ºC) or 20 (at 50ºC) min of reaction with phosphoric acid, the extracted CO₂ is primarily that of calcite and should not affect the measurements appreciably.

The ¹³C and ¹⁸O values of dolomite and siderite were measured at Geochron Laboratories. After CO₂ generated by dissolution of calcite (20 min at 50ºC) was withdrawn, the digestion vessels were again sealed and returned to the 50ºC water bath, where digestion of the dolomite or siderite fraction of the sediment could proceed. A series of CO₂-yield/digestion-time experiments were run at Geochron (N.M. Rodriguez, unpubl. data) to estimate the amount of time required for dolomite and siderite digestion to go to completion. Complete digestion of dolomite and siderite in these experiments took place in ~12 and ~48 hr, respectively. Thus, samples containing primarily dolomite underwent continued digestion from 20 min to 24 hr. Samples containing primarily siderite underwent continued digestion from 20 min to 72 hr. Digestion residues from selected samples, which initially contained either siderite or dolomite, were evaluated by SEM to confirm that complete carbonate digestion occurred.

All stable isotopic measurements are reported with the notation, relative to Peedee belemnite (PDB). We have chosen not to apply specific phosphoric acid fractionation factors that account for the difference in fractionation between calcite and dolomite, or calcite and siderite (Table 2). Although specific phosphoric acid-liberated CO₂ fractionation factors exist (at a wide range of temperatures) for both stoichiometric siderite (Carothers et al., 1988) and dolomite (Rosenbaum and Sheppard, 1986), the scientific community currently lacks consensus on how to apply them to compositionally varied samples. Fractionation factors for "impure" dolomites and siderite are uncertain due to the relationship between the chemical composition of a carbonate mineral and its phosphoric acid-liberated CO₂ fractionation factor (Land, 1980; Rosenbaum and Sheppard, 1986).

Pore-water samples were obtained by routine shipboard squeezing of whole-round sediment samples shortly after core recovery (Manheim and Sayles, 1974). Interstitial Ca²⁺ and Mg²⁺ concentrations, Sr²⁺ concentrations, and total alkalinity were measured by ion chromatography, atomic absorption spectrometry, and titration respectively (Gieskes et al., 1991; Paull, Matsumoto, Wallace, et al., 1996). Aliquots of pore water were flame sealed in glass ampoules and stored for shore-based analysis. Pore-water samples were analyzed for ¹³C of DIC by acidification with 33% H₃PO₄ to liberate CO₂, which was cryogenically separated and analyzed on a Delta E mass spectrometer at NCSU (Borowski et al., Chap. 9, this volume; Paull et al., Chap. 7, this volume).

Age estimates for Sites 994, 995, and 997 are based upon detailed nannofossil datums for each site (Paull, Matsumoto, Wallace, et al., 1996) and reported relative to the late Neogene timescale of Shackleton et al. (1995). Host sediment ages are then estimated by assuming a constant sedimentation rate between nannofossil datums.