Two suites of samples were selected, 52 (5-cm3) samples for geochemical analyses and 35 (20-cm3) samples for benthic foraminifer studies and stable isotope analyses. The 5-cm3 samples were chosen to sample the greatest range of variation in the carbonate cycles. The 20-cm3 samples are spaced at approximately equal time intervals, based on the shipboard biostratigraphy (Shipboard Scientific Party, 2000), and are from carbonate-rich levels.
Calcium carbonate was measured on 52 samples at the University of Rhode Island. Samples were freeze-dried and crushed. Analyses were made with a coulometer on 1-mg aliquots. Two replicates were analyzed per sample, and a third replicate was analyzed when these differed substantially (Table T1). Reported values are the average of the two closest analyses. In addition, estimates of carbonate are based on measurements of reflectance (550 nm) made every 4 cm according to the following logarithmic relationship (Fig. F2) (R2 = 0.89):
Organic carbon analyses of 19 samples were made at the University of Maine's Darling Marine Center (Table T1). Aliquots of 30 mg were finely powdered and subjected to HCl fumes to remove carbonate. Organic carbon and nitrogen were measured with a CHN analyzer. A single replicate analysis differs from the first analysis by 0.013 wt% (0.070 and 0.083 wt%). Because of the high carbonate and low OC concentrations, the method of Verardo et al. (1990), which involves dissolving carbonate with sulfurous acid, was attempted on an additional suite of samples. This technique failed when most of the aluminum boats developed holes before the carbonate was entirely dissolved.
Grain size, opal, and major and trace element analyses were conducted on five samples with carbonate ranging from 23 to 79 wt%. Carbonate was first removed with 25% acetic acid (URI/Hovan protocol) on a magnetic stirring hot plate for 2 hr. Samples were centrifuged and rinsed with deionized water three times. Grain sizes were then determined on the decarbonated samples by the X-ray sedigraph method (Table T2). A 0.5% Calgon solution was used to disperse the particles. Afterward, the samples were rinsed with deionized water and centrifuged. Samples were dried at 60°C and weighed. Opal was dissolved with Na2CO3 (URI/Hovan protocol). Dried samples were reweighed to determine opal concentrations (Table T2). Major and trace elements were measured on these five samples plus an additional five untreated samples at Washington University by X-ray fluorescence (Table T2).
Five samples with carbonate ranging from 14 to 74 wt% were prepared for X-ray diffraction analyses (Table T3). Carbonate was first removed with weak acetic acid on a magnetic stirring hot plate (Moore and Reynolds, 1997). Samples were rinsed with deionized water and centrifuged. Before size separation, Na phosphate was added to disperse the particles. To separate the clay (<2 µm) fraction, the samples were centrifuged three times until the supernatant cleared. Silt and clay-sized fractions were dried at 60°C. Thin-filter transfers were prepared from the silt fraction. This method did not work for the clay fraction; therefore, smear slides were made. Selective glycolation was used to identify smectite (d-spacing increases from ~15 to >17 Å).
Stable isotope ratios were measured in triplicate on six samples of bulk sediment with carbonate ranging from 29 to 65 wt% (Table T4). Samples were dissolved in phosphoric acid at 90°C and were analyzed at the University of Maine on a VG/Fisons Prism Series II mass spectrometer. Measured ratios were normalized to the Peedee belemnite (PDB) standard according to values of the working standard NBS-20.
The initial 35 plus an additional 23 Oligocene through lower Miocene samples (20 cm3) were prepared for benthic foraminifer studies and stable isotope measurements on the foraminifers. The samples were soaked in a Calgon solution then washed through a 63-µm sieve and dried in an oven at 60°C. Benthic species were separated from the >125-µm fraction in 35 samples and identified by Mimi Katz at Rutgers University (Table T5). The number of specimens retrieved per sample was small, in part because the mostly lithified samples failed to disaggregate despite the Calgon treatment and long duration of soaking. Twenty-three samples yielded sufficient quantities of the benthic foraminifer genera Cibicidoides spp. for stable isotope analyses (Table T6). The samples were sonicated to remove matrix material.
Biostratigraphic and paleomagnetic datums used to construct the depth-age plot are shown in Table T7.