All samples used for analysis consist of 15 cm3 of sediment taken at a general sampling frequency of 20 cm within each core. However, additional samples were taken in between to increase the sampling frequency at significant points (e.g., transitions between glacial periods [glacials] and interglacials). A total of 454 samples have been analyzed from Hole 1003A (Sections 166-1003A-1H-01 through 9H-CC) and a total of 293 samples were taken from Hole 1006A (Sections 166-1006A-1H-01 through 7H-7). A flow diagram to in Figure 2 shows the laboratory procedures.
The samples (5 cm3) were oven dried at 45°C, weighed, and then washed through a 63-µm stainless steel sieve. The coarse fraction was then further divided into subfractions (for details, see "Sediment Flux Analysis"). The 250- to 500-µm fraction was used to pick tests of the planktonic foraminiferal species Globigerinoides ruber. Care was taken to select only complete and undamaged tests to reduce the risk of picking the incorrect species. The tests were then soaked in methyl alcohol for several minutes and cleaned in an ultrasonic bath to remove adherent contaminants. Following ultrasonic cleaning, excess methyl alcohol was drawn off with tissue paper, and any residual alcohol was allowed to evaporate. The foraminiferal sample weights were typically <0.1 mg and composed of 5-8 planktonic specimens.
After cleaning, the foraminiferal tests were reacted in orthophosphoric acid (specific gravity = 1.9) at 90°C, and the resulting CO2 gas was analyzed using a Precision Isotope Ratio Mass Spectrometer (PRISM) located at the Geology and Geophysics Department of the University of Edinburgh. The samples were analyzed using an internal laboratory standard (marble reference SM1), and the values obtained converted to a PeeDee Belemnite (PDB) standard. Precision for the oxygen isotope analysis was 0.085° (standard deviation for 100 analyses of an "in-house" standard carbonate [SM1] conducted over several months) using SM1 sample weights of 0.05-0.1 mg.
The analysis was carried out on ~0.5 g of fine-fraction (<63 µm) sediment, and follows that described in various texts including Milliman (1974). Each sample was oven dried at 60°C before being ground for exactly 4 min in an agate mortar (with analytical grade acetone). This provides a more homogenized sediment, and thus reduces errors that might seriously affect the quality of the results. The grinding time has been documented as that required to provide an optimal peak intensity for X-ray diffraction (XRD) analysis (Milliman, 1974).The individual samples were then drawn up in a pipette, dispersed on a glass slide (2-cm diameter), smeared out to produce a sediment slurry of standard thickness and diameter, and dried at room temperature before being subjected to X-rays.
The samples were analyzed
at the Geology and Geophysics Department of the University of Edinburgh using a
Philips PW 1011/1050 automatic power diffractometer (PW 1808 sample charger) at
40 kV and 50 mA, through a scan from 25° to 45° 2-
,
at a low scan speed of 0.040° per second for optimal resolution. The peak
intensities were measured for aragonite, HMC, LMC, and dolomite, and the
presence of quartz was also recorded to provide a representative for terrigenous
components within the sediment.
The program MacDiff 3.1.5 was then used to determine the amount of calcite and aragonite within the sediment by measurement of the total peak intensity, which is directly related to the peak area (Milliman, 1974). Calcite was composed of HMC and LMC, one forming the major peak and the other a minor peak on the shoulder of the major peak. The intensity of the major peak was measured and subtracted from the total calcite intensity to provide the intensity of the second (minor) peak. The relative weight percentage of calcite and aragonite were then calibrated using an "in-house" calibration curve formed from a series of measurements carried out on two component standards. These consisted of a laboratory-produced synthetic calcite and a pure aragonite from a Red Sea coral. The relative weight percentage of dolomite was then calculated using the linear correlation by Milliman (1974). The quartz data was left in peak-height intensities.
Thermal ionization U-Th analyses have been applied to sediment from four horizons in Site 1006 and six from Site 1003 (Henderson et al., Chap 3, this volume). These horizons were selected from peaks in the interglacial aragonite stratigraphies to assign each package of aragonite-rich sediment to its correct highstand period. In general, U-Th dating of marine sediment is difficult because they contain appreciable amounts of initial 230Th bound in detrital clay minerals and scavenged from seawater. This initial 230Th gives the sediment a nonzero initial age and must either be removed or corrected for in order to arrive at a true age. The aragonite-rich sediments of the Bahamas are unusually high in U and low in initial Th so this problem is less acute than elsewhere (Slowey et al., 1996), but it must still be addressed. To reduce initial 230Th, analyses were performed on 63- to 250-µm pure-aragonite sediment separates. These procedures quantitatively remove the detrital material and reduce the concentration of scavenged Th. Remaining scavenged 230Th is corrected for by using the measured 232Th/230Th ratio and assuming a 232Th/230Th ratio for seawater. Full details of the sieving and heavy liquid protocols used for the separation and of blanks, chemical separation, and mass spectrometry can be found in Henderson et al. (Chap 3, this volume).
This procedure (Wolf and Thiede, 1991) consisted of freeze-drying 5-cm3 bulk sediment samples in a vacuum container. The temperature was maintained at -40°C in liquid nitrogen. Dehydrated sample weights were taken using a digital scale (precision to 0.0001 g). Samples were then left to form a deflocculated suspension in distilled water prior to sieving.
The 5-cm3 suspensions were washed through a 63-µm metal sieve (U.S. standard) with a spray of distilled water. The coarse fraction (>63 µm) was drained and rinsed into 10-cm3 containers for drying in fan-assisted ovens before weighing.
The fine fraction (<63 µm) was collected in large (5 L) beakers and left to settle. After settling, the water was siphoned off and the fine fraction was dried in fan-assisted ovens before weighing.
The dried, weighed coarse fraction (>63 µm) was subsequently split (using a hand-held sieve set) into five subfractions: 63-125 µm, 125-250 µm, 250-500 µm, 500-1000 µm, and >1000 µm (U.S. Standard) (Fig. 3). Subfractions were stored in preweighed glass vials and their weights were measured.
