Methods

Stable isotope analyses were conducted at the University of Florida (UF) and Scripps Institution of Oceanography (SIO). At UF, foraminifer tests were soaked in ~15% H₂O₂ for 30 min to remove organic matter. The tests were then rinsed with methanol and sonically cleaned to remove fine-grained particles. The methanol was removed with a syringe, and samples were dried in an oven at 50°C for 24 hr. The foraminifer calcite was loaded into individual reaction vessels, and each sample was reacted with three drops of H₃PO₄ (specific gravity = 1.92) using a Finnigan MAT Kiel III carbonate preparation device. Isotope ratios were measured online by a Finnigan MAT 252 mass spectrometer. Analytical precision was estimated by measuring eight standards (NBS-19) with each carousel containing 38 samples.

At SIO, foraminifer specimens were cleaned ultrasonically in distilled water until visually free of adhering particles and roasted under vacuum at 375°C for 1 hr to remove organic matter. Samples were then reacted in a common acid bath of H₃PO₄ using a Carousel-48 automatic carbonate preparation device. Isotope ratios of purified CO₂ gas were measured online with a Finnigan MAT 252 mass spectrometer. All isotope results are reported in standard delta notation relative to VPDB (Vienna Peedee belemnite) (Coplen, 1996).

Analytical results from SIO and UF were intercalibrated by measuring a common set of benthic foraminiferal samples from Site 1089. For ¹⁸O, UF results were systematically greater than those measured at SIO by ~0.4, whereas no offset was measured for ¹³C. The intralaboratory ¹⁸O offset is large relative to the precision measured by repeated analysis of NBS-19. At UF, the long-term mean value of NBS-19 is -2.13 ± 0.06 for ¹⁸O and 1.93 ± 0.03 for ¹³C, slightly greater than the literature value of -2.19 for ¹⁸O but not large enough to account for the 0.4 offset. For several reasons, we don't believe the offset is due to differential treatment of samples for organic matter (chemical vs. roasting). Mead et al. (1993) found no significant differences between standards that were treated by both methods. Furthermore, Cibicidoides samples used for the intracalibration experiment were first cleaned at UF using the peroxide method and then sent to Scripps for analysis. We observe a similar offset in ¹⁸O for samples analyzed at UF using the PRISM mass spectrometer and ISOCARB (common acid bath) preparation device compared to the Finnigan MAT252/Kiel III system. The difference appears to be related to using a "common acid bath" as opposed to a "single acid aliquot" device and may be dependent upon the reaction rate and how quickly the CO₂ is removed from the reaction vessel (equilibration with the H₂O and H₃PO₄). For consistency of isotopic results measured internally at UF and at SIO, we subtracted 0.4 from the UF ¹⁸O data that were produced using the MAT252/Kiel III system.

Weight percent CaCO₃ was determined for Sites 1088, 1089, and 1090 at UF by coulometric titration with a UIC Inc. model 5240 total inorganic carbon (TIC) preparation system and model 5011 coulometer (Englemann et al., 1985). Analytical precision is estimated to be ±1% based on repeated analysis of reagent grade (100%) CaCO₃.