Methods, Magnesium-calcium ratios 

The magnesium-calcium ratio was measured on two samples of 40 N. pachyderma (s.) from the 150-212 mm size fraction. We performed mechanical, oxidative and reductive cleaning following the procedure of [Boyle and Keigwin, 1985] with revisions by [Rosenthal et al., 1995] and [Boyle and Rosenthal, 1996], and where ultrasonication during the final acid leach was skipped due to the fragility of the samples. Samples were measured on a high-resolution sector-field Finnigan Element 2 inductively-coupled plasma mass spectrometer (ICP-MS). 

Results, Foraminifer Mg/Ca 

The Mg-Ca ratio of N. pachyderma (s.) from 51JPC are in the same samples as in the mass mixing model described in the previous section. At 1104 cm, the Mg/Ca is 0.7 mmol/mol, which is within the typical range for Quaternary planktonic foraminifera [Rosenthal et al., 1999]. At 864 cm, a sample with anomalously low d13C, the Mg/Ca is 72 mmol/mol. From the same mass-mixing model used to estimate d18O and d13C of authigenic carbonates, we estimate the Mg/Ca of the authigenic carbonates is 132 mmol/mol. The estimate of Mg/Ca of overgrowth could be affected by the cleaning procedure for ICP-MS analysis (the same chemical cleaning procedure is not applied to samples before analyzing for d13C and d18O). Fifty percent and 95% of the initial mass of the samples at 864 cm and 1104 cm were lost, respectively, during cleaning. For this procedure, loss of 50% of the mass of a sample is considered typical. If the authigenic carbonate is more susceptible to loss, then authigenic Mg/Ca is underestimated (it could be as low as 72 mmol/mol). If the foraminifer calcite is more suceptible to loss, then authigenic Mg/Ca is overestimated (it could be as high as 837 mmol/mol).