All samples were freeze-dried to remove water and divided into two portions. The first ~2-cm3 portion was crushed and used to determine bulk carbonate content. The other ~5-cm3 portion was weighed, wet-sieved at 38 Ám, and used to determine coarse size fraction, benthic foraminiferal abundance, and foraminiferal fragmentation. All four of these parameters have been used in previous investigations to make inferences about carbonate dissolution (e.g., Le and Shackleton, 1992; LaMontagne et al., 1996). In studies of Quaternary foraminiferal assemblages, the >63-Ám size fraction has generally been used for dissolution studies because it conveniently coincides with the sand size fraction (e.g., Broecker and Clark, 1999, 2001). However, 38 Ám was chosen to delineate coarse fraction in this study because Paleocene foraminifers are commonly very small.
Approximately 100 mg of dried, crushed sample was analyzed for carbonate content according to the "Karbonate-Bomb" method (Mueller and Gastner, 1971). Sample aliquots were reacted with ~3 mL of 10% HCl to produce CO2 gas in a closed vessel attached to a water-filled cylinder marked with 0.1-mL graduations. Carbonate content was determined by comparing the volume of CO2 gas generated to that produced from known masses of laboratory-grade CaCO3. Each sample was analyzed twice to ensure analytical precision. These replicate analyses were consistently within 2 wt%, and reported carbonate contents are the average of both measurements. A sample of Paleogene siliceous limestone (JCU Sample MS14) with a known carbonate content of 72 wt% was also analyzed twice to evaluate accuracy and precision. The measured carbonate content of MS14 was 71 wt% ▒ 2 wt%.
After wet sieving, the coarse (>38 Ám) component of the second sample portion was dried and weighed. This mass divided by bulk sample mass rendered the coarse (>38 Ám) size fraction. Carbonate dissolution within the lysocline commonly leads to fragmentation of foraminifer tests, which decreases the coarse size fraction of bulk sediment (Berger et al., 1982; Broecker and Clark, 1999). However, we were unsure if and how this proxy would register dissolution in Paleogene sediment from Shatsky Rise because nannofossils are the dominant contributors (>90%) to the sediment (see "Visual Core Descriptions" in Bralower, Premoli Silva, Malone, et al., 2002). Changes in the relative abundances of nannofossils and foraminifers could certainly give variations in grain size unrelated to dissolution.
A microsplitter was then used to separate at least 300 foraminifers and fragments from the coarse fraction onto a marked slide. Care was taken to distribute material evenly over the microsplitter to avoid sampling bias. The numbers of fragments, benthic foraminifers, and planktonic foraminifers were recorded. Where the sample was highly fragmented and whole foraminifers rare, at least 50 counts of whole foraminifers were made.
The relative abundance of benthic foraminifers (BENTH) is expressed as a ratio of benthic foraminifers to the sum of benthic and planktonic foraminifers. Benthic foraminifers are less susceptible to dissolution than planktonic foraminifers because the latter typically have porous chamber walls designed to maintain buoyancy in surface waters. Thus, the relative abundance of benthic foraminifers may serve as an index for carbonate dissolution at deepwater sites (Schlanger and Douglas, 1973; Thunell, 1976).
A foraminiferal fragment (F) is defined here and elsewhere (Berger et al., 1982) as a test portion less than two-thirds of its original size. The fragmentation index (FRAG) for each sample was calculated according to the following equation (Williams et al., 1985; Malmgren, 1987):
The number of fragments is divided by 8 because, on average, one foraminifer breaks into this number of fragments and it is the proportion of fragmented foraminifers, rather than the number of fragments themselves, that has a near-linear relationship with dissolution (Le and Shackleton, 1992). Benthic foraminifers are not included in the whole foraminifer count as their tests are more resistant to dissolution than planktonic foraminifers.