All carbon and silica analyses were conducted at Boise State University (BSU; Idaho, USA). Opal procedures followed Olivarez Lyle and Lyle (2002), who modified the methods of Mortlock and Froelich (1989) to address the problem of dissolution-resistant Eocene radiolarians. Biogenic opal is a hydrated form of silica sometimes averaging up to 10%–15% water. No corrections were made for water in the crystalline structure, and for this reason, data are reported as biogenic silica without any assumptions of water content. Carbon procedures followed Lyle et al. (2000). A second acidification was applied to samples enriched in calcium carbonate or dolomite. Before analysis, all samples were first freeze-dried, ground with mortar and pestle, and stored in glass vials placed in desiccant to prevent equilibration with ambient moisture.
To assess precision and accuracy, both silica and carbon analyses are performed with in-house standards included in each run. Two opal standards were included with silica analyses: a composite standard containing sediments from ODP Cores 199-1219A-19H through 22H, prepared at BSU, and a standard supplied by Steve Hovan at Indiana University of Pennsylvania (USA), a radiolarian ooze composite containing samples from several Hole 1219A core catchers. Analyses of Hovan's standard averaged 65.8 ± 4.4 wt% silica (N = 45) (Table T1). The BSU composite standard averaged 56.3 ± 1.7 wt% silica (N = 52). Carbon analysis is conducted with a "Midway" standard from the Multitracer site in the northeast Pacific, W8709A-5BC: 5–20 cm. The Midway standard averaged 2.64 ± 0.02 wt% total carbon (N = 523). Organic carbon (C-org) in the same standard averaged 0.85 ± 0.01 wt% (N = 570).