Diatom slides were prepared as simple smear slides by suspending sediment in less than 5 ml deionized water, transferring a drop of this solution to a cover slip, and allowing the cover slip to air dry (Sancetta, 1982). Cover slips were mounted on cleaned microscope slides using hyrax in toulene (refractive index: 1.7135). At least 300 diatom valves were identified in at least three random transects across each slide using a light microscope at magnifications from 1000x to 1250x (see Armand et al., 2005; Sancetta, 1979; Sancetta and Silvestri, 1986; Scherer, 1994). Partial valves were counted according to the methods of Schrader and Gersonde (1978). All diatoms were identified to the species level when possible. On the table attached, diatom counts are transformed into relative percent abundances.

Grain size was measured on samples collected in 2006 and 2007 using the Coulter LS200 laser diffraction particle analyzer. Pretreatment of sediments to remove organics, carbonates, biogenic opal, and to disperse clay particles follows Sakamoto (2005); however, samples were not centrifuged between chemical treatments to avoid losing very fine sediments. Instead chemical volumes were adjusted so that each treatment was neutralized before the next chemical was added. This method of grain size analysis returns the weight percent of particles in each of 92 bins ranging from 0.375 to 1822 μm. In addition, we report the median (Dx50) grain size and % clay, % silt, and % sand in each sample.