Samples were prepared in the typical manner (Sanfilippo et al., 1985; Boltovskoy, 1999) by treating the sediments with a solution of 10% H2O2 neutralized with sodium pyrophosphate to a pH of 7. Successive treatments were conducted until the radiolarian skeletons were clean. The carbonate fraction was removed with HCl. Sediments were sieved over a 45-m screen, and strewn slides were mounted with Canada balsam made with the >45-m fraction. Analyses of the slides were made using a Zeiss Photomicroscope I at 100. Selected species were counted along 1-3 transects until a total of 300 specimens were counted. Qualitative estimates of total radiolarian abundance and preservation were made using the following criteria:

A = >1000 skeletons per slide.
C = >500-1000 skeletons per slide.
F = >100-500 skeletons per slide.
R = >10-100 skeletons per slide.
T = 1-10 skeletons per slide.
B = barren of skeletons.
G = no sign of dissolution and only minor fragmentation.
M = dissolution and obvious fragmentation.
P = high degree of dissolution and very few skeletons intact.

Sixty-two species were counted, and species with well-constrained habitats were used to make environmental groups that reflect specific water masses. Species were assigned to a water mass based on published occurrences, which are cited in the species list. These water masses include the warm-water mass carried by the Angola Current (warm fauna), transition water in the Benguela Current (transition fauna), and intermediate waters upwelled along the continental margin by oceanic upwelling (intermediate fauna). A more detailed description of the hydrographic setting can be found in Chapter 1 of the Leg 175 Initial Reports volume (Wefer, Berger, Richter, et al., 1998). The majority of data cited are from sediment trap, plankton tow, and surface-sediment studies with concurrently collected hydrographic data, so the data are suitable for determining which water mass the species occur in.