Because of the occurrence of radiolarians throughout the upper Miocene to Pleistocene and good records of magnetic polarity sequences, Site 1082 would be a key station for constructing upper Cenozoic radiolarian biozonations and dating their bioevents. Table T2 summarizes radiolarian events that seem to be of potential stratigraphic significance in the vicinity of Site 1082, probably in the Benguela Current region. Table T1 and the previously published occurrence chart (Shipboard Scientific Party, 1998, table 5), which was produced by shipboard investigations on the basis of the core-catcher samples, are both used for identification of the bioevents listed in Table T2. Numerical ages of the samples and these bioevents estimated based on the age-depth model (Fig. F2) are also shown in Tables T1 and T2.

The number of specimens searched in an assemblage is an important factor in stratigraphic interpretations (Sanfilippo et al., 1985). It must be noted that the number of specimens searched in a sample is about 500 in the case of the present study and several thousand in the case of most of the core-catcher samples that were searched on board. Based on the present census assessment of radiolarian assemblages, abundance change of taxa can be seen, and such bioevents as LCOs, RIs, and SAs characterized by abundance changes can be defined. An absence recorded after searching through thousands of specimens has more meaning than one recorded after searching through only a few hundred. Thus, the higher the number of specimens searched, the higher the accuracy of absence and the reliability of definition of first and last occurrences of a taxon.

As seen in descriptions of radiolarian biostratigraphy by Wefer, Berger, Richter, et al. (1998), any one of the previously proposed zonal frameworks alone (e.g., Johnson et al., 1989; Caulet, 1991; Lazarus, 1992; Moore, 1995; Motoyama, 1996) is hard to fully apply to radiolarian assemblages at Leg 175 sites because of the absence or scarcity of known zonal marker species. Pisias and Moore (1978) applied the North Pacific biozones of Kling (1973) to the upper Miocene and Pliocene at DSDP Leg 40 Site 362, which is near Site 1082. Although occurrences of Axoprunum angelinum, Stichocorys peregrina, Lamprocyrtis heteroporos, and Spongurus pylomaticus are recognized in Hole 1082A, the other key species, including Eucyrtidium matuyamai, Sphaeropyle langii, or Dictyophimus robustus are absent from Hole 1082A, demonstrating that part of the North Pacific radiolarian zonation could not be used for the biostratigraphic study at Hole 1082A. In this study, radiolarian zones are given in the tropical zonation of Moore (1995) wherever possible, and the North Pacific zonation of Motoyama (1996) is adopted in part (Table T1). Comments are made below on radiolarian events, some of which are expected to be useful for future biostratigraphic studies in the southeastern Atlantic Ocean.

The LO of A. angelinum (= Stylatractus universus) is known to be a globally synchronous event (Hays and Shackleton, 1976). This event is one of the widely traceable datum levels in the southeastern Atlantic Ocean, as it is recognized in the cores RC13-205 and RC13-229 (Morley and Shackleton, 1978) and Holes 1075A, 1076A, 1077A, 1080A, 1081A, 1083A, 1084A, 1085A, and 1087A (Wefer, Berger, Richter, et al., 1998) as well as in Hole 1082A; thus, it extends across the Angola-Benguela Current System.

The LO of C. pliocenica is a noticeable bioevent because this species is an important marker species in the Miocene to Pleistocene of the Antarctic Ocean. C. pliocenica is known to be common in Miocene-Pliocene sediments of subantarctic and antarctic cores between 45S and 70S (DSDP/ODP Sites 265-269, 274, 278, 281, 513, 514, 594, 689-696, 736, 737, and 744-746) (Chen, 1975; Petrushevskaya, 1975; Keany and Kennett, 1975; Weaver, 1983; Caulet, 1986, 1991; Lazarus, 1990). The species has not been reported from sediments of the Northern Hemisphere, so it is endemic to the southern oceans. C. pliocenica has been recognized in Holes 1081A, 1082A, 1083A, 1084A, and 1085A of Leg 175 (Wefer, Berger, Richter, et al., 1998), giving its record from the lowest latitude (20S) so far. The LO of C. pliocenica of Hole 1082A is calculated to 2.08-2.11 Ma in age, suggesting that this species disappeared slightly earlier in the southeastern Atlantic than in the southern oceans where the species became extinct at 1.7-1.8 Ma (Lazarus, 1990; Caulet, 1991).

C. pliocenica first occurs in the upper Pliocene in Hole 1082A (3.3 Ma in age), whereas this species appears in Miocene time in the southern oceans (Chen, 1975; Petrushevskaya, 1975; Weaver, 1983; Caulet, 1991; Lazarus, 1992). Therefore, the FO of C. pliocenica is apparently diachronous between the southeastern Atlantic Ocean and the southern oceans.

Some bioevents characterized by abundance changes, such as the SA of Heliosoma? sp., SA of Corocalyptra kruegeri, RI of Actinomma boreale group, RI of C. davisiana, RI of Cycladophora sakaii, SA of Antarctissa group, and LCO of Stichocorys group, probably reflect changes in the ocean circulation.

Rare occurrences of C. davisiana in the lower sequence below 420 mbsf (~4.2 Ma) are probably downworked or contaminated specimens, judging from their better preservation than the other specimens in each sample. Consequently, the FO of this species in Hole 1082A was placed at the same level with the RI of the species (2.91-2.93 Ma).

Spongurus pylomaticus and Spirema sp. have been recorded in the Miocene to Pliocene submarine sequences of the North Pacific (Kling, 1973; Motoyama, 1996). The FO of S. pylomaticus, which defines the base of S. pylomaticus Zone, is given quite similar numerical ages in the North Pacific (5.2 Ma) (Motoyama and Maruyama, 1998) and in Hole 1082A (5.38-5.46 Ma). The Lower Pliocene LO of Spirema sp. may also be applicable in correlations between the North Pacific and the southeastern Atlantic.