The diatom zonal scheme applied to the Neogene section of Hole 1138A is modified from that proposed by Harwood and Maruyama (1992) (Fig. F3A, F3B). Published datum ages were adjusted to the Berggren et al. (1995) geomagnetic polarity timescale (GPTS), and some age calibrations were preferentially selected from sites on the Kerguelen Plateau. In the modified zonal scheme used here, the first occurrence (FO) of Thalassiosira miocenica is replaced by the FO of Thalassiosira oliverana as the base of the Nitzschia reinholdii Zone, following the convention set during Legs 178 (Barker, Camerlenghi, Acton, et al., 1999) and 188 (O'Brien, Cooper, Richter, et al., 2001). The published age for both of these datums is reported at 6.4 Ma. The FO of T. oliverana, however, has not been rigorously calibrated to the polarity timescale and is considered to be an approximate age. The modified N. reinholdii Zone used here is that of Harwood and Maruyama (1992) and not the original N. reinholdii Zone defined by Weaver and Gombos (1981); the N. reinholdii Zone of Weaver and Gombos (1981) utilizes the last abundant appearance datum (LAAD) of Denticulopsis simonsenii (synonymous with the former taxonomic usage of Denticulopsis hustedtii) as its base. We have also exchanged the FO of Thalassiosira spumellaroides (22.6 Ma) for the last occurrence (LO) of Rocella gelida (22.3 Ma) as the base of the T. spumellaroides Zone. In Leg 177 studies, the FO of Thalassiosira vulnifica was replaced by the FO of Fragilariopsis weaveri (Zielinski and Gersonde, 2002). We have not made this modification but are aware of possible diachronous occurrences for both the FO and LO of T. vulnifica in different regions of the Southern Ocean.
The names of several zones summarized and defined by Harwood and Maruyama (1992) have been changed to reflect recent taxonomic revisions. The former Pliocene-Pleistocene Nitzschia spp. zones have been changed to Fragilariopsis spp., following the taxonomy of Gersonde and Bárcena (1998). The names of the former Nitzschia hustedtii-Nitzschia grossepunctata and N. hustedtii Zones are also changed to the D. simonsenii-N. grossepunctata and D. simonsenii Zones, respectively, in order to reflect taxonomic clarifications in the Denticulopsis group by Yanagisawa and Akiba (1990).
A total of 23 single and combined diatom biozones are recognized in Hole 1138A (Tables T1, T2, T3). The datums used to define the Miocene diatom zones are listed in Table T2, and the Pliocene-Pleistocene zones are outlined in Table T3. The Pliocene-Pleistocene zones are better constrained because of a higher sampling frequency in the upper part of the Hole 1138A section.
Diatom biostratigraphy divides the Neogene section of Hole 1138A into well-defined time-rock units. A poorly preserved uppermost Oligocene-lowermost Miocene diatom assemblage is present in Sample 183-1138A-29R-CC (269.41 mbsf). Samples 183-1138A-28R-CC through 12R-3, 100-101 cm (261.81-106.30 mbsf), are assigned to the lower to upper Miocene; Samples 12R-2, 100-101 cm, through 7R-3, 25-26 cm (104.80-58.05 mbsf), are placed in the lower to upper Pliocene; and Samples 7R-2, 100-102 cm, through 1R-1, 70-71 cm (57.30-0.70 mbsf), are placed in the lower to upper Pleistocene (Tables T2, T3). Sub-epoch boundaries within the Miocene and Pliocene are also roughly defined based on diatom biostratigraphy (Tables T2, T3).
Several diatom zones could not be defined in Hole 1138A because of the absence or sporadic occurrence of marker taxa. These datums include the FO of Crucidenticula kanayae (the base of the C. kanayae Zone), the FO of N. grossepunctata (the base of the N. grossepunctata Zone), and the FO of Asteromphalus kennettii (the base of the A. kennettii Zone). The FOs of Nitzschia denticuloides and D. simonsenii are concurrent; thus, the D. simonsenii-N. grossepunctata Zone could not be defined. This is also the case for the FOs of Thalassiosira torokina and Actinocyclus ingens var. ovalis, which prevented recognition of the T. torokina Zone. These concurrent datums may be a result of low core recovery, large sampling intervals, or actual synchroneity of the datum events at this site.