The combined nannofossil, foraminifer, diatom, and radiolarian biostratigraphy at Site 1120 yielded 53 event levels with age significance, using the shipboard stratigraphic framework (see "Time Scale" in "Biostratigraphy" in the "Explanatory Notes" chapter). The levels are shown in Table T11 and consist of 24 first occurrence (FO) events; eight acme, first common occurrence (FCO), or last common occurrence (LCO) events, and 21 last occurrence (LO) events. FO events may have been estimated to be too shallow, based on limited sampling. The position of arrows in Figure F6 reflects the possibility that further work may extend these datums downhole. Last occurrence events may have been estimated to be too deep, again as a result of the limited sampling interval. The position of arrows in Figure F6 reflects the possibility that further work may extend these datums uphole. The more reliable events, from the point of view of stratigraphic range established at the site and for the age depth plot (Figure F6) are listed in Table T11. The dashed line in Figure F6 shows the preferred age depth model using subjective weighting for areas where different events appear inconsistent.
Biostratigraphic control varies downcore. Three intervals (0-13, 51-100, and 175-215 mbsf) have particularly abundant occurrence datums in good agreement and these form the major control for the age depth correlation shown in Figure F6. The interval between 13 and 51 mbsf is poorly constrained. The two radiolarian datums FO Prunopyle titan (14) and LO Helotholus praevema (16) and the LO of the diatom Hemidiscus karstenii (13) appear to be older than expected (see "Time Scale" in the "Explanatory Notes" chapter). The interval between 100 and 175 mbsf also has only a few apparently conflicting occurrence datums. The preferred line of correlation is constrained by the range of Praeorbulina circularis (37) and the FO of Orbulina saturalis (36). The FO of the diatom Denticulopsis dimorpha (34) and the radiolarian Eucyrtidium inflatum (35) are not considered reliable and may well be older than expected (see "Time Scale" in "Biostratigraphy" in the "Explanatory Notes" chapter). The LO of the nannofossil Sphenolithus heteromorphus (38) and the FO of the foraminifer Globorotalia praemenardii requires further scrutiny. Other bioevents that deviate from the best-fit line in Figure F6 may be a result of incomplete local range at Site 1120. These also require more scrutiny.
Summarizing the data and taking into account the above uncertainties, we conclude that average sedimentation took place in two major steps with a relatively rapid rate of ~20 m/m.y. from early to late Miocene and a slower sedimentation rate (~5 m/m.y.) from the late Miocene to present. Sedimentation rate values are simplified in Table T12. Bearing in mind that this sequence averages 94% CaCO3, fluctuations in sedimentation rate are not attributable to tectonics and terrigenous sediment dilution or starvation. The variation is most likely a result of variation in productivity and the current strengths of AAIW. The relatively shallow depth of the site also means that dissolution is unlikely to have played a role in the fluctuations. The major change is the decline in sedimentation rate at ~10 Ma for which further work must find an explanation. It may well be related to growth of the Antarctic ice sheet and possible development of a vigorous AAIW circulation across the plateau.