Multisensor track (MST) and color reflectance data (650-750 nm) collected from Holes 1090A-1090E were used to determine depth offsets in the composite section. Magnetic susceptibility, GRA bulk density, and color reflectance measurements were the primary parameters used for core-to-core correlation at Site 1090. GRA bulk density and magnetic susceptibility data were collected at 2-cm intervals on all APC cores recovered from Holes 1090A-1090D, and at 4-cm intervals on cores from Hole 1090E and Cores 177-1090B-21X through 43X. Color reflectance data were collected at 4- to 6-cm intervals on cores from Holes 1090A, 1090B, and 1090D and selected cores from Hole 1090E (see "Physical Properties" and "Lithostratigraphy" for details about MST and color reflectance data).
The data used to construct the composite section and determine core overlaps are presented on a composite depth scale in Figures F10, F11, and F12. The depth offsets that comprise the composite section for Holes 1090A-1090E are given in Table T2 (also in ASCII format in the TABLES directory).
The composite data show that the cores from Site 1090 provide a continuous overlap to at least 212 mcd and possibly as deep as 245 mcd (base of Core 177-1090B-25X). The tie between the base of Core 177-1090E-22H and the top of Core 177-1090D-23H is not firm as it is supported only by an ambiguous correlation in magnetic susceptibility data between the two cores. Neither color reflectance data nor GRA bulk density data could confirm the overlap.
Stretching and compression of sedimentary features in aligned cores indicate distortion of the cored sequence. Because much of the distortion occurred within individual cores on depth scales of <9 m it was not possible to align every feature in the MST and color reflectance records accurately by simply adding a constant to the mbsf core depth. Core-scale changes will require postcruise processing to align smaller sedimentary features. Only after allowing variable adjustments of peaks within each core can an accurate estimate of core gaps be made.
Following construction of the composite depth section for Site 1090, a single spliced record was assembled for the aligned cores over the upper 245 mcd (base of Core 177-1090-25X) primarily by using cores from Holes 1090D and 1090E (Figs. F13, F14). The composite depths were aligned so that tie points between adjacent holes occurred at exactly the same depths in mcd. Intervals having significant disturbance or distortion were avoided when possible. The Site 1090 splice (Table T3, also in ASCII format in the TABLES directory) can be used as a sampling guide to recover a single sedimentary sequence between 0 and 212 mcd. If the splice tie between the base of Core 177-1090E-22H and the top of Core 177-1090D-23H is valid (see discussion above), the sampling splice continues until 245 mcd.
Sediments recovered from Site 1090 provide a discontinuous Pleistocene through middle Eocene record. Pliocene-Pleistocene calcareous nannofossils are abundant to common, and good to medium preservation is observed for this interval of time. Early Miocene assemblage abundances are rare to common, and no biostratigraphic marker species have been recognized. Eocene-Oligocene calcareous nannofossils are abundant to few, and preservation is moderate to poor. Some barren intervals are observed (Table T4, also in ASCII format in the TABLES directory). Besides the Martini (1971) and Okada and Bukry (1980) standard zonations, we have considered some additional events for the Pliocene-Pleistocene interval, according to Raffi et al. (1993) and Wei (1993). For the Oligocene-Eocene interval, some biostratigraphic events defined by Crux (1991) are included (see "Biostratigraphy" in the "Explanatory Notes" chapter). For the Miocene-Eocene interval, the biochronology proposed by Berggren et al. (1995) is used.
The Pleistocene sequence at Site 1090 extends from 0 to 45.0 mcd. The first occurrence (FO) of Gephyrocapsa medium (4-5.5 µm), between 44.05 and 43.35 mcd, is the nannofossil event that approximates the Pliocene/Pleistocene boundary (Fig. F15). The acme of Emiliania huxleyi defines the base of Subzone NN21b and is recognized between 2.80 and 4.20 mcd. The FO of E. huxleyi is tentatively placed between 56.5 and 6.0 mcd, defining the base of Subzone NN21a. The last occurrence (LO) of Pseudoemiliania lacunosa is present between 16.75 and 115.85 mcd, and defines the base of Zone NN20. The top and base of the small Gephyrocapsa acme (see "Biostratigraphy" in the "Explanatory Notes" chapter) are not recognized from the samples analyzed; closer sampling and quantitative analyses could elucidate this acme whose occurrence has been recognized in both Sites 1088 and 1089. The LO and FO of Reticulofenestra asanoi are recognized at Site 1090 from 29.01 to 30.01 mcd and from 35.02 to 36.66 mcd, respectively. The reentrance of Gephyrocapsa medium is present between 33.51 and 35.02 mcd. The proximity between the reentrance of Gephyrocapsa medium and the FO of R. asanoi suggests the presence of a short hiatus (Fig. F15). The LO of large Gephyrocapsa (>5.5 µm) is observed from 36.66 to 37.61 mcd, whereas the FO of this morphotype is present from 39.61 to 40.41 mcd. The LO of Calcidiscus macintyrei is identified between 42.31 and 43.31 mcd (Tables T4, T5, both also in ASCII format in the TABLES directory).
Characteristic upper Pliocene assemblages are observed in Hole 1090B, and they allow us to identify Zones NN19 (LO Discoaster brouweri, 44.81-45.1 mcd) and NN18 (LO Discoaster pentaradiatus, 53.10-53.44 mcd). Typical markers of Zones NN16 and NN15 are absent or are extremely rare. Well-diversified and preserved assemblages are observed in Sample 177-1090B-8H-2, 140 cm (67.81 mcd), where Reticulofenestra pseudoumbilicus is dominant and P. lacunosa is absent. The lowest presence of the latter species (FO?) is recorded at 66.11 mcd, within Zones NN14-NN15, in agreement with Rio et al. (1990b) (4.0 Ma according to Gartner, 1990). R. pseudoumbilicus is absent at this depth. Previous workers reported a co-occurrence of R. pseudoumbilicus and P. lacunosa within Zones NN14-NN15, so we cannot exclude a hiatus between Samples 177-1090B-8H-2, 140 cm (67.81 mcd), and 8H-1, 120 cm (66.11 mcd).
A poorly preserved Miocene assemblage is observed in Hole 1090B, below 71.21 mcd. This material, comprised essentially of poorly preserved specimens of Discoaster deflandrei gr. and Cyclicargolithus floridanus, marks a hiatus from early Pliocene Zone NN15(?) to early Miocene Zones NN1-NN2(?).
The presence of long-ranging species and the absence of index species prevent the identification of biozones. Sphenolithus belemnos and intermediate forms between S. belemnos and Sphenolithus dissimilis (Rio et al., 1990a), as well as the presence of Calcidiscus sp. in a few samples, indicate an early Miocene age (Zones NN1-NN2[?]) (Table T4).
The Miocene interval is represented in Hole 1090B between 71.21 and ~140.00 mcd. The Oligocene/Miocene boundary at Site 1090 can be placed on the basis of the LO of Reticulofenestra bisecta (Zones NN1-NP25) between 175.49 and 179.49 mcd.
The events recognized within the Oligocene interval at Site 1090 include the LO of Reticulofenestra umbilica (Zone NP23) from 220.70 to 221.70 mcd, and the LO of Ismolithus recurvus (Zone NP22) between 221.70 and 222.16 mcd (Table T4).
The base of the Oligocene interval is not clearly defined in Hole 1090B because the LO of Reticulofenestra oamaruensis, the nannofossil event that approximates this boundary, is not recognized. In fact, this species is very rare and the poor preservation of the assemblages generally prevents the identification of this marker. Other nannofossil events approximate the Eocene/Oligocene boundary in Hole 1090B: the LO of Discoaster saipanensis (Zone NP20) from 262.09 to 262.69 mcd, and the acme of Clausicoccus spp. (Zone NP21) from 232.06 to 233.86 mcd (Table T5; Fig. F15).
A continuous record from the Eocene/Oligocene boundary (Zone NP21) to the middle Eocene (top of Zone NP15) is observed in Hole 1090B. The LO of Discoaster saipanensis (Zone NP21) is present between 262.69 and 262.09 mcd. Characteristic markers from Zone NP20 are not identified. The FO of Ismolithus recurvus (Zone NP19) is found between 301.49 and 302.99 mcd. In most cases, Chiasmolithus is not identified at the species level because of poor preservation; therefore, Zone NP18, whose base is defined by the FO of Chiasmolithus oamaruensis, is not recognized. The FO of R. bisecta (Zone NP17) is placed between 361.11 and 362.49 mcd and is coincident with the late/middle Eocene boundary. Between 309.39 and 361.11 mcd, Reticulofenestra reticulata, which is commonly in Zones NP16 through NP18, is recognized. The LO of Nannotetrina cristata is found between 384.57 and 390.39 mcd. The disappearance of genus Nannotetrina is commonly used to approximate the NP15/NP16 boundary (Table T4). Calibration of the above-mentioned events shows a very important diachronism (sometimes of million of years) between low and high latitudes (Berggren et al., 1995).
The sediments recovered at Site 1090 span two different time intervals. In the upper cores, from the sedimentary surface to ~74 mcd, the planktic foraminifer fauna is characterized by a Pliocene-Pleistocene assemblage. Below this depth, Miocene, Oligocene, and Eocene forms were recorded only occasionally (Table T6, also in ASCII format in the TABLES directory). In the Pliocene-Pleistocene sequences, planktic foraminifers are abundant and dominate the >63-µm fraction, and the preservation of planktic foraminifers is generally good to moderate. A slight decrease in preservation, which is marked by an increase in fragmentation and a decrease in the abundance of microperforate species, was recorded downhole. In general, Neogloboquadrina pachyderma (sinistral) is the dominant species, but Globigerina bulloides is often almost as abundant as N. pachyderma (sinistral) in the most well-preserved core-catcher (CC) samples. In addition to these two species, the planktic foraminifer fauna consists of Globigerina quinqueloba, Globigerinita glutinata, Globigerinita uvula, Globorotalia puncticulata, and Globorotalia puncticuloides. Globorotalia inflata was only recorded in the uppermost one or two cores in each hole. In addition, small, microperforate species (not shown in Table T6), mainly in the 63- to 150-µm fraction, make a significant contribution to the planktic foraminifer fauna. The abundance of microperforate species is lower in samples with a higher degree of fragmentation.
At Site 1090, the population of G. puncticuloides includes two distinct forms characterized by differences in the morphology of the last chamber. In addition to the more abundant form, which exhibits compressed margins (see Scott et al., 1990, their fig. 48), the Pliocene-Pleistocene assemblage at Site 1090 often contains relatively large specimens with a more broadly rounded final chamber. In this form the resemblance with G. puncticulata is stronger. The existence of different four-chambered morphotypes of G. puncticuloides in Pleistocene sequences from the mid- and high-latitude Southern Hemisphere gives rise to taxonomic confusion between this species and G. puncticulata, making it hard to establish the temporal distribution of these two species in different regions (e.g., Berggren, 1992).
Below the inferred hiatus, only rare occurrences of planktic foraminifers were recorded in the studied CC samples, and many of the samples were barren. In the Paleogene and late Neogene sediments, the preservation of planktic foraminifers was poor to moderate. In the poorly preserved CC samples many tests were partially broken or severely etched. The few specimens recorded can be divided into two groups: a lower Miocene assemblage characterized by the presence of Catapsydrax dissimilis, C. stainforthi, Globigerina woodi, and Globorotalia miozea; and a middle Eocene assemblage characterized by the presence of a few Acarinina species, Chiloguembelina cubensis, Mozozovella spinulosa, Pseudohastigerina micra, Subbotina angiporoides, S. brevis, S. linaperta, and Truncorotaloides spp. (Table T6). Because of the low abundance of planktic foraminifers and the low temporal resolution of the samples examined it is difficult to further divide the assemblages into any zonations. However, the co-existence of C. dissimilis and Globoquadrina dehiscens in Sample 177-1090D-9H-CC, 7-12 cm (90.48 mcd), suggests an early Miocene age for this sample if the warm subtropical zonation scheme by Jenkins and Srinivasan (1986) is applied (no Subantarctic or Antarctic zonation schemes use C. dissimilis as a marker species) (Fig. F15). Berggren et al. (1995) suggested a last appearance datum of C. dissimilis at 17.3 Ma. Furthermore, the presence of both Acarinia primitiva and P. micra in Sample 177-1090B-39X-CC, 12-22 cm (370.77 mcd) indicates that this sample belongs to Zone AP9 or AP10 in the zonal scheme of Stott and Kennett (1990) and that it is of middle Eocene age (Fig. F15). Future quantitative high-resolution studies of the Paleogene and late Neogene planktic foraminifer fauna at Site 1090 may provide important paleoceanographic information, especially if combined with stable isotopic studies and census counts of nannofossils. The abundance of planktic foraminifers varies markedly downhole (narrow bands of foraminifer sand are present), and meaningful biostratigraphic and paleoceanographic interpretations will only be possible from high-resolution sampling.
Benthic foraminifers at Site 1090 are highly variable, both in abundance and preservation. Above the lithologic transition that marks the early Miocene-Pliocene unconformity, benthic foraminifers generally constitute between 5% and 10% of the total foraminifer fauna from the >63-µm fraction; these values increase to >50% in most early Miocene-Eocene samples. Absolute foraminifer abundances are variable and exhibit a clear trend toward higher values above ~67 mcd in Hole 1090B, reaching a maximum of 220 specimens/cm3 in Sample 177-1090B-3H-CC, 13-18 cm. An increase in foraminifer abundance is also observed below 361.11 mcd in Hole 1090B, notably in Sample 177-1090B-41X-CC, 0-5 cm, where 170 specimens/cm3 were recorded (Table T7, also in ASCII format in the TABLES directory; Fig. F16). Preservation above ~67 mcd is moderate to good and generally improves uphole. Two significant barren intervals are present in Samples 177-1090B-20H-CC through 22H-CC and 33X-CC through 35X-CC, within the Oligocene and late Eocene, respectively. Although carbonate abundances are generally low throughout these intervals, they remain highly variable (see "Geochemistry"), and the low temporal resolution of the CC samples examined may have resulted in missed intervals containing foraminifers. Quantitative estimates of relative species abundance were made from Hole 1090B, with counts of up to 200 specimens per sample. Species richness is variable, with a maximum of 38 taxa recorded in Sample 177-1090B-2H-CC, 0-10 cm, and a minimum of four taxa in Samples 177-1090B-28X-CC, 16-21 cm, and 32X-CC, 13-18 cm. Some of this variability can be accounted for by sample size (see "Biostratigraphy" in the "Explanatory Notes" chapter), but the consistently low values from ~150 to 350 mcd represent an interval of generally poor preservation that corresponds to generally lower carbonate abundances.
Although certain intervals of the cored sequence appear to be far from ideal for high-resolution benthic foraminifer studies of the Paleogene/Neogene transition, the near-ubiquitous presence of extremely well-preserved specimens of Oridorsalis umbonatus at Site 1090 suggests that there is good potential for establishing a long benthic stable isotopic record here.
Assemblage 1 is present in Samples 177-1090B-1H-CC through 7H-CC, from the mudline to 65.59 mcd. Biostratigraphic differentiation of the Pliocene-Pleistocene at Site 1090 is limited to the LO of Stilostomella lepidula in Sample 177-1090B-4H-CC, 12-17 cm, which supports a late Pleistocene age above 35.02 mcd in Hole 1090B (e.g., Thomas, 1987). This depth is below the LO of A. ingens (0.64 Ma) at a depth of 23.45 mcd in Hole 1090B (see "Diatoms"). The benthic foraminifer assemblages of the Pliocene-Pleistocene sequence (Table T7) are the most diverse at Site 1090 (average number of species = 30) and are dominated by Alabaminella weddellensis, Epistominella exigua, Melonis pompilioides, and Pullenia quinqueloba. Additional taxa which are constrained to Assemblage 1 include Pyrgo murrhiana, Stainforthia loeblichi, Triloculina trihedra, and Uvigerina hispidicostata. As noted at Site 1089, these taxa, notably E. exigua and A. weddellensis, are known to respond to changes in phytodetritus supply within the flocculent layer of the ocean floor (Gooday, 1988). Because of the low temporal resolution of the CC samples examined, however, it has not been possible to establish the assemblage response through the glacial-interglacial cycles of the Pliocene-Pleistocene.
Assemblage 2 is present in Samples 177-1090B-8H-CC through 37X-CC, from 73.34 to 349.15 mcd. The benthic foraminifer assemblages are rather poor and generally very low in species diversity (average number of species = 12), suggesting that certain intervals may have been influenced by carbonate dissolution (see "Planktic Foraminifers"). Low species richness and high agglutinated (noncalcareous) to calcareous benthic foraminifer ratios throughout the upper Eocene-Oligocene sequence, from ~150 to 350 mcd in Hole 1090B, suggest that this interval has, indeed, been influenced by carbonate dissolution. The most abundant taxa are O. umbonatus, Pullenia spp., Stilostomella spp., and Ophthalmidium spp. The LO of Alabamina dissonata in Sample 177-1090B-31X-CC, 27-32 cm (303.43 mcd), which ranges from the late Paleocene to latest Eocene, is of limited biostratigraphic value, but it generally supports the assignment of this interval to Zone NP19 (>36 Ma). The LO of Nuttallides truempyi in Sample 177-1090B-36X-CC, 26-34 cm (342.11-342.19 mcd), which is considered by Berggren and Aubert (1983) to provide a useful marker of the Eocene/Oligocene boundary, may have been influenced by the barren intervals above and is not considered a reliable event. It is possible that detailed postcruise sampling throughout this sequence may provide information on the diachroneity of this LO from the late middle Eocene to latest Eocene reported by Tjalsma and Lohman (1983).
Assemblage 3 is present in Samples 177-1090B-38X-CC through 43X-CC, from 361.11 mcd to the bottom of the hole. Benthic foraminifers are moderately well preserved, with no evidence of calcite overgrowths, and both total abundances and diversity are higher in this interval (average number of species = 20), coinciding with higher carbonate abundances. The dominant taxon is N. truempyi, together with common occurrences of A. dissonata, Bulimina semicostata, Cibicidoides praemundulus, and Stilostomella subspinosa. The presence of Aragonia aragonensis in Sample 177-1090B-42X-CC, 22-27 cm (389.49 mcd), provides a potentially valuable indication of latest middle Eocene age.
The sediments recovered in four holes at Site 1090 consist of a Pleistocene to middle Eocene record that is disrupted by a hiatus around 69-70 mcd. The hiatus spans the lower Pliocene to lower Miocene. A few decimeters below this hiatus is a volcanic ash layer described in "Lithostratigraphy". Other hiatuses have been identified in the lower Pleistocene and upper Pliocene sequence (Fig. F15). To obtain a reliable biostratigraphy and to accurately identify hiatuses, we examined smear slides from Holes 1090B and 1090E in addition to the CC samples (Tables T5, T8, both also in ASCII format in the TABLES directory). In the Pliocene-Pleistocene sequence, we observed moderately preserved diatoms that generally range from few to common in occurrence. The sediments from a few meters above and ~10-13 m below the lower Pliocene/lower Miocene hiatus bear only trace to rare diatoms, and some intervals are barren. Between ~85 and 340 mcd, diatom assemblages are generally moderately to well preserved and are common to abundant in smear slides. A distinct abundance maximum of biosiliceous components, mainly diatoms, is present between 250 and 320 mcd in the upper Eocene sequence, as also indicated by opal measurements (see "Lithostratigraphy"). Middle Eocene sediments below 340 mcd are barren of siliceous microfossils. However, the presence of clinoptilolite in this interval indicates the former presence of biogenic opal that has been diagenetically transformed.
For the establishment of diatom biostratigraphic zonations at Site 1090 we used the zonations proposed by Gersonde and Bárcena (1998) and Gersonde et al. (1998) for Pleistocene to late Pliocene and Pliocene to Oligocene sediments, respectively. The stratigraphic age assignment for the Eocene was based on the zonation proposed by Fenner (1984) (see "Biostratigraphy" in the "Explanatory Notes" chapter). The diatom stratigraphic information from the four holes was combined and converted to the mcd scale. Events were placed at the midpoint of the investigated sample depth (Tables T5, T8).
The base of the upper Pleistocene Thalassiosira lentiginosa Zone, marked by the LO of Actinocyclus ingens, can be placed at 18.30 mcd (Table T9, also in ASCII format in the TABLES directory; Fig. F15). The T. lentiginosa Subzone b, which ranges from the base of marine isotope Stage (MIS) 11 to the top of MIS 7 is between 7.23 and 14.58 mcd. Thalassiosira elliptipora, which is a marker of the A. ingens Subzone c, could not be identified below the T. lentiginosa Zone. Fragilariopsis barronii, which is present in the A. ingens Subzone a, was only observed in one sample at 44.81 mcd (177-1090B-5H-6, 130 cm), near the top of a ~2.5-m-thick interval with normal polarity (see "Paleomagnetism") interpreted to represent the Olduvai Subchron (C2n) (Fig. F15). This interpretation is based on the occurrence of assemblages typical of the Proboscia barboi Zone identified in Samples 177-1090B-5H-1, 70 cm, 5H-CC, 14-19 cm, 177-1090E-5H-1, 70 cm, and 5H-2, 70 cm, between 45.01 and 47.69 mcd. This zone stratigraphically coincides with the Olduvai Subchron. The apparent lack of most of the A. ingens Subzone a indicates a hiatus at or close to the top of the Olduvai Subchron and the P. barboi Zone. This hiatus may have removed a time record ranging minimally from 1.3 to 1.8 Ma, and may also include the upper portion of the Olduvai Subchron. The sediment in the magnetically reversed interval below the Olduvai Subchron (between ~47.7 and 56 mcd) belongs to the Thalassiosira kolbei/Fragilariopsis matuyamae Zone, indicated by the co-occurrence of few to common P. barboi and F. matuyamae. The nominate taxon of this zone, T. kolbei, was encountered only in trace amounts. From 56 to 61.8 mcd, diatom assemblages are marked by few to common Fragilariopsis weaveri and the absence of Thalassiosira insigna. T. insigna is known to have an acme in the upper portion of the Gauss Chron in southern high latitudes (Burckle et al., 1990). This indicates an age range between 2.8 and 3.26 Ma in the lower portion of the T. insigna Zone. The top of this interval is marked by the LO of F. weaveri around 2.8 Ma, occurring in Subchron C2An.1n of the Gauss Chron, whereas the LO of T. insigna defines the top of the T. insigna Zone at the top of the Gauss Chron. The absence of the nominate taxon of the Thalassiosira vulnifica Zone (~2.5 to 2.63 Ma) and the absence of T. insigna indicate a hiatus lasting between 2.5 and 2.8 Ma during the lowermost portion of the Matuyama to upper Gauss Chrons. The absence of both taxa cannot be explained by paleoenvironmental conditions that would exclude their presence at Site 1090, as both taxa have been reported in few to common numbers in Subantarctic waters (Fenner, 1991). Below the sediments assigned to the T. insigna zone, we note the presence of the nominate species of the Fragilariopsis interfrigidaria Zone at 62.32 mcd, whereas F. weaveri that has its FO in the uppermost portion of the F. interfrigidaria Zone is absent. The F. interfrigidaria Zone ranges from 3.26 to 3.8 Ma, spanning the lower portion of the Gauss and upper Gilbert Chrons. At ~67.5 mcd, we found F. praeinterfrigidaria, the evolutionary precursor of F. interfrigidaria. This indicates that sediments between ~67.5 mcd and the lower Pliocene/lower Miocene hiatus are somewhat older than 3.8 Ma (Table T9; Fig. F15).
The lower Pliocene/lower Miocene hiatus around 69-70 mcd is marked by a mixture of Pliocene and lower Miocene diatom species and is underlain by sediments that contain poorly preserved lower Miocene assemblages that do not allow a distinct age assignment. Moderately to well-preserved diatoms were recovered in Samples 177-1090E-9H-1, 70 cm, and 9H-5, 70 cm, at 85.81 and 91.81 mcd, respectively. Few to common occurrences of Thalassiosira fraga place this sediment in the early Miocene T. fraga Zone, which ranges from 17.7 to 20.8 Ma. The absence of Thalassiosira spumellaroides, a taxon co-occurring with T. fraga in the middle and lower portions of the T. fraga Zone, suggests that sediments below the hiatus belong to this zone (<19 Ma). Co-occurrence of both taxa was found between 95.29 and 111.4 mcd, which indicates an age between 19 and 20.8 Ma for this interval. Assemblages of the underlying T. spumellaroides Zone were recovered to a depth of 126.5 mcd (Table T9; Fig. F15). Rare to few occurrences of Rocella gelida place the underlying interval to a depth of 182.6 mcd in the R. gelida Zone, which ranges from 22.58 to 26.5 Ma and straddles the early Miocene/late Oligocene boundary. The co-occurrences of Rocella vigilans and R. gelida place the sediments between 157.6 and 182.6 mcd in the lower portion of the R. gelida Zone, between 25.5 and 26.5 Ma. The diatom biostratigraphic zonation thus indicates that the Miocene/Oligocene boundary, at 23.8 Ma (Berggren et al., 1995), is present between 126.5 and 157.6 mcd. This is consistent with the age assignments based on radiolarians (Table T9; Fig. F15). Diatom assemblages belonging to the Lisitzinia ornata and the Azpeitia gombosii Zones (early late Oligocene) have been found between 192 and 211.3 mcd.
The early late and late early Pliocene assemblages between ~60 mcd and the hiatus are composed of diatoms that indicate rather warm sea-surface temperatures (SSTs). Taxa such as Azpeitia nodulifer, Azpeitia tabularis, Hemidiscus cuneiformis, Fragilariopsis fossilis, F. reinholdii, Thalassionema nitzschioides, Thalassiosira oestrupii, and T. eccentrica are prominent, whereas typical Antarctic taxa, such as F. barronii and Thalassiosira inura that are present in high numbers in mid-Pliocene sediments in more southern latitudes than Site 1090 (Abelmann et al., 1990; Gersonde and Burckle, 1990), are rare or absent. This supports Dowsett et al.'s (1996) conclusion based on a global survey that mid-Pliocene SSTs were higher than modern. Site 1090 was probably located north of an oceanographic system that separated more southerly cold waters from warmer waters in the north during the mid-Pliocene, comparable to the present Subtropical Front.
Radiolarian biostratigraphy at Site 1090 is based on the examination of 39 CC samples (Table T10, also in ASCII format in the TABLES directory). Radiolarians are abundant to common and generally well preserved in Site 1090 sediments, except for the lower part of Hole 1090B (Samples 177-1090B-38X-CC, 8-18 cm, 39X-CC, 12-22 cm, and 43X-CC, 29-39 cm; below 349.12 mbsf [361.20 mcd]) where radiolarians are absent. Preliminary investigations indicate that at least two stratigraphic units are present throughout the recovered sequence.
The upper unit which is between Samples 177-1090C-1H-CC, 14-19 cm (2.80 mbsf, 3.16 mcd), and 1090D-7H-CC, 17-22 cm (62.90 mbsf, 68.98 mcd), is Pleistocene to late Pliocene in age. The four Samples 177-1090B-1H-CC, 20-25 cm (4.20 mbsf, 4.25 mcd), 1090C-1H-CC, 14-19 cm (2.80 mbsf, 3.35 mcd), 1090D-1H-CC, 12-17 cm (7.37 mbsf, 7.76 mcd), and 1090E-1H-CC, 8-13 cm (8.65 mbsf, 12.08 mcd), are correlative with the mid-latitude Botryostrobus aquilonalis Zone and high-latitude Omega Zone based on the absence of Stylatractus universus. The precise stratigraphic positions of other samples are not well understood because of the absence of index species, but Samples 177-1090B-7H-CC, 12-17 cm (59.29 mbsf, 65.64 mcd), 1090C-7H-CC, 8-13 cm (58.88 mbsf, 66.75 mcd), and 1090D-7H-CC, 17-22 cm (62.90 mbsf, 69.03 mcd), are older than 2.61 Ma based on the first appearance datum (FAD) of Cycladophora davisiana. Radiolarian assemblages from the upper unit are similar to known Antarctic fauna, but application of Antarctic radiolarian zonations appears to be hampered by the transitional character of the species composition.
The lower unit below the hiatus ranges in age from early Miocene to late Eocene. All Miocene samples contain Lychnocanoma conica, and some are correlative with the lower Miocene Cyrtocapsella longithorax Zone or underlying Cycladophora antiqua Zone. The distinction of the two zones at Site 1090 is impossible because the FAD of C. longithorax, which defines the base of the Cyrtocapsella longithorax Zone (Abelmann, 1990, 1992), seems to be below the Miocene/Oligocene boundary as suggested by Caulet (1991). Samples 177-1090B-8H-CC, 18-23 cm (69.63 mbsf, 73.39 mcd), 1090C-8H-CC, 0-5 cm (68.80 mbsf, 76.66 mcd), and 1090D-8H-CC, 17-22 cm (73.97 mbsf, 79.93 mcd), whose horizons are probably situated immediately below the hiatus, contain reworked radiolarians in moderate number. The Miocene/Oligocene boundary can be placed below Sample 177-1090B-13H-CC, 0-5 cm (117.12 mbsf, 122.60 mcd), based on the FAD of Cyrtocapsella tetrapera (Takemura and Ling, 1997). The Antarctic Paleogene radiolarian zones established by Takemura (1992) (i.e., Eucyrtidium spinosum, Axoprunum irregularis, and Lychnocanoma conica Zones in ascending order) are recognized in Hole 1090B. The boundary between the upper Oligocene Lychnocanoma conica Zone and lower to upper Oligocene Axoprunum irregularis Zone can be identified between Samples 177-1090B-21X-CC, 15-25 cm (194.09 mbsf, 202.20 mcd), and 22X-CC, 0-10 cm (203.69 mbsf, 211.40 mcd). This boundary is also recognized in Holes 1090D and 1090E. The boundary between the Axoprunum irregularis Zone and the upper Eocene to lowest Oligocene Eucyrtidium spinosum Zone is not well recognized yet. At present, it is certain that Sample 177-1090B-27X-CC, 13-18 cm (251.53 mbsf, 263.60 mcd), and the interval below are correlative with the Eucyrtidium spinosum Zone. As a whole, the assemblages from the lower unit are very similar to those reported by Abelmann (1990, 1992), Takemura (1992), and Takemura and Ling (1997).
Archive halves of APC and XCB cores recovered at Site 1090 were measured using the shipboard pass-through magnetometer. Measurements were made at 5-cm intervals. Sections obviously affected by drilling disturbance were not measured. Hole 1090A (Cores 177-1090B-1H through 20H) and Hole 1090D were measured after alternating-field demagnetization at peak fields of 0 (natural remanent magnetization [NRM]), 5, 10, 15, and 20 mT. Cores 177-1090B-21X through 40X were measured at peak fields of 0, 10, 20, and 25 mT. Cores 177-1090E-1H through 14H were measured at peak fields of 0, 10, and 20 mT. Cores 177-1090E-15H through 25H were measured at peak fields of 0 and 20 mT. Cores 177-1090D-4H through 24H and 177-1090E-5H through 24H were oriented using the Tensor tool, except for Cores 177-1090D-18H and 19H and 177-1090E-10H, 11H, and 17H, which were not oriented because of technical difficulties.
At Site 1090, NRM intensities vary with lithologic unit (see "Lithostratigraphy"). In Unit I, the white calcareous oozes above the prominent tephra layer at ~65 mbsf in Hole 1090B (see "Lithostratigraphy") have NRM intensities between 4 × 10-2 and 1 × 10-4 A/m. In Unit II, the upper Eocene to lower Miocene red/brown oozes below the tephra layer have NRM intensities between 2 × 10-2 and 8 × 10-2 A/m. In Unit III, comprising middle Eocene oozes at the base of the recovered section, the intensities are generally >1 × 10-1 A/m.
Above the tephra layer at ~65 mbsf within lithologic Unit I in Hole 1090B (see "Lithostratigraphy"), the inclinations of magnetic remanence (after 20-mT peak field demagnetization) indicate identifiable polarity zones (Fig. F17; Table T9). The inclination records from the three holes are, however, highly discontinuous because of poor recovery and/or drilling disturbance. The polarity interpretation given in Table T9 and Figures F15 and F17 is based on the record from Hole 1090C. The two normal polarity zones within the Matuyama Chron are interpreted as the Jaramillo and Olduvai Subchrons. Correlations from Hole 1090C to the other holes imply offsets of several mbsf, which should be reconciled when the data are converted to the mcd scale.
Below 70 mbsf in Holes 1090B, 1090D, and 1090E, the inclination records indicate clearly defined polarity zones for the APC sections from Holes 1090D and 1090E, and for the APC and XCB sections from Hole 1090B (Fig. F18). Close to the top and base of the XCB section in Hole 1090B, at ~185 and ~340 mbsf, respectively, inclination values are highly scattered because of drilling-induced core deformation; however, the inclination pattern in the rest of the XCB section is coherent and allows polarity zones to be identified. Correlation of polarity zones to the GPTS is ambiguous because of the lack of distinctive "fingerprints" in the polarity zone pattern. The correlation must, therefore, await detailed postcruise biostratigraphy.
A 397-m-thick sedimentary section spanning the interval from the Pleistocene through the middle Eocene was recovered at Site 1090. The base of Site 1090 lies at about the boundary of calcareous nannofossil Zones NP16/NP15 of Martini (1971) or the lowermost Zone CP14 of Okada and Bukry (1980). Therefore, the basal age at Site 1090 is between 43 and 44 Ma according to the Berggren et al. (1995) time scale. The stratigraphic record is interrupted by a major hiatus at ~70 mcd, which separates lower Pliocene from upper Miocene sediments and spans ~14 m.y. Other hiatuses may occur in the Pleistocene and upper Pliocene sections of Site 1090 (Figs. F15, F19).
Holes 1090A-1090E were cored with the APC to 7, 173, 69, 226, and 237 mbsf, respectively. Hole 1090B was drilled with the XCB to a TD of 397 mbsf (Fig. F15). The combined MST and color reflectance data indicate that a continuous sedimentary section was recovered to 212 mcd and possibly as deep as 245 mcd at the base of Core 177-1090B-25X (Figs. F10, F11, F12).
Age assignment and calculation of sedimentation rates for Site 1090 are based on calcareous nannofossil, diatom, and radiolarian biostratigraphy, as well as measurements of geomagnetic polarity reversals (Table T9; Figs. F15, F19).
The uppermost 44 mcd corresponds to the Pleistocene. All biostratigraphic datums yield consistent age assignments to the base of the Jaramillo Subchron at ~35 mcd. Sedimentation rate averages ~35 m/m.y. (Fig. F19B). A combination of physical properties variations and biostratigraphic data allows the identification of MISs 4 to 12 in the upper 18 mcd (Figs. F10, F11, F12). The proximity of two stratigraphic datum levels, the LO of P. lacunosa (0.46 Ma at ~16.30 mcd) and the top of the A. ingens Zone (0.65 Ma at ~18.30 mcd) (Table T9), indicates a short hiatus during this interval (Figs. F15, F19). Cyclicity in color reflectance and GRA bulk density records documents the shift from the "100-k.y.-world" to the "40-k.y.-world" around 30 mcd at Site 1090 (Figs. F13, F14). We observe a mismatch between the diatom and nannofossil age assignments between ~35 and 43 mcd, and thus between the Jaramillo and Olduvai Subchrons within the Matuyama Chron. Diatom ages indicate continuous high sedimentation rates of ~33 m/m.y. below the Jaramillo Subchron, followed by a hiatus at ~43 mcd that spans the time interval between ~1.3 and 1.8 Ma. In contrast, calcareous nannofossil datums show a drop in sedimentation rates below the Jaramillo Subchron and a possible hiatus at ~38.5 mcd (Figs. F15, F19B). However, this disagreement could be related to inaccurate age calibration of the microfossil data. Improvement of the age calibrations will be possible only after integration of all Pliocene-Pleistocene stratigraphic data obtained during Leg 177.
The upper Pliocene sequences have been deposited at sedimentation rates of ~10-12 m/m.y. (Fig. F19B; Table T9). Diatom data indicate another hiatus at ~55 mcd, which approximately spans the time interval between 2.5 and 2.8 Ma. The lowermost portion of the Pliocene, below ~3.8 Ma, is missing because of a hiatus present at 70 mcd. This hiatus is underlain by a tephra layer (see "Lithostratigraphy").
The youngest sediments below the hiatus are early Miocene in age and were assigned to the T. fraga diatom Zone and C. longithorax-C. antiqua radiolarian Zone, which indicate an age of ~19 Ma or slightly younger for this interval. On the basis of diatom and radiolaria occurrences, it is estimated that the Miocene/Oligocene boundary is at ~150 mcd. Calcareous nannofossil and radiolarian data place the Oligocene/Eocene boundary at ~230 mcd. The late/middle Eocene boundary was interpreted to be at ~360 mcd. This age interpretation results in sedimentation rates of ~10 m/m.y. in the early Miocene, Oligocene, and middle Eocene, whereas calculated average sedimentation rates are ~30 m/m.y. in the late Eocene (Fig. F19A; Table T9). However, the upper Miocene to middle Eocene biostratigraphic age assignments are very preliminary because not all available stratigraphic data points have yet been identified and included in the age model. The same is true for the paleomagnetic polarity record, which displays clearly defined polarity zones in the APC as well as in the XCB sections. More detailed combined biostratigraphic and paleomagnetic studies will allow the establishment of a well-defined age model for the early Miocene to middle Eocene interval. The preliminary data hold much promise for correlating calcareous and siliceous zonations to the GPTS, which will considerably improve our knowledge of early Miocene to middle Eocene stratigraphy in southern high latitudes.