Preliminary age assignments were based on biostratigraphic analyses of calcareous nannofossils and planktonic foraminifers. Core catchers from all holes at each site were examined for calcareous nannofossils and planktonic foraminifers, whereas for benthic foraminifers, usually only core catcher samples from the hole with the deepest penetration were used. In addition to core catcher samples, one to six samples per core (sample spacing = 0.25–1.5 m) were examined for shipboard determination of stratigraphic positions of nannofossil datums. More detailed sampling was conducted across critical intervals such as the P/E and K/P boundaries. The preservation, abundance, and zonal assignments for each microfossil group have been recorded in the stratigraphic site summary sheets and entered into Excel spreadsheets (see the "Supplementary Material" contents list).
The zonal scheme of Bukry (1973, 1975) with the CN and CP zonal code notation as added by Okada and Bukry (1980) was used for Cenozoic calcareous nannofossil biostratigraphy. These zonations provide a general framework for the biostratigraphic classification of mid- to low-latitude nannofossil assemblages and are correlated with Martini's (1971) NN and NP nannofossil zones (Fig. F4). In the lower Eocene, the Zone CP11/CP10 boundary was not recognized because its marker species, Toweius crassus, was recorded in older sediments (within Subzone CP9b) at several Leg 208 sites. The lower Paleocene marker species Ellipsolithus macellus, used in the standard zonations, is very rare in the lower part of its range and was not used for biostratigraphic classificiation. Nannofossil taxonomy follows that of Perch-Nielsen (1985). Cenozoic biostratigraphic events, including Okada and Bukry's (1980) zonal indicators and additional markers, are listed in Table T1. Age estimates of biohorizons are all adjusted to the timescale of Leg 208 (see "Age Model and Mass Accumulation Rates").
The zonal scheme of Sissingh (1977; CC zones), as modified by Perch-Nielsen (1985) and Roth (1978; NC zones), was used for the Upper Cretaceous. These zonations provide a general framework for the biostratigraphic classification of mid- to low-latitude nannofloral assemblages (Table T1). Cretaceous nannofossil taxonomy follows that is used in Perch-Nielsen (1985) and Bown (1999). Numerical ages reported by Erba et al. (1995) have been adopted for the Maastrichtian events (Table T1).
The tropical planktonic foraminiferal zonal scheme for the Cenozoic follows Berggren et al. (1995a, 1995b) (Fig. F4; Table T2). For the Miocene, the zonation for transitional austral regions was used in addition to the tropical zonation because of the overall absence of several tropical/subtropical marker species (Berggren et al., 1995a, 1995b). Age estimates of the datums are all adjusted to the timescale of Leg 208 (see "Age Model and Mass Accumulation Rates"). Numerical ages and references for the Cenozoic planktonic foraminiferal datum levels are presented in Table T2. Cenozoic taxonomic concepts selectively follow Postuma (1971), Kennett and Srinivasan (1983), Bolli and Saunders (1985), Toumarkine and Luterbacher (1985), Spezzaferri and Premoli Silva (1991), Chaisson and Leckie (1993), Spezzaferri (1994), Chaisson and Pearson (1997), Pearson and Chaisson (1997), and Olsson et al. (1999). The zonation used for Cretaceous planktonic foraminifers is based on the tropical zonal schemes of Caron (1985).
According to the Astronomical Tuned Neogene Timescale 2004 (Lourens et al., in press), the top of the Hirsutella cibaoensis range of 4.16 Ma falls within Zone PL2 and can therefore not be used to define Subzone PL1a within the lowermost Pliocene (Fig. F4). Furthermore, a number of key taxa used in the standard tropical biozonation for the Neogene are either extremely rare or absent in Leg 208 sites, thereby negating their stratigraphic utility. Some of these ecologically excluded Neogene taxa are Globigerinoides fistulosus, Globorotalia tumida, Menardella exilis, Menardella pertenuis, Menardella multicamerata, Pulleniatina primalis, and Pulleniatina spectabilis. Later members of the biostratigraphically useful Fohsella lineage are missing from the middle Miocene as well. The absence of these marker taxa limits biostratigraphic resolution within the Neogene sections from Walvis Ridge. For instance, Zones M2 and M3 of the lower Miocene could not be differentiated because of the absence of the marker species Globigerinatella insueta.
Ecological exclusion of tropical marker species occurs within the Paleogene as well. Moreover, intervals of intense dissolution and extensive reworking throughout much of the middle Eocene through Oligocene have disrupted the Paleogene stratigraphy, and species of the marker taxon Hantkenina spp. are rare or absent. The marker Cribrohantkenina inflata is not present in any of the uppermost Eocene sections, making it impossible to differentiate Zones P16 and P17. Hence, the standard biozonation of Berggren et al. (1995a, 1995b) has been amended by omitting Zone P17 (Fig. F4). The absence of such dissolution-susceptible taxa as Planorotalites palmerae and Hantkenina nuttalli precludes direct identification of Zones P9 and P10, and alternative biostratigraphic criteria are used. To this end, Zone P9 is approximated by an acme in the relative abundance of Morozovella caucasica. Scarcity of Morozovella formosa and Morozovella velascoensis within well-preserved assemblages from the lower Eocene is evidence for atypical environmental conditions in this subtropical area.
The boundary between Zone P5 and Subzone P6a could not be determined with confidence at Leg 208 sites. This boundary is defined by the uppermost occurrence of M. velascoensis (54.31 Ma), which occurs above the benthic foraminiferal extinction event (BEE) at the base of the carbon isotope excursion (CIE) marking the P/E boundary (55 Ma) (Norris and Röhl, 1999; Luterbacher et al., 2000). At all Leg 208 sites, the uppermost occurrence of M. velascoensis was observed close to the BEE, with only rare and discontinuous occurrences above that event. The last appearance datum of M. velascoensis is thus diachronous, and the species has a shorter range at higher latitudes. We used the recalibrated numerical ages (Norris and Röhl, 1999; Röhl et al., 2000, 2003) within a narrow stratigraphic interval bracketing the P/E boundary (55.0 Ma). For the Upper Cretaceous, we used the numerical age for zonal boundaries as given in Erba et al. (1995), Bralower et al. (1995), and Premoli Silva and Sliter (1999) (Table T1).
We used the depth zonation of van Morkhoven et al. (1986) for benthic foraminifers (Table T3).
Benthic generic classification follows mainly Loeblich and Tappan (1988), except where modified by Hayward (2002) for uniserial taxa. Species classification mainly follows Boltovskoy (1978), Tjalsma and Lohmann (1983), Thomas (1985, 1990), van Morkhoven et al. (1986), Boltovskoy and Boltovskoy (1989), various papers on agglutinated foraminifers in Hemleben et al. (1990), Müller-Merz and Oberhänsli (1991), Boltovskoy and Watanabe (1993), Boltovskoy et al. (1995), Nomura (1995), Thomas and Shackleton (1996), Schmiedl et al. (1997), Widmark (1997), and Alegret and Thomas (2001).
Because of time constraints aboard the ship, assemblages could be examined only superficially and not enough specimens could be counted per sample for full evaluation. Deep-sea benthic foraminifers are diverse, and at least 200 specimens must be counted for Neogene assemblage evaluation (300 for Paleogene assemblages). Only ~50–100 specimens per sample could be counted, and the paleodepth estimates, ranges of taxa, and remarks on dominance and species richness are thus preliminary.
The lower Oligocene is zoned in greatest detail by calcareous nannofossils (Fig. F4), although the Eocene/Oligocene (E/O) boundary is recognized by the last occurrence (top [T]) of the planktonic foraminiferal genus Hantkenina. The E/O boundary is present within uppermost Chron C13r at the planktonic foraminifer Biozone P16/P18 boundary and within the calcareous nannofossil Subzone CP16a (NP21).
Calcareous nannofossil datum levels within 2 m.y. of the E/O boundary (33.7 Ma) include the uppermost occurrence (T) of Reticulofenestra umbilicus (31.7 Ma), the T of Ericsonia formosa (32.9 Ma), and the onset (bottom [B]) of the acme of Ericsonia obruta (33.7 Ma). The bottom of the acme of E. obruta in the mid- to low latitudes seems to approximate the E/O boundary. The uppermost Eocene is easily determined by the presence of the rosette-shaped discoasterids Discoaster saipanensis (T at 34.0 Ma) and Discoaster barbadiensis (T at 34.2 Ma). The boundary interval should be constrained between the T of D. saipanensis and D. barbadiensis and the T of E. formosa.
Among the planktonic foraminifers, the tops of the stratigraphic ranges of Globigerinatheka spp. (~34.3 Ma) and the Turborotalia cerroazulensis lineage (~33.8 Ma) can be used to approximate the E/O boundary. Unfortunately, the genera Hantkenina and Turborotalia are particularly susceptible to dissolution. Hence, the top of the thick-shelled globigerinathekids is used to approximate the E/O boundary. In addition, the top of the genus Pseudohastigerina (32.0 Ma) delimits the top of the lowermost planktonic foraminiferal zone in the Oligocene.
The P/E boundary has been placed at the planktonic foraminiferal Zone P5/P6 boundary in the middle part of calcareous nannofossil Zone NP10 (CP9a) (Aubry et al., 1996). The International Subcommission on Paleogene Stratigraphy on the Criterion for the Recognition of the Paleocene/Eocene Boundary redefined this boundary and placed it at the base of the CIE. The CIE, and by definition the P/E boundary, occurs at the same level as the top of the benthic foraminifer Stensioeina beccariiformis, within reversed Magnetochron C24r (Luterbacher et al., 2000). From a biostratigraphic point of view, the P/E boundary falls within planktonic foraminiferal Zone P5 and in the upper part of calcareous nannofossil Zones CP8 and NP9.
The P/E boundary can be approximated by a series of calcareous microfossil datum levels (Fig. F4). These events include a major extinction event among benthic foraminifers at bathyal to abyssal depths (including S. beccariiformis, Aragonia velascoensis, Paralabamina lunata, Paralabamina hillebrandti, and Osangularia velascoensis), which has been correlated precisely to the base of the CIE (e.g., Thomas and Shackleton, 1996). At all Leg 208 sites, the T of S. beccariiformis occurs directly below the base of a pronounced clay layer. Our data thus do not agree with Tjalsma and Lohmann (1983), who described the last occurrence of this species as occurring earlier at abyssal depths.
The postextinction interval can be recognized by the presence of low-diversity, small, and thin-walled foraminifers on Walvis Ridge dominated by Nuttallides truempyi and various abyssaminid taxa (Boltovskoy and Boltovskoy, 1989; Müller-Merz and Oberhänsli, 1991; Thomas and Shackleton, 1996). In noncarbonate sequences, this postextinction interval can be recognized by the dominance of Glomospira spp., the so-called earliest Eocene "Glomospira event" in agglutinated assemblages (Kaminski et al., 1996; Galeotti et al., in press).
The P/E boundary can also be identified by the presence of the planktonic foraminifer "excursion fauna" that includes Acarinina africana, Acarinina sibaiyaensis, and Morozovella allisonensis (Kelly et al., 1996, 1998). The planktonic foraminifers Pseudohastigerina wilcoxensis and large specimens of Chiloguembelina wilcoxensis have lowermost occurrences (bases) close to the boundary as well (Speijer and Samir, 1997). Furthermore, the P/E boundary is typically associated with a marked increase in the relative abundance of the genus Acarinina (Kelly, 2002). The P/E boundary is followed by the T of M. velascoensis (~54.31 Ma) (Norris and Röhl, 1999).
The B of Rhomboaster cuspis, the oldest representative of the Rhomboaster-Tribrachiatus lineage, occurs at the base of the clay-rich interval. A decrease in the relative abundance of the calcareous nannofossil genus Fasciculithus together with an increase in abundance of Zygrhablithus spp. is recorded just above the P/E boundary. We do not use the B of Campylosphaera eodela, the marker for the lower boundary of Subzone CP8b, because specimens of C. eodela occur far below the lowermost occurrence of Discoaster multiradiatus (the base of Biozone CP8) (Bralower and Mutterlose, 1995). All these biostratigraphic events are present in the long interval of reversed polarity, Chron 24r. According to Cande and Kent (1995), Chron C24r has a duration of 2.557 m.y. Recent cyclostratigraphic work (Norris and Röhl, 1999; Röhl et al., 2000, 2003) indicates that the base of the CIE, and therefore the P/E boundary, occurred ~1 m.y. after the end of Chron C25n, at 55.0 Ma. We have used this age for the P/E boundary in the Leg 208 timescale.
The K/P boundary is marked by one of the largest mass extinctions in Earth history. The extinction level is about halfway through Magnetochron C29r and occurred ~250 k.y. before the magnetic reversal at the base of Chron C29n. The top of the Maastrichtian is recorded in the extinction of ~95% of planktonic foraminifers, including the globotruncanids, rugoglobigerinids, and the large serially coiled planktonic foraminifers. The nearly exclusive presence of minute planktonic foraminifers (which are typically <63 µm in diameter) is characteristic of the lowermost Paleogene (Danian) sediments above the K/P boundary. The Danian species include Parvulorugoglobigerina eugubina, Guembelitria cretacea, and Heterohelix globulosa. Previous zonations have identified Zone P0 as dominated by G. cretacea, between the K/P boundary and the lowermost appearance of P. eugubina (~64.97 Ma). However, reexamination of the classic K/P boundary section and Global Standard Stratotype Section and Point at El Kef, Tunisia, has demonstrated that rare specimens of P. eugubina are present directly above the extinction level of the Cretaceous foraminifers, suggesting that Zone P0 reflects an increase in abundance of P. eugubina rather than its evolutionary lowermost appearance (Norris et al., 1999).
Calcareous nannofossils also display a significant extinction across the K/P boundary. The lowermost Danian is characterized by an increase in abundance of Thoracosphaera spp. and the B of Biantholithus sparsus and Cyclagelosphaera reinhardtii. The K/P boundary is also bracketed by the B of Cruciplacolithus primus (64.8 Ma) and Cruciplacolithus tenuis (64.5 Ma) above the boundary and the lowermost occurrence (base) of Micula prinsii (65.4 Ma) below the boundary.
Benthic foraminifers did not suffer significant extinction at the K/P boundary but in many locations show short-lived changes in assemblage composition (Culver, 2003).
Calcareous nannofossils were examined in smear slides using standard light-microscope techniques under crossed nicols and transmitted light at 1000x magnification. The following abbreviations were used to describe nannofossil preservation:
The total abundance of calcareous nannofossils in a smear slide was estimated as follows:
Abundances of calcareous nannofossil taxa were estimated, and their abundance levels were recorded as follows:
Benthic and planktonic foraminifers were extracted from unlithified ooze by washing samples over a 63-µm sieve. More indurated samples were broken into small (<1-cm diameter) pieces, soaked in a 3% solution of hydrogen peroxide with a small amount of Calgon, warmed on a hot plate, and then washed with tap water over a 63-µm sieve. A smaller sieve size (38 µm) was used for stratigraphic intervals in which key marker taxa are diminutive (e.g., lowermost Danian and parts of the lower Eocene). All samples were dried in a low-temperature oven at ~50°C.
Species identification was generally made in the >63-µm size fraction. Because of time constraints imposed by ongoing coring, the relative abundances of foraminiferal taxa were estimated semiquantitatively. In some instances, only presence/absence data were compiled for major marker species, distinctive morphotypes, and/or dominant species in a sample.
The preservation status of the planktonic and benthic foraminifers was estimated as follows:
The abundance of planktonic foraminifers in a given sample is expressed as follows:
The abundance of planktonic and benthic foraminiferal species was expressed as follows:
The abundance of benthic foraminifers was expressed as follows (P/B = number of planktonic foraminifers/number of benthic foraminifers):