We observed well-preserved nannofossils throughout the cores (Appendix B, back pocket) except in the middle Miocene intervals where some samples contain poorly preserved nannofossils or are barren. This interval is characterized by dissolved assemblages and corresponds to the "middle Miocene Carbonate Crash" interval, which was also observed in the eastern equatorial Pacific cores (Lyle et al., 1995). Stratigraphic positions of datums recognized in the three holes are shown in Table 3, Table 4, Table 5.
The base of this zone is defined by the first occurrence (FO) of Emiliania huxleyi. Two other datums, the base of the acme of E. huxleyi and the last occurrence (LO) of Helicosphaera inversa, have not been determined. The former event was difficult to identify because the abundance of E. huxleyi increased gradually; Helicosphaera inversa was rare and had a sporadic stratigraphic occurrence in the three holes.
The top of this zone is defined by the FO of E. huxleyi, and the base is approximated using the FO of Gephyrocapsa parallela. Bukry (1973, 1975) defined the base of Zone CN14 by the FO of Gephyrocapsa oceanica, but the base of this zone might correspond to the FO of G. parallela (Takayama and Sato, 1987). The FO of G. parallela almost coincides with the re-entrance of medium-sized Gephyrocapsa (Raffi et al., 1993) and the FO of Gephyrocapsa sp. 3 (Rio, 1982); the latter taxon corresponds to "G. omega - G. parallela morphotypes" (Raffi et al., 1993). G. oceanica, used as the definition of the base of Zone CN14 by Bukry (1973, 1975), is considered to be G. parallela because Bukry (1973) reported only the presence of "G. omega" within the Gephyrocapsa oceanica Zone (CN14 by Okada and Bukry, 1980).
Two other useful datums, the LOs of Pseudoemiliania lacunosa and Reticulofenestra asanoi, lie in Zone CN14. The LO of P. lacunosa defines the boundary between Subzones CN14b (Ceratolithus cristatus Subzone) and CN14a (Emiliania ovata Subzone).
This interval lies between the LO of Discoaster brouweri and the FO of G. parallela. Some of the datums defined by Takayama and Sato (1987) can be used to subdivide this interval and to correlate between the three sites. For example, the FO of Gephyrocapsa caribbeanica (= FO of medium Gephyrocapsa spp. by Raffi et al. [1993]) defines the boundary between Subzones CN13b (Gephyrocapsa caribbeanica Subzone) and CN13a (Emiliania annula Subzone) and approximates the Pliocene/Pleistocene boundary. Thus, the Pliocene/Pleistocene boundary should be placed close to the intervals corresponding to Samples 165-998A-4H-5, 70 cm; 165-999A-4H-7, 30 cm; and 165-1000A-4H-4, 50 cm.
This zone lies between the LO of Reticulofenestra pseudoumbilicus and the LO of Discoaster brouweri. It is subdivided into four subzones by the subsequent LOs of Discoaster pentaradiatus, D. surculus, and D. tamalis (Bukry, 1973, 1975; Okada and Bukry, 1980). All of these datums have been determined in the Caribbean cores (see Table 3, Table 4, Table 5). In addition, the FO of Discoaster triradiatus, a three-rayed discoaster, lies in the uppermost part of the Pliocene (Takayama, 1970). Similarly, Backman and Shackleton (1983) used the increase in abundance of D. triradiatus relative to D. brouweri as an event in the uppermost Pliocene. This species was observed in the uppermost part of Subzone CN12d (Calcidiscus macintyrei Subzone) in the Caribbean sites, but it is too rare to determine the abundance increase event defined by Backman and Shackleton (1983).
Two other events can be identified in this zone in the Caribbean sections: the LOs of Reticulofenestra ampla and a circular form of Reticulofenestra minutula. These events were described by Sato et al. (1991) in sections from the North Atlantic and the Indian Oceans. R. ampla is a small form of Reticulofenestra pseudoumbilicus with a diameter >5 µm (Fig. 5, Fig. 6). This species is one of the "medium reticulofenestrids" defined by Flores et al. (1995) at Sites 849 and 852 in the eastern equatorial Pacific. The LO of R. ampla lies at about the same level as the LO of Discoaster tamalis, which defines the top of Subzone CN12a (Discoaster tamalis Subzone). The occurrence of the circular form of R. minutula is limited to the lower part of Subzone CN12a in the Caribbean. The size of the circular form of R. minutula is >5 µm in diameter.
The LO of Sphenolithus spp, including Sphenolithus abies and Sphenolithus neoabies, lies in the lower part of Subzone CN12a in all of the studied sections. This event, however, lies slightly above the LO of Reticulofenestra pseudoumbilicus, which defines the base of Zone CN12. This event was used by Bukry (1991) for the further subdivision of Subzone CN12a in Subzones CN12aA and CN12aB.
This zone lies between the LO of Amaurolithus spp. and the LO of Reticulofenestra pseudoumbilicus. One of the most prominent datums in the Pliocene sequence at the Caribbean sites is the LO of R. pseudoumbilicus (Table 3, Table 4, Table 5). Many authors define this event as the LO of large (>7 µm in diameter) Reticulofenestra specimens (e.g., Raffi and Flores, 1995). R. pseudoumbilicus and Sphenolithus spp. are both abundant in Zone CN11. Bukry (1973) described the beginning of abundant occurrence of Discoaster asymmetricus within Zone CN11 and used it for the subdivision of Zone CN11. We were able to recognize this datum only in Hole 998A.
This zone lies between the LO of Discoaster quinqueramus and the LO of Amaurolithus spp. Originally, Bukry (1973) used the LOs of Amaurolithus tricorniculatus and Amaurolithus primus as the boundary definition for the top of Zone CN10. In Holes 998A and 1000A, the uppermost specimens of Amaurolithus observed belong to A. primus, whereas in Hole 999A, they are A. tricorniculatus specimens.
Zone CN10 can be subdivided into three subzones at the sites investigated. The boundary between Subzones CN10c (Ceratolithus rugosus Subzone) and CN10b (Ceratolithus acutus Subzone) is defined by the FO of C. rugosus as well as the LO of C. acutus (Bukry, 1973). The FO of C. rugosus is more easily determined in the Caribbean sections, however, Raffi and Flores (1995) indicated that this event is slightly diachronous with respect to magnetostratigraphy.
C. acutus has a continuous distribution in the lower Pliocene section in the three holes and the FO of this species is used to define the boundary between Subzones CN10b and CN10a (Triquetrorhabdulus rugosus Subzone). This boundary is also defined by the LO of T. rugosus (Bukry, 1973), although this datum lies just below the FO of C. acutus in Holes 998A and 999A. Moreover, T. rugosus has a sporadic occurrence in Hole 1000A.
The Miocene/Pliocene boundary is placed within Subzone CN10a (e.g., Mazzei et al., 1979; Raffi and Flores, 1995).
This interval lies between the FO of Discoaster berggrenii and the LO of D. quinqueramus. Both species are common in upper Miocene sediments in the Caribbean. Three datums, based on the FO and LOs of species of Amaurolithus, can be used to subdivide this zone. Bukry (1973, 1975; Okada and Bukry, 1980) divided Zone CN9 into Subzones CN9b (Amaurolithus primus Subzone) and CN9a (Discoaster berggrenii Subzone) by the FO of Amaurolithus primus. Raffi and Flores (1995) used the FO of Amaurolithus spp. to subdivide this interval into two subzones. A. primus is rare in the sections studied, but the FO of Amaurolithus delicatus can be determined precisely in Holes 998A and 1000A, and thus we use this event to determine the base of Subzone CN9b. In addition, the FO and LO of Amaurolithus amplificus lie within Subzone CN9b and can be used to subdivide this subzone at the Caribbean sites into three biostratigraphic units (CN9bA, CN9bB, and CN9bC in ascending order) following Raffi and Flores (1995).
Specimens of Reticulofenestra >7 µm in diameter (R. pseudoumbilicus) are observed above the lower part of Subzone CN9b in Holes 998A and 1000A and above the middle part of Subzone CN9a in Hole 999A (Fig. 5, Fig. 6). The temporary disappearance of this taxon in the lower part of Zone CN9 and the upper part of Zone CN8, the small Reticulofenestra event or the paracme interval of R. pseudoumbilicus, has also been observed in the equatorial Indian Ocean (Rio et al., 1990; Young, 1990), in the western equatorial Pacific Ocean (Takayama, 1993), the eastern equatorial Pacific (Raffi and Flores, 1995), and the western equatorial Atlantic (Backman and Raffi, 1997). The reappearance of R. pseudoumbilicus (>7 µm in diameter) in Zone CN9 is used by Raffi and Flores (1995) to define the top of the paracme interval of R. pseudoumbilicus.
The LO of Minylitha convallis lies in Subzone CN9a in Holes 998A and 999A; this event is difficult to determine in Hole 1000A because of poor preservation.
Hole 998A: interval 165-998A-14H-3, 100 cm, to 15H-3, 100 cm
Hole 999A: interval 165-999A-26X-2, 100 cm, to 29X-1, 22 cm
Hole 1000A: interval 165-1000A-31H-6, 50 cm, to 36X-6, 50 cm
This interval lies between the LO of Discoaster hamatus and the FO of D. berggrenii. Zone CN8 is subdivided into Subzones CN8b (Discoaster neorectus Subzone) and CN8a (Discoaster bellus Subzone) by the FO of Discoaster loeblichii. This datum can be determined in Holes 998A and 999A although D. loeblichii is rare in both holes. It is hard to distinguish this species from other poorly preserved species of Discoaster in Hole 1000A.
In Holes 998A and 999A, the FO of M. convallis lies near the LO of D. hamatus. This event was recognized near the base of Zone CN8 in the eastern equatorial Pacific and the Indian Ocean (Raffi et al., 1995).
Hole 998A: interval 165-998A-15H-4, 20 cm, to 16H-5, 100 cm
Hole 999A: interval 165-999A-29X-1, 100 cm, to 29X-CC
Hole 1000A: interval 165-1000A-36X-CC, to 40X-6, 50 cm
Zone CN7 (Discoaster hamatus Zone) corresponds to the total range of Discoaster hamatus. D. hamatus is abundant in the three holes and thus its range can be determined precisely.
Bukry (1973) used the FO of Catinaster calyculus to define the boundary between Subzones CN7b (Catinaster calyculus Subzone) and CN7a (Helicosphaera carteri Subzone). However, in Holes 999A and 1000A this event lies below the FO of D. hamatus. A similar relationship was observed in the eastern equatorial Pacific (Raffi and Flores, 1995), in the western Indian Ocean (Rio et al., 1990), and in the Ceara Rise (Backman and Raffi, 1997). The LO of C. calyculus lies in Zone CN7 in all three holes, but this event is not considered to be reliable because of its sporadic occurrence.
The base of this zone corresponds to the FO of Catinaster coalitus. This species is common in Holes 998A, 999A, and 1000A, therefore, the lower limit of CN6 can be determined precisely. The LO of Coccolithus miopelagicus lies in Zone CN6 in all three holes; this datum lies just below the top of the zone in Holes 998A and 1000A.
Nannofossils are typically poorly preserved in this zone, which corresponds to the "carbonate crash" (Lyle et al., 1995).
This interval lies between the LO of Spenolithus heteromorphus and the FO of Catinaster coalitus. Bukry (1973) also used the LO of Discoaster kugleri to define the top of Zone CN5. We do not use this event as the zonal boundary for the top of Zone CN5 because the LO of D. kugleri lies below the FO of C. coalitus in the three holes studied.
The FO of Discoaster kugleri has been used to subdivide Zone CN5 into Subzones CN5b (D. kugleri Subzone) and CN5a (Coccolithus miopelagicus Subzone) (Bukry, 1973, 1975). Because D. kugleri is rare in the lower part of its range, we cannot determine its FO precisely. We can determine the abundance increase of D. kugleri, an event that has been used by Raffi and Flores (1995).
Other stratigraphically useful bioevents in Zone CN5 in the Caribbean sites include the LOs of Cyclicargolithus floridanus and Coronocyclus nitescens. The LO of C. floridanus was defined by Bukry (1973) as an alternate marker for the base of Subzone CN5b. However, this event lies in Subzone CN5a, just above the base of Zone CN5. This event was also recognized in higher levels in the North Atlantic (Takayama and Sato, 1987; Gartner, 1992). The LO of C. nitescens was shown to be a useful event by Gartner and Chow (1985) and Fornaciari et al. (1990). This datum lies below the abundance increase of D. kugleri in Holes 998A and 999A, but it cannot be determined in Hole 1000A because of poor preservation.
This interval lies between the LO of Helicosphaera ampliaperta and the LO of Spenolithus heteromorphus. Both datums can be determined precisely in the three holes. The LO of Discoaster deflandrei lies in this zone. The abundance of Discoaster signus and D. variabilis increases abruptly in the upper part of CN4.
This interval lies between the FO of Sphenolithus heteromorphus and the LO of Helicosphaera ampliaperta and can be recognized in Holes 998A and 999A. Bukry (1973, 1975, 1978) suggested the use of the FO of Calcidiscus macintyrei to define the boundary between Zones CN4 and CN3. However, this species is completely absent from both zones in the Caribbean sections. Another datum used to define the upper boundary of CN3 by Bukry (1973, 1975) is the top of the acme of D. deflandrei. However, this event is highly dependent on preservation and lies below the LO of H. ampliaperta in the three holes studied, the same as observed in the Atlantic (e.g., Olafsson, 1991; Gartner, 1992) and the eastern equatorial Pacific (Raffi and Flores, 1995).
This interval lies between the FO of Spenolithus belemnos and the FO of S. heteromorphus. Discoaster deflandrei, Coccolithus miopelagicus, Cyclicargolithus floridanus, and Sphenolithus moriformis are common in this zone.
This interval lies between the LO of Reticulofenestra bisecta and the FO of Sphenolithus belemnos. Because Discoaster druggii and Cyclicargolithus abisectus have sporadic distributions in the Caribbean sections, we cannot subdivide this zone into subzones as proposed by Bukry (1973, 1975). The LOs of Triquetrorhabdulus carinatus and Sphenolithus delphix can be determined precisely in Hole 998A.
The LO of Sphenolithus delphix lies close to the Oligocene/Miocene boundary (Fornaciari and Rio, 1996).