Table 3 lists several apparently diachronous datums at Site 999, which casts doubt on the reliability of the ages of the marker species used in construction of the age model. Until independent means of dating the record are used (e.g., complete magnetostratigraphy, isotopic stratigraphy, multiple radiometrically dated ash layers), the age model presented here can only be regarded as provisional. The diachronies are generally on the order of 105 years, which permits some broad conclusions about the relationship between the timing of biotic events at this site and paleoceanographic and tectonic events in the Caribbean region.
The gradual shoaling of the Central American Isthmus and the consequent diversion of the Caribbean Current (= North Equatorial Current) northward into the Gulf of Mexico and discontinuation of surface-water flow from the eastern Pacific into the Caribbean seems to have left several marks on the biostratigraphic record of Site 999.
Left-coiling Neogloquadrina pachyderma were present at several Leg 165 sites during the late Miocene (~6.5-5.6 Ma) (Sigurdsson, Leckie, Acton, et al., 1997). The presence of sinistrally coiled Neogloboquadrina pachyderma at a tropical site is unusual in itself. The fact that these specimens sometimes constituted a significant portion of the assemblage suggests the presence of unusually cold water, probably on a seasonal basis (Keigwin, 1982). Further investigation is needed to first establish the existence and then determine the cause of this late Miocene upwelling in the western Caribbean. Such research may help to better constrain the timing of tectonic uplift in the Central American region.
Globoconella miozea, Globoconella panda, and Globoconella praescitula were encountered in the fine fraction and more rarely in the coarse fraction (>125 µm) of some samples at Site 806 (Chaisson and Leckie, 1993), but the specimens at Site 999 (particularly Gc. panda) are often quite large (>315 µm), which (along with their relative abundance) suggests they were well within their limits of environmental tolerance. Their presence therefore suggests that the surface waters of the western Caribbean Sea were connected to the cooler waters of either the California or Peru Current system (depending on the contemporary position of the intertropical convergence zone) until at least the end of the middle Miocene. Norris (1998) also found temperate latitude globoconellids at Leg 159 sites in the eastern equatorial Atlantic where the Benguela Current could have brought them up the east coast of Africa.
It is generally known (Lamb and Beard, 1972; Stainforth et al., 1975; Kennett and Srinivasan, 1983) that several species of Menardella are confined to the tropical Atlantic. The timing of the first appearances of these species, beginning at 4.77 Ma (date at this site; 4.45 Ma at Ceara Rise) with the FO of Menardella exilis, follow a period of reorganization of carbonate deposition in the eastern Pacific (Farrell et al., 1995) and the Atlantic (Haug and Tiedemann, 1998) that has been linked to the closing of the Central American Seaway and isolation of the tropical Atlantic. Isolation is an important factor in the development of endemic species.
Several species of planktonic foraminifers were absent from Atlantic sites, including Site 999, for much of the Pliocene. The best documented members of this group are the pulleniatinids, which were absent between 3.5 and 2.3 Ma (Bolli and Saunders, 1985). Globorotalia tumida was largely absent at Ceara Rise between 3.7 and 2.1 Ma (Chaisson and Pearson, 1997) and between 5.6 and 2.0 Ma at Site 999. Globorotaloides hexagona was absent from the Site 999 record between 3.1 and 1.4 Ma. At Ceara Rise this species was largely absent between 3.1 and 1.5 Ma, but was found in one sample in that interval. Norris (1998) reports that it is absent from a similarly bracketed interval in the eastern tropical Atlantic at Site 959 (Leg 159). Finally, there are no sinistrally coiled Menardella menardii at either Ceara Rise or Site 999 through this period, although they are found at Pacific sites at this time (Chaisson, 1996).
A less pervasive phenomenon than the "Atlantic hiatus" is the very early (deep stratigraphic) occurrence of several species at Site 999, which suggests that they may have evolved in the western Caribbean and then spread to other regions. The most significant members of this group are Globigerinoides extremus and Candeina nitida. Both of these species have published dates of first appearances at 8.1 Ma, but they appear at depths corresponding to much greater ages in the Site 999 (and Site 1000; Sigurdsson, Leckie, Acton, et al., 1997) record.