Results are presented of a quantitative biostratigraphic study of the calcareous nannofossils from the upper Miocene Metochia section (Gavdos Island, Greece) located in the eastern Mediterranean. The study allows the completion of the calcareous nannofossil Zonal Scheme for the Mediterranean Neogene. Parts of the scheme have already been proposed earlier for the Pliocene-Pleistocene (Rio et al. 1990b) and for the latest Oligocene to early late Miocene (Fornaciari and Rio 1996, Fornaciari et al. 1996).

The Metochia section proved to be a suitable section for high-resolution biostratigraphic studies. It has a detailed planktonic foraminiferal biostratigraphy and a robust chronology based on the astronomical calibration of the sedimentary cycles (Krijgsman et al. 1995, Hilgen et al. 1995). Age estimates for the nannofossil biohorizons recognized in the section were obtained through direct correlation to this astrochronology. Five out of seventeen biohorizons were selected for biostratigraphic purposes and to define the (sub)zonal boundaries. They are (in stratigraphic order): LO (Last Occurrence) of Discoaster hamatus Martini and Bramlette 1963, equivalent to the base of Subzone MNN10a, dated at 9.53 Ma; FCO (First Consistent Continuous Occurrence) of Discoaster pentaradiatus (Tan 1927) Bramlette and Riedel 1954, equivalent to the base of Subzone MNN10b, at 9.37 Ma; PB (Paracme Beginning) of Reticulofenestra pseudoumbilicus (Gartner 1967) Gartner 1969, equivalent to the base of Zone MNN11a, at 8.76 Ma; FO (First Occurrence) of Amaurolithus spp. (A. primus (Bukry and Percival 1971) Gartner and Bukry 1975), equivalent to the base of Subzone MNN11b, at 7.42 Ma, and FO of Nicklithus amplificus (Bukry and Percival 1971) Raffi Backman and Rio 1998, equivalent to the base of Subzone MNN11c, at 6.69 Ma. We compared our biochronology with data available from other Mediterranean sections (Sprovieri et al. 1996, Negri et al. 1999, Negri and Villa 2000, Bonomo 2001), and with the nannofossil biochronology from low-latitude oceanic reference sections (Raffi et al. 1995, Backman and Raffi 1997). Our results evidence differences and similarities: some of the recently proposed events in the Tortonian-Messinian (e.g., FO and FCO of Reticulofenestra rotaria Theodoridis 1984, FO of Amaurolitus delicatus Gartner and Bukry 1975) proved not to be reliable and do not improve the biostratigraphic resolution by calcareous nannofossils in the time interval under consideration. On the other hand, the five selected biohorizons can easily be reproduced elsewhere in the Mediterranean, and are considered useful for correlations to the extra-Mediterranean realm.