The geographic and stratigraphic distribution of the oysters described above allows some interesting, although as yet preliminary, paleobiogeographic conclusions to be made. In the first place, as observed in figure 9, Turkostrea damboreneae nom. nov. And Cubitostrea primordialis n. sp. occur considerably earlier in southern South America than their close relatives from other parts of the world. Clear morphological differences in the shell show that by the late Maastrichtian they were already well established and that the origin of the Turkostreini should probably be sought for in slightly earlier rocks. Where exactly this could have taken place must necessarily be a matter of speculation, but some clues may be drawn from their presence in northern Patagonia in such comparatively old rocks. In fact, both genera are well known in younger rocks in other parts of the world, particularly the northern hemisphere. Cubitostrea seems to have been quite common throughout the Eocene in the Atlantic and Gulf coast of the United States, while occurring as well in the European Eocene. Turkostrea , although common in central Asia, from where it was originally described, is particularly abundant in the Eocene of northern Africa and Europe (Strougo, 1973, 1976, and references therein), while occurring also in Texas, Trinidad and Mexico according to Perrillat and Vega (1993), and possibly in Jamaica (Trechmann, 1923; Strougo, 1976). In fact, in northern Africa it is known from Paleocene rocks exposed in the Western Desert of Egypt, the age being confirmed by planktonic forams (Luger, 1985; Malchus, pers. com.). In South America, this genus has not been described previously, and the two species described here, although quite common in certain units and localities of the Neuquén Basin, have been largely overlooked. There can be no doubts as to the generic identity of our material and, in such a case, a southern origin for this group of oysters should be seriously considered.
In the case of Cubitostrea , the earliest recorded species of the genus appears also in Maastrichtian rocks in northern and central Patagonia. Cubitostrea primordialis n. sp. clearly belongs in that genus, becoming the oldest species presently identified. Like in the case of Turkostrea , this genus is also quite common in the Northern Hemisphere, particularly in Eocene rocks of Europe (Karagyuleva, 1961; Malchus, 1995) and the Gulf coast and eastern seaboard of the United States (Stenzel, 1949; Dockery and Nystrom, 1992; Willoughby and Nystrom, 1992). This genus shares many characters with Turkostrea and was therefore placed in the Turkostreini by Malchus (1990, p. 196). It seems likely that one of the two genera is ancestral to the other or else that they share a common ancestor not too far back in the Cretaceous. Cubitostrea primordialis n. sp. is, again, the earliest record of Cubitostrea .
While common in Paleocene rocks of Patagonia and western Argentina (Casadío, 1998, and references therein), Cubitostrea seems to be absent in Eocene and Oligocene rocks here. This apparent absence is probably a consequence of various concurrent factors. In the first place, the Eocene is very scantily represented in Argentina. The only reasonably good exposures are those of the Río Turbio and Man Aike Formations, in addition to the poorly fossiliferous Leticia Formation exposed off the coast of Tierra del Fuego. In none of the three units has material referable to Cubitostrea -or Turkostrea - been collected so far, despite the abundant mollusk fauna (including oysters) in some of them. A second cause for the lack of these two genera in Eocene or Oligocene rocks may be the particular ecological requirements that they may have had and about which we can only speculate. The third reason - i.e ., the actual disappearance of Cubitostrea in southern latitudes at the close of the Paleocene while surviving in younger rocks elsewhere- seems likely as well were it not for the fact that species of Cubitostrea , one of them very similar to the Paleocene species, appear again in the widespread early and late Miocene rocks exposed throughout Patagonia and central Argentina (Entre Ríos). In addition to a non described Cubitostrea from the early Miocene Chenque Formation (Bellosi, 1995) in the area surrounding Comodoro Rivadavia (central Chubut), Cubitostrea alvarezii (d'Orbigny, 1842) (d'Orbigny, 1842, p. 134, pl. 7, fig. 19; other references) appears in the late Miocene Puerto Madryn Formation exposed in the area around Peninsula Valdés in northern Patagonia and in the equivalent Paraná Formation in western Entre Ríos (del Río and Martínez, 1998 and references therein). It is also known from several drillings on the Pampas of central Argentina (Wahnish, 1939; Carral Tolosa, 1942; Camacho, 1967). These Miocene occurrences, as young as Tortonian for Cubitostrea alvarezii (Scasso et al ., 2001), are the youngest for any member of this genus in South America. Thus, the time range of Cubitostrea in South America spans the Maastrichtian through the Miocene, albeit the absence of records in the Eocene-Oligocene. Consequently, in this region the Turkostreini appear to have survived at least up to the late Miocene. Malchus (1995, p. 230) stated that Cubitostrea ? latimarginata (Vredenburg, 1908) from Asia and Saudi Arabia and Cubitostrea ? coxi (Gardner, 1945) from Florida were two Miocene species that could be placed in this genus. He also suggested the possibility that the Cubitostrea lineage may be extant in the Indo-West Pacific and Mediterranean.
Summarizing, the available data point towards a southern origin for Cubitostrea and Turkostrea , a fact that would render South America the place or origin of the entire tribe Turkostreini. The presence and abundance of members of this taxon in younger seas of the northern hemisphere must thus be accounted for. In connection with this, the Tethyan affinities of other marine groups during the Maastrichtian-Danian have been documented (Camacho, 1992; Feldmann et al ., 1995; Casadío, 1998). Such affinities, revealed in this case by the oysters of this particular group, are reinforced by the fact that oysters belonging in the Turkostreini have not been recorded from either Maastrichtian, Danian or Eocene (or younger) age in other areas of the southern hemisphere. This is particularly striking in the case of the extremely abundant and well preserved Maastrichtian-Eocene mollusk faunas from Marambio (Seymour) Island in Antarctica (Sharman and Newton, 1894, 1897; Wilckens, 1910, 1911; Zinsmeister, 1984; Zinsmeister and Macellari, 1988; Stilwell and Zinsmeister, 1992) -and also in the case of the Eocene fauna from MacMurdo described by Stilwell (2000)- which have yielded not one specimen that could be even tentatively placed either in Turkostrea or Cubitostrea . Turkostrea also appears to be absent from the very diverse and well known New Zealand Tertiary faunas, which share some particular groups in rocks ranging from the pre Cenozoic to the Pliocene and in some instances share taxa of the present day faunas (Beu and Griffin, 1997; Beu et al ., 1997). Cubitostrea is represented there by the Eocene Cubitostrea gudexi (Suter, 1917), but the generic placement of this species is doubtful, as it does not seem to be a true Cubitostrea and does not resemble the South American species at all.
If the two genera (or either one of them) appeared first in west-central Argentina, then the northward migration must be explained, of course considering the currently accepted patterns of landmass and current distribution in the Atlantic Ocean during the end of the Cretaceous and beginning of the Cenozoic. Such an explanation, as yet premature, would probably be of great heuristic value when dealing with other members of the shallow marine fauna inhabiting the southernmost tip of the American Continent.
This study was partially supported by Universidad Nacional de La Pampa and Fundación Antorchas to S.C. and CONICET to M.G. We also thank N. Malchus and M. Machalski for constructive comments. M. Sander kindly allowed access to Fritzsche's material housed in the Goldfuß-Museum, Bonn. A.C. Riccardi granted access to the material housed in the Museo de La Plata, La Plata, Argentina.