The fossiliferous levels of the Guadix–Baza Depression are situated in the SW sector. According to Martín Suárez et al. (2012), we use the term “Guadix–Baza Depression” instead of “Guadix–Baza Basin” because the palaeogeographic configuration of the “Guadix–Baza Basin” (sensu stricto) started in the Late Tortonian (Soria et al., 1999 and Viseras et al., 2004, and references therein), and earlier deposits correspond to a different palaeogeographic context.

The fossil locality of La Peza-4A (PEZ-4A) is located in a dark lutite level having a thickness of about 60 cm at the base of a formation mainly constituted by lacustrine limestones with lutite levels (Fig. 1). These limestones show fenestral porosity, which can be related with shallow lake level stages. These limestones alternate with continental lutites (with aquatic gastropods) and some lignite levels. Above this limestone unit, a thick lutite deposit can be observed. The transition between this lutite formation and that of the limestones is progressive, with alternation of lutite levels between the limestones. On top of the limestone unit, continental lutites crop out, and these change progressively into marine marls. Although Vera (1968) suggested a Burdigalian age with much doubts, Navarro-Vilá and García Dueñas (1977), using stratigraphic criteria, suggested a Serravallian age for these deposits. The succession clearly shows a shallowing trend (from lutites to porous limestones), and afterwards a deepening trend (from porous limestones to continental lutites and finally marine marls).

The Otura section is situated in the eastern sector of the Granada Basin (Fig. 1). This section has yielded three different fossiliferous levels with a Latest Tortonian-Earliest Messinian age: OTU-4 (Latest Tortonian), OTU-1 and OTU-3 (Earliest Messinian) (García-Alix et al., 2008). The chronology of this section was based on small fossil mammals (García-Alix et al., 2008). The studied specimen appeared in OTU-1 (Earliest Messinian age). The predominant sediments in this section are fine sands and some intercalated lutites, frequently with lignites. These deposits are related with a braided system, and the fossil levels are located in lutites and lignitiferous levels of the floodplain area.

The studied fossil localities (PEZ-4A and OTU-1) were intensively sampled to collect small mammal remains, and about 200 and 1440 kg of sediment were extracted and processed respectively. Sediment was washed and sieved with different mesh sizes (2.5 mm, 0.7 mm, and 0.5 mm). After that, bone and tooth remains were manually collected under a binocular microscope. The teeth of Cyprinus were finally examined and described. A scanning electronic microscope was used for preparing the pictures.

In the recent Cyprinus carpio (Linnaeus), each pharyngeal bone bears five teeth distributed in three rows (Fig. 2 and Fig. 3). The main row, or inner row, is composed of the largest teeth. The anterior one (noted 3 by Rutte, 1962), which is cylindrical, ends distally in a hook. The second one (noted 2), which is molariform, is flattened posteriorly. Its grinding surface is typically ornamented by three or four concentric scalloped grooves. The posterior tooth (noted 1) is peg-shaped. The single tooth of the middle row (noted 1a) is quite smaller and shorter than the teeth of the mesial row. Its grinding surface is ornamented by three or four transverse grooves. The single tooth of the lateral row (noted 1b) is still smaller and shorter; when it is not worn, three transverse grooves are present on its grinding surface.

The largest and most significant fossil tooth is that from a left pharyngeal bone, which has been found in the Granada Basin, in the lower lignite level (OTU-1) of the Otura section (Fig. 4). Its crown measures 3.15 mm transversally and 2.2 mm longitudinally. Its more or less oval grinding surface is ornamented by three transversal wavy grooves that are arched backwards. This tooth is very similar to the median tooth of the main (inner) row (noted 2).

The two smaller fossil teeth (PEZ-4A) have somewhat different shapes: one, which has a quadrangular crown, has a length of 1.45 mm and a width of 1.3 m. Its grinding surface is ornamented by two transversal grooves: one slightly convex forward, and the second almost straight (Fig. 5). The second tooth has a subcircular section. Its grinding section is ornamented with two transverse parallel wavy grooves. These two teeth may be interpreted as being the single tooth of the median row (noted 1a).

The fact that the fossil pharyngeal teeth described above do not significantly differ from those of the common carp suggests that it is not impossible that they may have belonged to the recent species Cyprinus carpio Linnaeus. However these remains are too much fragmentary to ascertain this hypothesis.

The find of three pharyngeal teeth of the genus Cyprinus Linnaeus in the Upper Miocene of South Spain has a considerable biogeographical significance if we consider the recent distribution of this genus. In fact, the carp, Cyprinus carpio Linnaeus, 1758, is presently a widespread Eurasiatic teleost species which, in the native state, is present in the “Rivers of the Pacific Ocean basin and in East Asia from the Amur to and Burma” (Myanmar), and also in the “basins of the Black, Caspian and Aral Seas” and, eventually, “partly perhaps western Europe” (Berg, 1949/1964: 835). Additionally, because of domestication, it is also widely distributed in ponds throughout the world for the purpose of aquaculture.

In the fossil record, the genus Cyprinus Linnaeus is rather poorly known. Sytchevskaya (1989) described as C. mediocris two pharyngeal teeth from the Upper Miocene of central Asia and the subspecies: Cyprinus carpio altaicus Sytchevskaya, based on two incomplete pharyngeal bones from the Upper Pliocene of the Altaï territory (Siberia). Remains of Cyprinus carpio were also identified by Tarashchuk (1962: 17) in the Late Pliocene of the Kherson oblast (Ukraine), whereas Kovalchuk (2013) extended recently its stratigraphical occurrence to the Upper Miocene of Ukraine.

Although it is unknown in the fossil record before the Pleistocene in central and western Europe, the genus Cyprinus Linnaeus was several times wrongly reported from the Oligo-Miocene of Germany. The first quotation was from Bronn (1828), who called Cyprinus papyraceus a small cyprinid fish from the Late Oligocene of the Siebengebirge, near Bonn (Germany), which was soon after referred by Agassiz (1839) as belonging to the genus Leuciscus Cuvier. Later, von Meyer (1850) created the species Cyprinus priscus for fishes from the Lower Miocene of Unterkirchberg (Württemberg, Germany). The same species was later identified by Weitzel (1933) in the Middle Miocene of Beuern (Hessen, Germany). However, Obrhelová (1970) rightly suggested that these teleosts are really directly related to the recent Crucian Carp, Carassius carassius (Linnaeus) and belong to the recently created genus Palaeocarassius Obrhelová.

Consequently, the genus Cyprinus Linnaeus is presently unknown in the European fossil record before the Pleistocene in the countries westward of Ukraine where its distribution in the fossil state has been recently re-examined by Kovalchuk (2013) who noted the occurrence of Cyprinus cf. carpio Linnaeus in the Lower Turolian (MN 11 mammal zone) of Lysa Gora-2 (Zaporizhzhia oblast) and that of C. carpio Linnaeus, in the Late Pliocene of the Kherson oblast, previously reported by Tarashchuk (1962). All the other Ukrainian occurrences of the genus Cyprinus Linnaeus are either Pleistocene (such as Lysa Gora-1: Rekovets et al.,2014) or Holocene in age.

In central and western Europe, the genus Cyprinus was only correctly reported from the Pleistocene (undetermined ante-Mindel warm period) of Stuttgart-Bad Cannstadt, Germany (Böttcher, 1994). However, it cannot be totally excluded that Cyprinus remains may exist in hitherto unknown Cenozoic localities of central and western Europe.

The find of pharyngeal teeth of the genus Cyprinus in the Late Miocene of Andalusia raises a quite unexpected palaeobiogeographical problem for explaining their occurrence in South-West Europe. In fact, it can be hardly admitted that this genus was widely distributed in central and western Europe during the Miocene and that its geographical distribution area became subsequently disconnected. The reason is that the Miocene freshwater teleosts of central and western Europe have been intensively studied during the last fifty years, so that many localities of articulated skeletons are known in the Miocene of Bulgaria (Gaudant and Vatsev, 2006 and unpublished data), Serbia (Gaudant, 1998), Hungary (Gaudant, 2005), Austria (Böhme, 2002, Gaudant, 1994, Gaudant, 2000a and Gaudant, 2010), Czech Republic (Obrhelová, 1967, Obrhelová, 1969, Obrhelová, 1970 and Obrhelová, 1990), Germany (Böhme, 2010, Gaudant, 1980 and Gaudant, 1989, Gaudant and Reichenbacher, 1998, Keller et al., 2002, Reichenbacher, 1988 and Reichenbacher, 2000), Switzerland (Gaudant et al., 2002 and Reichenbacher, 1993) and France (Gaudant, 2000b and Mein et al., 1983) but none of them has yielded the slightest remain of the genus Cyprinus. Additionally, no pharyngeal tooth of Cyprinus has ever been found when washing and screening lacustrine sediments of central and western Europe for collecting vertebrate remains.

An alternative, highly improbable, hypothesis might have been that the carps that have been living in Andalusia during the Late Miocene and had been able to penetrate from North Africa into South-West Europe, like the characiforms and channids that have been identified in the Middle Miocene of Sansan, a well-known mammalian fossiliferous locality of the northern piedmont of the Pyrenees. However, no Cyprinus remain, has ever been reported from the Neogene of North Africa. Moreover, the genus Cyprinus is presently unknown in the native state in North Africa. Consequently, the palaeobiogeographical significance of the occurrence of this genus in the Late Miocene of Andalusia remains presently unexplained.