The nomenclature used in the descriptions of the teeth of Sciuridae is that of de Bruijn [5] and Cuenca [9]; length and width have been measured as defined by van de Weerd [22] and amended by Reumer and van den Hoek Ostende [18]. In all cases, length and width are perpendicular. The measurements followed by a star are estimates because of the poor preservation state of the teeth.

Measurements were taken with a Wild M7S binocular microscope, equipped with a Sony Magnescale LM12 digital measuring device. The measurement data were processed by a computer program provided by the last author. Measurements are given in units of 1 mm. Photographs were made with the FEI ESEM QUANTA 400 of the ‘Centro Andaluz de Medio Ambiente’ (Granada, Spain). The specimens are kept in the Departamento de Estratigrafía y Paleontología de la Universidad de Granada, Spain.

Heteroxerus Stehlin and Schaub, 1951 is a medium-sized Xerini. It differs from Atlantoxerus Forsyth Major, 1893 mainly because the teeth of Atlantoxerus are larger than those of Heteroxerus are, and because the lower teeth of Heteroxerus usually have an anterior cingulum. Heteroxerus can be also distinguished by the absence of a mesostyle (except for some rare specimens of H. grivensis), which is usually present in Atlantoxerus, except in A. adroveri. The species of Atlantoxerus have higher crowns and cusps, more rounded conules and wider crests than Heteroxerus [9].

The teeth of Pliopetaurista Kretzoi, 1962 are large with sinuous crests; rugose surfaces are frequent. The upper teeth have a very well developed metaconule and metaloph and protoloph converge towards the protocone [15].

Order Rodentia Bowdich, 1821

Family Sciuridae Fischer von Waldheim, 1817

Subfamily Sciurinae Fischer von Waldheim, 1817

Genus Heteroxerus Stehlin and Schaub, 1951

Species Heteroxerus mariatheresae Adrover, Mein and Moissenet, 1993 (Fig. 2A)

Locality: JUN-2B

Material: 1 M^1–2

Description: M^1–2 (1.83 × 2.33) with rectangular outline. The parastyle is slightly inflated. There is no mesostyle. Protocone and paracone are connected by the protoloph. There is no protoconule sensu stricto. Metacone and a large metaconule are connected by the metaloph, and the metaconule is in weak contact with the posteroloph. There are three roots.

Discussion: Molars of Atlantoxerus are larger than the JUN-2B specimen is, and they usually have a mesostyle, except for Atlantoxerus adroveri (de Bruijn and Mein, 1968); however, A. adroveri differs from our specimen by its large size and by its outline. Heteroxerus is poorly represented since the Vallesian. The M^1,2 of Heteroxerus grivensis (Forsyth Major, 1893), the most similar Vallesian species, has usually a metaconule that is less developed than in our specimen, because the metacone is, in general, larger than the metaconule (in contrast with our specimen). The metaconule is frequently reduced to an inflated end of the metaloph in H. grivensis. Our specimen has no mesostyle, which is sometimes present in the M^1–2 of H. grivensis. The parastyle is slightly more developed in our specimen. The outline of our M^1–2 is more rectangular than that of H. grivensis.

Another Vallesian species, H. molinensis Lacomba, 1986, is larger than our specimen is, and the anterior contour of its M^1–2 is different, due to the very lingual position of the anteroconule; that cusp is situated more labially in our M^1–2.

The shape and the size of our specimen are within the variation range of H. mariatheresae from the Turolian of Teruel (La Gloria-5). It resembles H. mariatheresae, among other features, in the degree of development of the metaconule, in the position and development of anteroconule and anteroloph, and in the absence of mesostyle and protoconule. Therefore, it is determined as Heteroxerus mariatheresae.

Genus Atlantoxerus Forsyth Major, 1893

Species Atlantoxerus margaritae Adrover, Mein and Moissenet, 1993 (Fig. 2B–D)

Localities: Purcal-13 (PUR-13) and Gorafe-1 (GOR-1)

Material: 1 P₄, 1 M_1–2 and 3 M^1–2

Description: In the locality PUR-13, we find:

P₄ (2.49 × 2.03): Trapezoidal outline, wider posteriorly than anteriorly. Protoconid and metaconid are connected by a posterior metalophid and separated anteriorly by a narrow groove. The posterolophid is in contact with the entoconid, but the short entolophid is not. Two roots.

M_1–2 (2.68 × 2.73): Worn and digested specimen. Sub-squared outline. The four main cusps and the hypoconulid are well developed; the mesoconid and the anteroconulid are less developed. The metalophid is continuous and low. There is a small depression enclosed by the anterolophid and the metalophid. There is a vestigial mesostylid. Four roots.

M^1–2 (2.31 × 2.96*; 2.46 × 2.98): Sub-rectangular outline. The four main cusps, the anteroconule and the metaconule are well developed. The anteroconule is labially displaced and not aligned with protocone and hypocone. Metaconule and metacone have the same size. Metaconule and posteroloph are in weak contact. There is no protoconule. The complete posteroloph is parallel to the metaloph. There is a little inflation in the mesostylar crest, forming a reduced mesostyle. Three roots.

Discussion: The M^1–2 from GOR-1 (2.63 × 3.15) is very similar to the PUR-13 specimens; it only differs in being a little larger, and having a medium-sized mesostyle.

The studied specimens are larger than those of Heteroxerus, A. cuencae Aguilar, Calvet and Michaux, 1995, and A. martini Aguilar, 2002, and smaller than Atlantoxerus huvelini Jaeger, 1977.

The size of the specimens is close to the lower values of Atlantoxerus tadlae (Lavocat, 1961) from the Middle Miocene locality of Beni-Mellal [12], but the mean values are smaller. Our specimens have a complete metalophid in the lower molars, whereas it is interrupted in A. tadlae; the posteroloph is weakly connected to the metaconule in our M^1–2, and isolated in A. tadlae.

Peláez-Campomanes and Aguilar [2] and [16] considered Atlantoxerus idubedensis Cuenca, 1988 to be a synonym of A. blacki (de Bruijn, 1967); the teeth of the populations that were ascribed to A. idubedensis are smaller than ours. The size of our teeth is close to A. blacki (not considering A. idubedensis), and A. rhodius de Bruijn, Dawson and Mein, 1970. Our lower teeth have a complete metalophid, which forms a closed depression together with anterolophid and anteroconule; the complete metalophid and this closed depression are not present in A. blacki. In our material, the valley of the talonid is less developed and the crests are wider than in A. rhodius; our P₄ has a metalophid and an entolophid, which are not observed in A. rhodius.

Our material resembles most Atlantoxerus adroveri and A. margaritae. The size of the specimens from PUR-13 is within the upper range of A. adroveri from its type locality [8] and of the most abundant samples of this species studied by van de Weerd [22] (Los Mansuetos and Concud-3, among others), except for the length of P₄, which exceeds the values given by van de Weerd [22]. A. adroveri from the type locality Los Mansuetos [8] is, in general, smaller than the M^1–2 from GOR-1; our single specimen is longer than those studied by van de Weerd [22], and its width agrees with the upper range of the M^1–2 from Teruel [22]. The size of the specimens from PUR-13 and GOR-1 agrees with that of A. margaritae. The posterolophid of the P₄ of A. adroveri does not join the entoconid [8], so it has a lingually open valley between entoconid, entolophid and posterolophid, whereas in our material, and in the three P₄ of A. margaritae from Teruel [1], the posterolophid is in contact with the entoconid, and closes this valley. There is no mesostylid in the M_1–2 of A. adroveri [8], but our specimens and those of A. margaritae from Teruel (17 specimens) do have a small one [1]. The metalophid of the M_1–2 of A. adroveri is interrupted [8], so the depression between anterolophid and metalophid is open; in our specimens and in those of A. margaritae from Teruel [1], the metalophid is continuous and this depression is closed. Our P₄ and those of A. margaritae from Teruel (three specimens) have a reduced and wide entolophid [1]. The upper molars of A. margaritae from Teruel [1] and ours have a mesostyle or mesostylar crest, which is absent in A. adroveri [8]. The comparison with A. margaritae from La Gloria-4 is based on material deposited in the Departamento de Estratigrafía y Paleontología of Granada University. On the basis of these features, we determine our molars as A. margaritae.

Species Atlantoxerus cf. margaritae Adrover, Mein and Moissenet, 1993 (Fig. 3)

Locality: Purcal-4 (PUR-4)

Material: 1 M_1–2, 2 D⁴, and 2 M^1–2

Description of the material from PUR-4: M_1–2 (2.34 × 2.70*). One can only recognize the four main cusps and the shallow depression of the talonid, because the specimen is much worn. There is a reduced mesostylid and four roots.

D⁴ (2.69 × 2.26): The parastyle is separated from paracone and protocone by a wide valley. The protocone is large, connected to the paracone by a long protoloph. There is a mesostyle. Metacone and metaconule are separated, or interconnected by a low crest, and each of them is in contact with the posteroloph through a low crest. Roots are not preserved.

M^1–2 (2.38 × 2.85; 2.35 × 2.89*): Subrectangular outline. One specimen is worn, the other one digested. There is a large mesostyle and a large and lingually displaced protoconule. A large metaconule seems to connect with the posteroloph by a low crest. Three roots.

Discussion: Our specimens are larger than those of Heteroxerus, A. cuencae and A. martini are, and smaller than those of Atlantoxerus tadlae and Atlantoxerus huvelini are.

Atlantoxerus blacki, A. rhodius and A. adroveri are similar in size to our material; the size of the specimens from PUR-4 is in the lower range of A. adroveri; our specimens are similar in size or slightly smaller than those of A. rhodius. The populations attributed to A. idubedensis by Cuenca, 1988 have smaller teeth than ours. The main differences are found in M^1–2: our specimens have a very well developed protoconule, which is displaced forward, whereas, in the other species, it may be reduced or absent, and not displaced forward; our specimens also have a large mesostyle.

The upper molars of A. margaritae have the same shape as our specimens (M^1–2 subrectangular with parallel anterior and posterior borders), a mesostyle, and a large metaconule. The M_1–2 has a reduced mesostylid too. However, our specimens differ from A. margaritae, and from the other species mentioned above, by a very well developed and forward displaced protoconule; moreover, the D⁴ from PUR-4 has a more prominent parastyle than in A. margaritae. This parastyle is similar to that of the D⁴ of Atlantoxerus cf. margaritae II from Villalba Alta [1]. Our specimens are smaller than those of A. margaritae from La Gloria-4 [1], PUR-13 and GOR-1, and similar to those from Orrios-1, Orrios-B, and La Judería [1]. The comparison with A. margaritae from La Gloria-4 is based on material deposited in the Departamento de Estratigrafía y Paleontología of the Granada University. The specimens from PUR-4 resemble A. margaritae, but the material is scarce and badly preserved; we identify it as Atlantoxerus cf. margaritae.

Species Atlantoxerus sp.

Material: 1 D⁴ and 1 M^1–2

One D⁴ from PUR-24A (2.17 × 2.31) and one not measurable fragment of M^1–2 from AGU-1C cannot be identified at the species level.

Subfamily Pteromyinae Brandt, 1855

Genus Pliopetaurista, Kretzoi, 1962

Species Pliopetaurista pliocaenica (Depéret, 1897) (Fig. 4A–E)

Localities: Purcal-24A (PUR-24A), Purcal-25 (PUR-25) and Purcal- 4 (PUR-4).

Material: 1 P₄, 1 M_1–2, 1 D⁴, 3 M^1–2 and 1 M³

Description: In PUR-24A, we have found one M³ (3.13 × 2.84). A weak anteroloph connects anterocone with paracone. The parastyle is absent. Paracone and protocone are very well developed. The high protoloph delimits two closed depressions with rugose surface. There is a weak mesostyle. The endoloph has a rugose lingual face and connects anteroconule, protocone, and hypocone. Three roots.

PUR-25 has yielded a non-measurable fragment of M_1–2; one may observe the protoconid, mesoconid, hypoconid, part of the entoconid and a large mesoconid. There is no hypoconulid. The bottom of the central depression is rugose. Roots are not preserved.

In PUR-4, we have found one fragment of P₄, one D⁴, and two M^1–2.

P₄: Only the posterior part is preserved. The bottom of the central depression is rugose. There is a vestigial hypoconulid. The posterolophid is wide. Roots are not preserved.

D⁴ (2.54 × 2.55): Sub-triangular outline. Parastyle and anteroloph are connected. A deep and open valley separates parastyle and protoloph. The large and high protocone is connected with the paracone by a slightly curved protoloph. There are a large mesostyle and a mesostylar crest. The metacone is larger than the metaconule; they are connected by a low and curved crest and their bases connect with the posteroloph. Metaconule and protocone are connected by a low and curved crest. Hypocone and posteroloph are low. Roots are not preserved.

M^1–2 (2.91 × 3.14; 2.86 × –) Only one specimen is complete; it has a sub-rectangular outline. The parastyle is connected to the paracone by a low and narrow crest. There is a mesostylar crest with an inflation (mesostyle) joined to the paracone. There is a sinuous protoloph. A low crest connects the large metacone and the metaconule, each of which is connected independently to the posteroloph; the metaconule connects with the protocone too. The hypocone is present. There are two main depressions: one delimited by anteroloph–protoloph, and the other one delimited by metacone–metaconule–protoloph. There are three smaller depressions, delimited by the crests that connect metacone–metaconule–posteroloph. The bottom of these depressions and the lingual face of the molar are rugose. Roots are not preserved.

Discussion: These specimens are characteristic for the subfamily Pteromyinae: large teeth with many sinuous accessory crests that give a labyrinthic appearance. The convergence of the lophs and the very well developed metaconule in the upper teeth are typical features of Pliopetaurista [15].

The size of our material is larger than that of other Neogene species of Pliopetaurista from Europe: P. bressana Mein, 1970 from its type locality (Soblay) [15] and from the Austrian locality of Eichkogel [10], P. dehneli Sulimski, 1964 from Wèze-1 [21] (type locality), Rebielice Królewskie [3], Hautimagne [15], Maramena [6] and Wèze-2 (personal comparison with specimens in Coll. Mein, Lyon-1 University), P. cf. dehneli from Podlesice [3] and P. meini Black and Kowalski, 1974 from Zalesicki (type locality) [3] (Fig. 5). The record of P. rugosa Qui, 1991 from Mongolia is very poor. The upper teeth of P. rugosa have no mesostyle, are shorter, wider and their cones are larger than in our material.

In Serrat d’en Vacquer, the type locality of P. pliocaenica, no upper teeth are preserved, but we have compared with other populations. The size and the main features of our material agree with P. pliocaenica from La Gloria-2, Desvío and Arquillo-3 [1], Wölfersheim [15] and Ivanovce (personal comparison with specimens in Coll. Mein, Lyon-1 University). The M_1–2 from PUR-25 cannot be measured, but its estimated size is close to P. pliocaenica's one.

Although size and morphology are quite variable in the genus Pliopetaurista, especially in P. pliocaenica [1], some features are more developed in P. pliocaenica and less developed or absent in P. dehneli (the most similar species in size and morphology to P. pliocaenica), like the hypocone, the parastyle, the mesostyle or the mesostylar crest, and the hypoconulid [15]; moreover, the size of P. pliocaenica is larger than that of P. dehneli [6] and [15]. Our material shares these characteristics, so we determine it as P. pliocaenica.

Species Pliopetaurista cf. pliocaenica (Depéret, 1897) (Fig. 4F–G)

Locality: Tollo de Chiclana-1B (TCH-1B)

Material: 1 D₄, 1 P⁴ and 3 M^1–2

Description: The specimens are very damaged.

D₄ (2.33 × –): Digested specimen. High and fused metaconid and protoconid. A curved posterolophid connects entoconid and hypoconid. A reduced mesoconid is present. The degree of wear impedes the observation of entolophid and hypoconulid. Roots are not preserved.

P⁴ (3.46 × 3.29): Worn specimen. The lingual border is shorter than the labial one. A large parastyle that is similar in size to the paracone and the metacone protrudes over the outline of the molar. The parastyle has a small postero-labial cusp and a low crest that connects with the paracone. There is a mesostylar crest and a large mesostyle that protrudes over the outline of the molar. Metacone, paracone and posteroloph are connected by broad crests. Due to wear, protocone, hypocone, and metaconule cannot be observed. Three roots.

M^1–2 (3.09 × 3.48; 3.15 × 3.41; – × 3.54): Sub-rectangular outline. The parastyle is separated from the paracone and much inflated. Large and high paracone and metacone. There is a mesostylar crest with a distinguishable mesostyle that protrudes over the outline of the molar. The metacone is connected with the metaconule and the posteroloph. The large and triangular metaconule is connected with the posteroloph and the protocone by a crest. Hypocone and protocone are large and there is a small anteroconule. The endoloph has a rugose lingual face. The hypocone is present. As in the specimens from PUR-4, there are two main depressions and three smaller ones, but all of them are attenuated due to wear. There are three roots with rugose face.

Discussion: Our specimens are larger than those of other European species of Pliopetaurista (P. bressana, P. dehneli, and P. meini). The upper teeth of the Asiatic, P. rugosa, are shorter, wider and their cones are larger than in TCH-1B. However, the main difference of the specimens from TCH-1B is the very much developed mesostyle sensu stricto in P⁴, and its presence in M^1–2. The main features of our material (hypoconulid, large hypocone, large parastyle and mesostyle) are very similar to P. pliocaenica. These features are absent or less developed in P. dehneli. The main difference with P. pliocaenica is that our material is larger; the M^1–2 from TCH-1B are larger than the ones from Wölfersheim [15], Ivanovce (personal comparison with specimens in Coll. Mein, Lyon-1 University), La Gloria-2, Desvío, Arquillo-3 [1] and PUR-4 (Fig. 5 and Fig. 6). On the other hand, their size is similar to that of P. cf. pliocaenica from Zamkowa Dolna [3] and Escorihuela-B [1] (Fig. 5 and Fig. 6).

The M^1–2 from TCH-1B have a mesostyle sensu stricto that protrudes over the molar outline, clearly more developed than in P. pliocaenica from Wölfersheim [15], Arquillo-3 [1] and PUR-4. The single M^1–2 of P. cf. pliocaenica from Escorihuela-B [1] also has a mesostyle sensu stricto [1]; these authors noted that the M^1–2 of P. cf. pliocaenica from Escorihuela-B and Orrios-3 have an ‘isolated mesostyle’; they did not refer to a cusp that is completely isolated from the mesostylar crest, but they referred to an individualized or distinguishable cusp, which may have a connection with the mesostylar crest. We emphasize this point because the figured specimen of M^1–2 from Orrios-3 [1] has a well-developed mesostyle that may have a connection with the mesostylar crest. These specimens of P. cf. pliocaenica have a mesostyle sensu stricto, in contrast with the inflated mesostylar crest in the M^1–2 of P. pliocaenica. Pliopetaurista pliocaenica from PUR-4, Wölfersheim [15], La Gloria-2, Desvío and Arquillo-3 [1] is older than P. cf. pliocaenica from TCH-1B and Escorihuela-B [1] (see Table 1). Because of these differences, we determine this material as Pliopetaurista cf. pliocaenica. The scarce and poorly preserved material does not allow more precision.

Atlantoxerus and Heteroxerus are ground squirrels. The record of these genera in the Turolian (Upper Miocene) is very scarce and discontinuous. In the Lower and Middle Miocene, the record is more complete and some phylogenetic hypotheses have been established [2] and [9]. Heteroxerus is present in Europe since the Late Oligocene [7] and [9]. Atlantoxerus appears in the Iberian Peninsula in the Aragonian (MN4) [2], [7] and [9], but it is present in France in MN3 [2]. The oldest record is from the Agenian of China (MN1) [17]; apparently, it is an immigrant that reaches France before the Iberian Peninsula; however, the absence in Spain in MN2 may be due to a lack of data.

There is an ancestor–descendant relationship between Heteroxerus paulhiacensis–H. rubricati–H. grivensis [9], and H. mariatheresae from La Gloria-5 (Upper Turolian) is the youngest representative of this lineage [1] (Fig. 7). The specimens from the Granada basin extend the register of H. mariatheresae to the end of Middle Turolian.

Heteroxerus vireti and H. rubricati were considered synonyms [9], but according to Aguilar [2], H. vireti from the French locality of Estrepouy must be classified as Atlantoxerus vireti, the predecessor of A. martini (Fig. 7). Atlantoxerus martini may be the ancestor of A. blacki [2], which has no relationship with younger species [12]. Atlantoxerus idubedensis and A. blacki, as said above, are considered as synonyms [2] and [16]. Atlantoxerus tadlae (Middle Miocene), A. rhodius (Mio-Pliocene boundary) and A. huvelini (Lower Pleistocene) are ancestors, and descendants [12] (Fig. 7).

Atlantoxerus margaritae is only known in the Lower Ruscinian of the Teruel [1] and Granada basins; it occupies an indeterminate position in the phylogeny of Atlantoxerus [1] (Fig. 7), and it does not seem to be related with A. adroveri, because the mesostyle is not present in A. adroveri. The phylogenetic relationships of A. adroveri are not clear either; it might be related with the extant species, A. getulus [12] (Fig. 7), but until now no intermediate record is known.

The genus Pliopetaurista belongs to the subfamily Pteromyinae. They are giant flying squirrels, forest dwellers that plane between trees.

Pliopetaurista shows a great variability in size and morphology. There are two lineages that may have a common origin in Forsythia or Pliopetaurista bressana [3] and [6]. The first line includes the small forms, which are only known from Poland [6]: Pliopetaurista sp. from Podlesice and P. meini from Zaleziaki (see Fig. 5). The species of the second line P. bressana–P. dehneli–P. pliocaenica show a tendency to size increase [6] and [15], and have a large geographic distribution [6] (see Fig. 4 and Fig. 5). Pliopetaurista rugosa from the Upper Turolian of Mongolia represents a central Asiatic branch of this group [6]. In the locality of Podlesice, two species of Pliopetaurista coexist: Pliopetaurista sp. (small size), and P. dehneli (large size) [3] (see Table 1 and Fig. 5).

This scheme is complicated due to the record of P. pliocaenica from the Granada Basin, in the localities of PUR-24A, PUR-25 (Upper Turolian) and PUR-4 (Lower Ruscinian), which are contemporaneous with central and eastern European localities containing P. dehneli (see Table 1 and Fig. 5). This finding constitutes the oldest record of Pliopetaurista in the Iberian Peninsula, and the oldest record of P. pliocaenica in Europe; it extends the biogeographic range of the genus to the South of the Iberian Peninsula.

As we said above, the lineage of P. bressana–P. dehneli–P. pliocaenica shows an increase of size through time [6] and [15]. Within P. pliocaenica, size increases too, and we find the biggest teeth of Pliopetaurista in localities like Escorihuela-B (Teruel, Spain, MN15) [1], TCH-1B (Guadix, Spain, MN15), and Zamkowa Dolna, (Poland, MN17) [3]; these are usually determined as P. cf. pliocaenica.

The oldest record of a large Pliopetaurista is from the Granada Basin (Mio-Pliocene boundary). The mesostyle is more developed in P. pliocaenica than in P. dehneli [6]. The M^1–2 of P. pliocaenica from PUR-4, Wölfersheim [15], La Gloria-2, Desvío, Arquillo-3 [1], have apparently a less developed mesostyle than those of P. cf. pliocaenica from Escorihuela-B [1] and TCH-1B. The mesostyle in P. pliocaenica is reduced to an inflated mesostylar crest or a protuberance in this crest, whereas in P. cf. pliocaenica from Escorihuela-B and TCH-1B, it is a well-distinct cusp that protrudes over the outline of the teeth, and may have a connection with the mesostylar crest. There are no references to the mesostyle of P. cf. pliocaenica from Zamkowa Dolna [3].

We agree with de Bruijn [6] in that the evolution of P. pliocaenica towards P. cf. pliocaenica is marked by size increase. The apparent increase in the development of the mesostyle from the oldest populations of P. pliocaenica to the youngest ones of P. cf. pliocaenica should be corroborated by a more complete record. The new data show that the larger-sized specimens of Pliopetaurista are recorded first in Spain (P. pliocaenica in PUR-24A, PUR-25 and PUR-4 from the Mio-Pliocene boundary; P. cf. pliocaenica in Escorihuela-B and TCH-1B from the Upper Ruscinian), and afterwards in the rest of Europe (see Table 1), whereas in contemporaneous central and eastern European localities (see Table 1), P. dehneli (in localities from the Mio-Pliocene boundary) and P. pliocaenica (in localities of the Latest Ruscinian) are present. Therefore, P. pliocaenica and P. cf. pliocaenica in Central Europe might be immigrants from the Iberian Peninsula.

The record of Sciuridae at the Mio-Pliocene boundary is very scarce, and although our material is poor too, it constitutes the most complete record of fossil Sciuridae in the South of Spain, with five species that have never been found before in this area.

The new data extend the record of H. mariatheresae to the Middle Turolian and enlarge its geographic range to the South of Spain. The biostratigraphic record of A. margaritae is extended to the Lower Ruscinian of the South of Spain, where it coexists with another lineage, which is represented in the Lower Ruscinian locality PUR-4 (A. cf. margaritae).

The appearance of P. pliocaenica at the Mio-Pliocene boundary of the Granada Basin extends the biostratigraphic record of P. pliocaenica to the Upper Turolian and Lower Ruscinian; this constitutes the oldest occurrence of the species, and its first record in the South of Spain. P. cf. pliocaenica, which is larger than P. pliocaenica, is recorded in other Upper Ruscinian localities of Spain. We suppose an emigration of these forms to Central Europe in two waves: one between the Mio-Pliocene boundary and the Late Ruscinian and another one during the Late Ruscinian and/or the Early Villafranchian.