All the fossil remains reported in this paper come from the Vallès-Penedès Basin (Fig. 1) in NE Iberian Peninsula. This basin is a NNE-SSW-oriented, asymmetric half-graben limited by the Littoral and Pre-littoral Catalan Coastal Ranges, which was generated by the rifting of the NW Mediterranean region during the Neogene (Bartrina et al., 1992, Cabrera et al., 1991, Cabrera et al., 2004, de Gibert and Casanovas-Vilar, 2011 and Roca and Guimerà, 1992). Besides some Early and Middle Miocene shallow marine and transitional sequences, most of the basin infill consists of distal-marginal alluvial fan sediments, with a rich fossil record of late Middle and Late Miocene terrestrial vertebrates (Agustí et al., 1985 and Casanovas-Vilar et al., 2011a). On the basis of the previously-available data for this basin (Agustí et al., 1997, Casanovas-Vilar et al., 2011a and Casanovas-Vilar et al., 2011b), the following correlations with Mammal Neogene (MN) units are available for the localities that have yielded Ptychogaster remains: CC and EC (MN4; Early Aragonian, Early Miocene); HPI (MN7 + 8; Late Aragonian, Middle to Late Miocene); HP (MN7 + 8 to MN9; Late Aragonian to Early Vallesian, Middle to Late Miocene); CM (late MN7 + 8 to MN9; Late Aragonian to Early Vallesian, Middle to Late Miocene); HPS (MN9; Early Vallesian, Late Miocene); EDAR24 (MN9; early Vallesian, late Miocene); CR (MN9–MN10; Early to Late Vallesian, Late Miocene); LT, SMT, VC and TS (MN10; Late Vallesian, Late Miocene). It should be taken into account that HP, HPI, HPS, CM and VC are not paleontological localities in a strict sense, but rather loosely-defined areas that mix material from different stratigraphic horizons that cover a relative long time interval (Alba et al., 2006, Alba et al., 2010 and Alba et al., 2011). The exact stratigraphic position of the new material from ACM is not recorded, but its provenance from sector ACM/BDA, on the basis of biostratigraphic and magnetostratigraphic data for the whole ACM series (Alba et al., 2006, Alba et al., 2011, Casanovas-Vilar et al., 2011b and Moyà-Solà et al., 2009), indicates an early MN7 + 8 (Late Aragonian, Middle Miocene) age.

Order: TESTUDINES Linnaeus, 1758

Infraorder: CRYPTODIRA Cope, 1868

Superfamily: TESTUDINOIDEA Batsch, 1788

Family: GEOEMYDIDAE Theobald, 1868

Subfamily: PTYCHOGASTERINAE De Stefano, 1903

Genus Ptychogaster Pomel, 1847

Type species: Ptychogaster emydoides Pomel, 1847.

Included subgenera: Ptychogaster (Ptychogaster) Pomel, 1847; Ptychogaster (Temnoclemmys) Bergounioux, 1958.

Emended diagnosis: Small to medium-sized geoemydid characterized by the following features: long peripherals; loose coincidence between the pleurals and the peripherals at the posterior part of the shell; posteriorly narrow suprapygal 1; sulci betwen marginal 12 and vertebral 5 situated on pygal; reduced anal notch; somewhat elongated internal ventral underlap of cervical; dorsal epiplastral lip long and narrow; entoplastron pyriform to rhomboid, longer than broad, and with rounded posterior margins; humeropectoral sulcus overlapping with the entoplastron, from about its midlength to three-quarters of the latter; wide dorsal overlap of the plastral scutes; narrow anterior edge of the anterior plastral lobe; vertebral 1 not completely covering the nuchal; irregular neural series formed mainly by octagonal and hexagonal plates with short sides in front; and nuchal longer than wide, markedly concave internally, and with a well-developed pair of internal thickenings. Further characterized by the following apomorphies: reduced axillary buttresses (approximately one-third of pleural 1 width); absent inguinal buttresses; completely ligamentous union between hypoplastron and peripherals; presence of a hinge between hyoplastron-hypoplastron; octagonal neural 2; well-developed pair of internal thickenings on the ventral surface of nuchal.

Differential diagnosis: Ptychogaster differs from other Geoemydidae by displaying a well-developed pair of internal thickenings on the nuchal plate. It differs from all other ptychogasterines (sensu Hervet, 2006) in the six apomorphic features mentioned above in the diagnosis. It further differs from all other ptychogasterines except “Geoemyda” saxonica Hummel, 1935 and Clemmydopsis Boda, 1927 in the position of sulci between marginal 12 and vertebral 5 on pygal—although this feature cannot be evaluated in Geiselemys Khosatzky and Młynarski, 1966. It also differs from other ptychogasterines except Clemmydopsis in the longer peripherals (sometimes with marked anterior expansions), the entoplastron longer than broad and with more rounded margins, and the markedly narrower anterior border of the plastral lobe. Ptychogaster also differs from Hummelemys Hervet, 2004b and Merovemys Hervet, 2006 in the loose coincidence between the pleurals and peripherals at the posterior portion of the shell; from Merovemys and Clemmydopsis in the posteriorly narrow suprapygal 1; from Hummelemys in the somewhat narrower and much longer dorsal epiplastral lip, the somewhat longer internal ventral underlap of the cervical, and the wider and more extensive dorsal overlap of the plastral scutes; from Merovemys in the larger overlap between the humeropectoral sulcus and the entoplastron, and in the presence of anal notch (albeit reduced); and from Clemmydopsis, in the less reduced axillary buttresses, less elongate nuchal (not longer than wide), the lack of fused costal and vertebral scutes, and in the fact that the posterior vertebrals are not wider than the costals.

Subgenus Ptychogaster Pomel, 1847

Other included species: Ptychogaster (Ptychogaster) laurae (Förster and Becker, 1888); Ptychogaster (Ptychogaster) sansaniensis (De Stefano, 1903); Ptychogaster (Ptychogaster) grepiacensis (Bergounioux, 1935); Ptychogaster (Ptychogaster) ronheimensis Groessens-Van Dyck and Schleich, 1985.

Diagnosis: As for genus. See below for a differential diagnosis with subgenus Temnoclemmys.

Ptychogaster (Ptychogaster) sp.

(Fig. 2 and Fig. 3)

Material: IPS64018 (Fig. 2 and Fig. 3), partial right and left epiplastron (possibly also a little portion of entoplastron) from EC; IPS64019 (Fig. 2 and Fig. 3), left partial peripheral 7 from CC; IPS64022 (Fig. 2 and Fig. 3), right epiplastron from HPI; IPS64023 (Fig. 2 and Fig. 3), partial nuchal from HP.

Stratigraphic range in the Vallès-Penedès Basin: MN4 (Early Miocene) to MN7 + 8–MN9 (Middle to Late Miocene).

Description of the new material: Based on the dimensions and structure of the peripheral plate and the epiplastron (in the material from EC) and the partial nuchal and right epiplastron (in the material from els Hostalets area), the shell would have been relatively large (> 20 cm). The dermal grooves and bony sutures are visible.

The peripheral 7 of IPS64019 (Fig. 2 and Fig. 3), which is nearly complete, is approximately rectangular in shape. In internal view, it displays a rough elongated area for the cartilaginous union with the inguinal process. It hosts the sulcus between marginals 7–8 and the costo-marginal sulcus, which is situated close to the pleuro-peripheral suture.

Despite some fractures affecting the anterior edge of the right epiplastron just medial to the right edge of the epiplastral lip, the ptychogasterid spike seems to be present but small in IPS64018 (Fig. 2 and Fig. 3); it is developed in posterior direction, slightly medially, although close to the gulohumeral sulcus, where it is poorly developed, not protruding dorsally and weakly defined. The epiplastral dorsal lip is only partially preserved, but its shape was most likely subrectangular. Its dorsal surface is somewhat concave medially, slightly thickened and anteroposteriorly long, imperceptibly shortening in medial direction so that its posterior edge is only slightly concave. The remnants of the anterior edge indicate that it was clearly concave, but for preservational reasons it is not possible to assess the depth of the concavity. On the basis of the small preserved portion of the gular area, it is possible to state that this element was extremely short and triangular, with rounded lateral margins, and probably not covering the anterior portion of the entoplastron.

On the basis of the right epiplastron IPS64022 (Fig. 2 and Fig. 3), it seems that the anterior plastral lobe was anteriorly truncated and slightly concave. The covering of the scutes on the dorsal surface of the epiplastron is well developed, resulting in an anteroposteriorly long epiplastral lip. The ptychogasterid spike is partially preserved. The area of the epiplastral lip posterior to the spike, and located medially to the gulohumeral sulcus, is roughly triangular and distinctly elongated, but not significantly protruding dorsally. The epiplastral dorsal lip of the isolated epiplastron is slightly wider than long. Its posterior edge is approximately straight and its dorsal surface is moderately concave medially. The gulars are proportionally long (covering the anterior portion of the entoplastron), triangular-shaped, with slightly rounded lateral margins. The dorsal overlap of the humeral scute on the epiplastron is well developed.

IPS64023 (Fig. 2 and Fig. 3) is hexagonal, probably slightly wider than long and highly concave ventrally. The anterolateral edge is vaguely concave. In lateral view, the nuchal is faintly concave, and the posterior edge of the cervical corresponds to a change in slope. The cervical is very large, triangular, and posteriorly quite wider, with a concave posterior edge and convex lateral margins. The internal overlap of the cervical is long and well developed. The left marginal 1, which is partially present, was apparently rectangular and slightly longer than wide. The vertebral 1, which is partially preserved, does not completely cover the nuchal because it is fairly narrow anteriorly; the small preserved portion of the left lateral vertebral-costal sulcus is slightly concave, possibly indicating that the vertebral was lyre-shaped.

Subgenus Temnoclemmys Bergounioux, 1958

Type species: Ptychogaster (Temnoclemmys) batalleri (Bergounioux, 1931).

Other included species: Ptychogaster (Temnoclemmys) cayluxensis Lydekker, 1889; Ptychogaster (?Temnoclemmys) francofurtanus Reinach, 1900 (nomen dubium); Ptychogaster (Temnoclemmys) bardenensis Murelaga et al., 1999.

Emended diagnosis: The subgenus Temnoclemmys differs from Ptychogaster s.s. in the following features: smaller size (adult shell length around 10–16 cm vs. 20–30 cm); much shorter cervical both dorsally and internally (not preserved in P. cayluxensis); trapezoidal (instead of rounded) margin of the anterior lobe and more tapering posterior lobe of the plastron (the latter only preserved in the type species); thicker epiplastral lip (at least medially, close to the gular sulci and the ptychogasterid spikes); shorter peripherals 1–2 without marked anterior expansions (only preserved in the type species); and shorter nuchal (i.e., shorter than wide; only preserved in the type species). The two latter features cannot be ascertained for P. (T.) bardenensis.

Ptychogaster (Temnoclemmys) sp.

(Figs. 2I–L, 3I–L)

Material: Two shell fragments of the same individual from CC, including: IPS64020 (Fig. 2 and Fig. 3), partial left epiplastron (a little portion of entoplastron is possibly present); and IPS64021 (Fig. 2 and Fig. 3), left peripherals 2–3.

Stratigraphic range in the Vallès-Penedès Basin: MN4 (Early Miocene).

Description of the new material: Based on the dimensions of the plates, the shell was small and thin. Most of the dermal grooves are visible, although the sutures are fairly obliterated.

The shape of the peripherals 2–3 IPS64021 (Fig. 2 and Fig. 3) cannot be assessed with confidence. The sulci delimiting the marginals 2–4 are visible. Marginal 3, the only complete scute, is longer than wide. In internal view, the overlap of the marginals is moderately developed.

The epiplastron of IPS64020 (Fig. 2 and Fig. 3) is moderately thin and lacks its posterolateral portion. Due to the anteroposterior length of this element and the absence of clearly visible sutures, it is not possible to exclude that a portion of the entoplastron is present. The ptychogasterid spike is preserved and distinctly protruding in anterior direction. The anterior edge of the epiplastron is slightly convex in dorsal view. The overlap of the gular and humeral scutes on the dorsal surface of the epiplastron is relatively well developed. The epiplastral dorsal lip is 2.8 times wider than long. In dorsal view, it is very little concave medially and slightly thickened and longer medially to the gulohumeral sulcus. The posterior edge of the lip is concave, being the anteroposterior width decreasing in medial direction. The gular is triangular, slightly longer than wide. Its relationships with the entoplastron are unclear due to the absence of visible sutures.

Ptychogaster (Temnoclemmys) batalleri (Bergounioux, 1931)

(Fig. 4, Fig. 5, Fig. 6, Fig. 7 and Fig. 8)

1926 Emys egarensis Bataller: 159 (nomen nudum).

1926 Testudo sp.: Bataller: pl. 3 figs. 1–2.

1931 Clemmys batalleri Bergounioux: 72, figs. 1–5 (original description).

1937 Clemmys batalleri Bergounioux, 1931: Bergounioux: 793.

1938 Clemmys batalleri Bergounioux, 1931: Bergounioux: 283.

1956 Clemmys batallery [sic] Bergounioux, 1931: Bataller: 21.

1957a Temnoclemmys batalleri (Bergounioux, 1931): Bergounioux: 1237 (new combination with invalid genus designation).

1957a Temnoclemmys vallesensis Bergounioux: 1237 (original description).

1957a Temnoclemmys elongata Bergounioux: 1237 (original description).

1957b Temnoclemmys batalleri (Bergounioux, 1931): Bergounioux: 40.

1957b Temnoclemmys vallesensis Bergounioux: 40.

1957b Temnoclemmys elongata Bergounioux: 40.

1957c Temnoclemmys batalleri (Bergounioux, 1931): Bergounioux: 164.

1957c Temnoclemmys vallesensis Bergounioux: 164.

1957c Temnoclemmys elongata Bergounioux: 164.

1958 Temnoclemmys batalleri (Bergounioux, 1931): Bergounioux: 185, figs. 18–19, pls. 35–36 (new combination with valid genus designation as its type species).

1958 Temnoclemmys vallesensis Bergounioux: 191, figs. 22–23, pls. 38–39.

1958 Temnoclemmys elongata Bergounioux: 194, fig. 24, pl. 40.

1958 Testudo celonica Bergounioux: 171, figs. 13–14, pls. 30–31 (original description).

1974 Testudo (Testudo) celonica Bergounioux, 1958: Auffenberg: 198.

1976 “Mauremys” batalleri (Bergounioux, 1958): Jiménez: 123.

1977 Ptychogaster (Temnoclemmys) batalleri (Bergounioux, 1931): de Broin: 239, fig. 100, pl. 23 figs. 1–4 (new combination).

1977 Ptychogaster (Temnoclemmys) cf. batalleri (Bergounioux, 1931): de Broin: 239, pl. 23 figs. 9–14.

1991 Ptychogaster (Temnoclemmys) batalleri (Bergounioux, 1931): Jiménez Fuentes and Martín de Jesús: 100.

1991 Ptychogaster celonica (Bergounioux, 1958): Jiménez Fuentes and Martín de Jesús: 102 (new combination).

2000 Ptychogaster (Temnoclemmys) batalleri (Bergounioux, 1931): Lapparent de Broin: 243, fig. 9 (emended diagnosis).

Emended diagnosis: A species of Temnoclemmys displaying all the features included in the diagnosis of this subgenus above, and further characterized by a trapezoidal, very thick and medially slightly concave dorsal epiplastral lip. It therefore differs from P. (T.) bardenensis by the medially less concave and much thicker dorsal epiplastral lip throughout its width, and from P. (T.) cayluxensis by the more trapezoidal and non-protruding contour of the anterior plastral lobe (due to the straight instead of convex gular angles), and the shorter dorsal epiplastral lip without a thickened inverse triangle.

Syntypes: Bergounioux (1931) described this species on the basis of two fossil remains that are currently lost (see below): an anterior portion of plastron (entoplastron and both epiplastra; Bataller, 1926: pl. 3 figs. 1–2; Bergounioux, 1931: figs. 1–2; see also our Fig. 4 and Fig. 5); and a shell fragment including two peripheral plates (left peripherals 7–8; Bergounioux, 1931: figs. 3–4; see also our Fig. 4 and Fig. 5). It is uncertain whether these two plates belong to a single individual. Given that Bergounioux (1931) did not designate a holotype, these two specimens must be considered syntypes. In order to clarify the taxonomic identity of this taxon, we think it is necessary to designate the most complete specimen (Fig. 4 and Fig. 5) as the lectotype, so that the fragment with the marginal plates (Fig. 4 and Fig. 5) must be considered a paralectotype. The lectotype was referred to as MGSM39663 by Jiménez Fuentes and Martín de Jesús (1991), and included in the catalogue of the MGSM published by Calzada and Urquiola (1992: 107). However, as noted by Jiménez Fuentes in Adserà and Calzada (2009: 22), this was a mistake, since Jiménez Fuentes and Martín de Jesús (1991) were referring to a different specimen from Sant Quirze, whose correct catalogue number is MGSB31663 (B-27). According to Bergounioux (1931), the syntypes were housed at the Institut Catholique de Toulouse. In 1985, the curators of the MSGB were informed by Drs. Bilotte and Cugny from the Institut catholique de Toulouse that this material was lost (Adserà and Calzada, 2009). However, given that there is no evidence that the lectotype has been destroyed, and that it was adequately described and figured in the original description of the species, we refrain from designating a neotype.

Type locality: Terrassa (MN10, Late Miocene).

Previously-described material: Partial shell (“specimen A” of Bergounioux, 1958: fig. 18, pl. 35; see also our Fig. 4 and Fig. 5) from VC, which is currently lost; partial shell (“specimen B” of Bergounioux, 1958: fig. 19, pl. 36; see also our Fig. 4 and Fig. 5) from VC, which is currently housed at the MGB (MGB V32); IPS696, partial shell from HPS (Bergounioux, 1958: fig. 24, pl. 40; see also our Fig. 4 and Fig. 5), which is the holotype of Temnoclemmys elongata; IPS2071 (field number VP387), partial shell from VC (Bergounioux, 1958: figs. 22–23, pls. 38–39; see also our Fig. 4 and Fig. 5), which is the holotype of Temnoclemmys vallesensis; MGSB25319A, partial posterior shell from CR (Bergounioux, 1958: fig. 13–14, pl. 30–31; see also our Fig. 4 and Fig. 5), which is the holotype of Testudo celonica; MGSB25319B, partial shell from CR (Fig. 4 and Fig. 5; previously not figured by Bergounioux, 1958), which is a paratype of T. celonica.

New material: IPS766a–h (Fig. 6 and Fig. 7), partial shell from EDAR24; IPS64024 (Fig. 6 and Fig. 7), right partial anterior plastral lobe from HP; IPS35267 (Fig. 6 and Fig. 7), partial anterior plastral lobe from VC; IPS20439 (Fig. 6M–N), anterior plastral lobe from ACM/BDA; IPS35288 (Fig. 6 and Fig. 7), partial left hypoplastron from LT; MGSB31699 (Fig. 6 and Fig. 7), epiplastron from CM; MGSB31692 (Fig. 6 and Fig. 7), entoplastron from SMT; IPS36059 (Fig. 6 and Fig. 7), posterior partial shell from VC; MGSB31662A (Fig. 6 and Fig. 7), partial anterior shell from CR; MGSB31662B (Fig. 6 and Fig. 7), partial carapace from CR. The most complete new material corresponds to the three partial shells IPS766, IPS36059 and MGSB31662A.

Stratigraphic range in the Vallès-Penedès Basin: early MN7 + 8 (Middle Miocene) to MN10 (Late Miocene).

Description of the new material from EDAR24: IPS766a (Fig. 6 and Fig. 7) preserves most of the anterior half of the plastron and most of the carapace, but the posterior area of the latter is in poor condition and not complete; the right xiphiplastron is misplaced and preserved within the matrix filling the carapace, so that only its posterior tip is visible among the right pleural fragments. The two epiplastra IPS766b–c (Fig. 6 and Fig. 7), left peripherals 9–10 IPS766d–e (Fig. 6 and Fig. 7), right peripheral 11 IPS766f (Fig. 6 and Fig. 7), right distal pleural 3 IPS766g (Fig. 6 and Fig. 7) and left peripheral 7 IPS766h (Fig. 6 and Fig. 7) were found associated to IPS766a and belong to the same individual.

The shell is relatively small, low and elongated in lateral view. The decoration of the carapace and plastron surfaces is smooth. The carapace is oval and quite slender, approximately 16 cm long and 13.5 cm wide once reconstructed. IPS766 probably represents a male individual due to the concavity of the hyoplastron and entoplastron. All the scute sulci are visible and the sutures are not obliterated. The carapace does not have evident carinae. As shown by the entoplastron and the hyoplastron, the anterior lobe of the plastron is distinctly bent in dorsal direction.

The nuchal is hexagonal, wider than long and not much concave ventrally. This plate is slightly elongated (ca. 21% of shell length), with the anterior part significantly widened and little emarginated. On the internal surface, there is a transversal moderately developed thickening in correspondence of the edge of the marginal overlap. In lateral view, the dorsal profile of the nuchal is straight, and the anterior edge is slightly arched in anterior view. The neural series is irregular: neural 1 quadrangular, neural 2 octagonal, neurals 3–6 hexagonal with short sides behind, neural 7 is hexagonal, with short sides in front, and neural 8 hexagonal with short sides in front. The pleurals are rectangular, with the length of the lateral edge being equal to that of the medial edge. There is no coincidence (alignment) between pleurals and peripherals in the posterior part of the shell. The peripherals 1–2 are longer than wide, slightly trapezoidal, but without clear strong anterior expansions. Peripherals 3–7 are involved in the bridge and develop a lateral ridge. Peripheral 7 is rectangular, and in internal view it displays the rough, elongated area for the cartilaginous union with the inguinal process. Peripheral 11 is rectangular with an oblique and slightly concave anteromedial border (this is reflected in the shape of the suprapygal 2; see below). The anterior and posterior peripherals are fairly equal in size.

The cervical is relatively small (covering 27% to 30% of the nuchal length), trapezoidal in shape, longer than wide, with the posterior border convex, and with straight lateral edges. The overlap of the cervical on the ventral surface of the nuchal is well developed but relatively short anteroposteriorly; the ventral surface of the cervical is wider than the dorsal one. The vertebral series is present, quadrangular, and slightly narrower than the costal series. Vertebral 1 is lyre-shaped, not covering the lateral corners of the nuchal and slightly narrower (at its widest point) than the nuchal. Vertebrals 2–4 are quite square, although slightly wider anteriorly than posteriorly. Vertebral 5 is fairly wider posteriorly than anteriorly. The costal series are rectangular, wider than long. The margino-costal sulcus is regularly situated on the peripheral, distal to the pleuro-peripheral suture, except at the level of the bridge (marginals 5, 7 and 9), where the sulcus contacts and in some cases crosses the pleuro-peripheral suture. Marginal 1 is dorsally wider than long; its length exceeds more than half of the anterolateral side of the nuchal. Marginals 2–4 are shorter than marginal 1 and rather square; marginal 4, in particular, has a costo-marginal sulcus situated rather close to the costo-peripheral suture. The ventral overlap of the marginals is well developed.

The anterior plastral lobe is relatively widened, slightly rounded and not completely filling the anterior opening of the shells. The anterior edge is truncated, straight and narrowed if compared to the width of the anterior lobe. The dorsal overlap of the humeral scutes is significant.

The epiplastron is vaguely pentagonal and approximately as wide as long. In dorsal view, the anterior edge of the anterior lobe is approximately rectilinear and laterally bordered by projections anteriorly directed, the ptychogasterid spikes, which are located within the gular shield and triangular, well-developed and distinctly protruding from the epiplastral outline. In anterior view, the edge is located close to the ventral border of the element and it is slightly convex medially. The lateral border of the epiplastron is overall straight but slightly curved close to the epi-hyoplastral suture. In dorsal view, the epiplastral dorsal pad is approximately rectangular (with a concave posterior edge), thrice wider than long; it is characterized by being longitudinally thickened in correspondence of the spike (therefore slightly medial to the gulohumeral sulcus) and moderately concave medially where the thickness of the pad equals that of the plate. However, a depression is developed on the dorsal surface of each epiplastron posteromedially to the gular pad, but laterally to the medial symphysis. In dorsal view, the posterior border of the pad is slightly concave medially (the maximum anteroposterior length corresponds to the longitudinal thickening of the spike), but it is, however, far from the anterior edge of the epiplastron. In lateral view, the epiplastral pad distinctly slopes in anteroventral direction. The area of the ventral surface covered by the gular is distinctly depressed compared to the humeral area.

The entoplastron is only partly developed in the anterior lobe. It is roughly pyriform in shape: the epi-entoplastron sutures are straight and form an angle of about 90°. The entoplastron has a roundish posterolateral edge. The humeropectoral sulcus crosses slightly below mid-height the ventral surface.

The hyoplastra are slightly wider than longer. They contact posteriorly the hypoplastron through a hinge and laterally with peripherals 3–6. The epi-hyoplastral suture is oblique and its lateral end is located rather close to humeropectoral sulcus. The hyo-hypoplastral hinge is straight and approximately transversal; it is located approximately at the mid of peripheral 6.

The gulars are approximately triangular-shaped, fairly short medially and narrow, so that they do not overlap the entoplastron.

The humerals are vaguely pentagonal, slightly longer than wide. The medial length of the humeral is much higher than the medial length of the gular. The humeropectoral sulcus is fairly straight medially but laterally concave anteriorly.

The pectorals are approximately rectangular and wider than long. They contact laterally with peripherals 3–6. The pectoro-abdominal sulcus is medially straight but laterally a little concave in posterior direction. It is located close (from 5 to 7 mm far) to the hyo-hypoplastral hinge but does not contact it. The anterior portion of the abdominal is therefore preserved on the hyoplastron.

Description of the new material from the area of Viladecavalls: IPS35267 (Fig. 6 and Fig. 7) from VC preserves the epiplastron and part of the entoplastron (possibly also a little portion of the left hyoplastron). The sutures are not visible. Taking into consideration the difference in size, the general morphology of this specimen is congruent with the corresponding region of IPS766b and IPS766c, with the exception of the gulars, which are proportionally wider, more developed in posterior direction and with straighter lateral edges. In anterior view, the ventral edge of the epiplastron is approximately straight and not slightly convex medially as in IPS766. The area of the ventral surface covered by the gular is not depressed if compared to that covered by humeral.

IPS36059a–d (Fig. 6 and Fig. 7) from VC is a partial posterior shell preserving part of the posterior half of the carapace and the left half of the posterior lobe of the plastron (which is fragmented and partially disarticulated). Both the sutures and the scutes sulci are visible. The specimen is highly fractured, but, thanks to the matrix filling of the shell, all the fragments are not significantly displaced. It includes: IPS36059a, left peripherals 8–11, pygal, suprapygals 1–2, neurals 7–8, left pleurals 6–8, right peripherals 10–11 and right pleural 8 (Fig. 6 and Fig. 7); IPS36059b, partial right pleurals 5–6 (Fig. 6 and Fig. 7); IPS36059c, right peripherals 7–9 and distal portion of pleurals 4–7 (Fig. 6 and Fig. 7); and IPS36059d, left hyoplastron and xiphiplastron (Fig. 6 and Fig. 7).

The shell is small, slightly high and domed in lateral view. The decoration of the carapace and plastron surfaces is smooth. In dorsal view, the preserved portion of the carapace seems to indicate that it was probably oval and quite slender; at least in the posterior sector of the carapace there are no evident carinae. The neural series is incomplete and the general morphology of this specimen is congruent with the corresponding region of IPS766a with the exception of neural 7, which is octagonal. The shape of the suprapygal 1 is quadrangular and slightly narrowed posteriorly. Suprapygal 2 is triangular, wider than long with rounded latero-posterior edges. The pygal is rectangular, slightly wider than long. In IPS36059a the costo-marginal sulcus shortly intersects the pleural-peripheral suture at the level of the marginals 9 and 11. Vertebral 5 is trapezoid, markedly wider posteriorly than anteriorly (only partially preserved in IPS766a) and partially covers the pygal posteriorly. The sulcus between vertebrals 4-5 laterally does not cross the suture between pleurals 7–8. Costal 4 is trapezoidal, slightly wider distally. The left portion of the posterior lobe of the plastron (IPS36059d) is narrowed with gently rounded borders. The posterior tip of the xiphiplastron is only partially preserved, but the anal notch should have been reduced (less than 15% of xiphiplastron length), and possibly as wide as long. The dorsal overlap of the plastral scutes is moderately developed, mainly in the area of the femoral scute. The inguinal buttresses are partially preserved and were probably reduced because of the presence of the hinge. The left hypoplastron is rectangular, slightly wider than long. It contacts laterally with peripherals 6–7 trough a completely ligamentous union in the inguinal process. The left hypoplastron-xiphiplastron suture is medially straight and laterally slightly curved in posterior direction. The left xiphiplastron is trapezoidal, slightly longer than wide.

The portion of the left abdominal developed on the hypoplastron is entirely preserved and rectangular. The abdominofemoral sulcus is straight (IPS36059d) and antero-laterally does not reach the inguinal process, being located slightly below it. The left femoral is rectangular, wider than long, with the dorsal overlap well developed. The femoroanal sulcus is oblique and faintly sinuous and laterally displays a wide notch. The left anal is triangular, fairly wider than long, and with rounded lateral borders.

IPS35288 (Fig. 6 and Fig. 7) from La Tarumba only preserves a left posterior partial fragment of hypoplastron that is basically congruent with IPS36059d, from which it differs because the abdominofemoral sulcus is weakly concave anteriorly. Given the scarce morphological details preserved, only a tentative taxonomic attribution is possible for this specimen.

MGSB31692 (Fig. 6 and Fig. 7) from San Miquel de Taudell is a complete entoplastron much smaller than IPS766a but morphologically congruent with it. However, the humeropectoral sulcus is placed more posteriorly in the former than in the latter; the ventral surface is only very slightly curved and therefore the anterior lobe of the plastron to which it belonged was less inclined in dorsal direction.

Description of the new material from the area of Hostalets: IPS64024 (Fig. 6 and Fig. 7) from HP preserves the right epiplastron (possibly also a little portion of the left epiplastron) and a fragment of the right hyoplastron. The epi-hyoplastral suture is not visible and the ptychogasterid spike is not preserved. The general morphology of this specimen is congruent with the corresponding region of IPS766c. However, IPS64024 is slightly smaller and less thick than IPS766c; the dorsal projection of the gular pad is less thick; the gulars are a little longer, with straight lateral edges. The preserved portion of the anterior lobe indicates that this was distinctly bent in dorsal direction.

The anterior plastral lobe IPS20439 (Fig. 6 and Fig. 7) from ACM/BDA preserves the epiplastron, entoplastron and left and partial right hyoplastron. The sutures are not visible because the skeletal elements are heavily fused together. Taking into consideration the difference in size, the general morphology of this specimen is congruent with the corresponding region of IPS766a with some exception: the gulars are proportionally wider and more developed in posterior direction and with straighter lateral edges; in dorsal view, the anterior edge of the epiplastron is slightly rounded and not straight as in IPS766a; the hyo-hypoplastral hinge is slightly thinner than IPS766 and oblique (not completely straight as in the latter); the humeropectoral sulcus is slightly sinuous and less curved laterally in anterior direction than IPS766a; the pectoro-abdominal sulcus is slightly oblique and sinuous instead of being completely straight as in IPS766a; and finally, IPS20439 shows the anterior lobe distinctly bent in dorsal direction.

MGSB31699 (Fig. 6 and Fig. 7) preserves both epiplastra and a little portion of the left entoplastron of the same specimen. Even if the general morphology of this specimen is similar to the corresponding region of the much larger IPS766b–c, the following features can be outlined: the gulars are proportionally narrower and slightly less developed in posterior direction and with sinuous lateral edges; in dorsal view, the anterior edge of the epiplastron is slightly rounded and not straight as in IPS766b–c; in ventral view, the ventral edge of the epiplastral pad has a slightly convex profile.

Description of the new material from Sant Celoni: MGSB31662A (Fig. 6 and Fig. 7) from CR preserves the anterior half of the shell and most of the posterior lobe of the plastron, which is housed mostly inside the shell and poorly preserved (not visible in the figure). The two hypoplastra are misplaced and preserved within the matrix filling the carapace, and therefore only their posterior tip is visible in posterior view. The right portion of the dorsal shell is partly covered by a hard concretion. All the scute sulci are visible whereas the sutures are obliterated. Except for its different size and the shell shape, which is relatively high and slightly domed, the general morphology of MGSB31662A is congruent with that of IPS766. The shape of the carapace is less oval and slender than in IPS766, being approximately 10 cm wide and approximately 12 cm long once reconstructed. MGSB31662A probably also corresponds to a male individual, given the concavity of the entoplastron and the hyoplastron. The anterior lobe of the plastron is more distinctly bent in dorsal direction than in IPS766. The nuchal is significantly widened in anterior view, more emarginated and much concave ventrally than in IPS766. In lateral view, the dorsal profile of the nuchal is less straight, much arched at the end of the cervical. The sutures and sulcus of the neural and peripheral series are not visible, and therefore differences in the neural pattern cannot be determined. Peripherals 1–2 are less prominent in anterior direction than in IPS766, but morphologically congruent with it. The cervical is relatively smaller than in IPS766 and has a narrower anterior edge narrower; the overlap of the cervical on the ventral surface of the nuchal is significantly longer anteroposteriorly. The vertebral series are somewhat narrower and vertebral 1 is less lyre-shaped. Marginal 1 is square-shaped and rather rectangular as in IPS766. The anterior plastral lobe of MGSB31662A is slightly narrower than in IPS766, somewhat rounded and not completely filling the anterior opening of the shell. In anterior view, the anterior edge of the anterior plastral lobe is located close to the dorsal border of this element and it is medially quite convex. In dorsal view, the anterior edge of the anterior lobe is slightly round and laterally bordered by ptychogasterid spikes, which are poorly developed and not protruding from the epiplastral outline. The gulars are triangular-shaped, proportionally longer in posterior direction and with straight lateral edges. The area of the ventral surface covered by the gulars is not depressed if compared to IPS766. In dorsal view, the epiplastral dorsal pad is vaguely trapezoidal, slightly longer in posterior direction than in IPS766. In lateral view, the epiplastral pad distinctly slopes anteroventrally, forming an angle slightly over 45°. The pectoro-abdominal sulcus is medially straight and close to, but not overlapping, the hyo-hypoplastral hinge; laterally it is located close to the hyo-hypoplastral hinge, contacting it at least on the left side.

MGSB31662B (Fig. 6 and Fig. 7) only preserves the central part of the carapace and the anterior part of the plastron, which is completely eroded. In the dorsal region, it preserves completely the vertebrals 2–3 and partially the posterior part of the vertebral 1 and the anterior part of vertebral 4. Vertebral series is rectangular, wider than long and not quadrangular as in IPS766. Furthermore, vertebrals 2–4 are less wide anteriorly. On the basis of the small preserved portion of the C1-C3 area, these elements seem rectangular.

The following Ptychogasterinae are recorded in the Vallès-Penedès Basin: Ptychogaster (Ptychogaster) sp. (MN4, MN7 + 8 and MN9?), representing the first record of subgenus Ptychogaster from this basin; P. (Temnoclemmys) sp. (MN4), indicating that the first appearance datum of subgenus Temnoclemmys is at least as old as that of the nominotypical one; and P. (T.) batalleri (MN7 + 8 to MN10), including most of the described material, and unequivocally showing the presence of this taxon also in the late Middle Miocene. Although some of the newly reported remains attain a larger size than previously recorded, the morphology of the studied specimens supports that a single species of Temnoclemmys is present in the Vallès-Penedès Basin during the Late Aragonian and Vallesian. Previously-unpublished material further indicates that Testudo celonica is another junior synonym of P. (T.) batalleri, like the nominal taxa Temnoclemmys elongata and Temnoclemmys vallesensis. In contrast, Temnoclemmys cordiformis is considered a nomen vanum of uncertain taxonomic affinities, whereas Ptychogaster francofurtanus is tentatively assigned to subgenus Temnoclemmys and provisionally considered a nomen dubium.