Measurements: L, mesiodistal length; B, labiolingual breadth; L/B, length/breadth index.

Institutions and fossil collections: ICP, Institut Català de Paleontologia Miquel Crusafont, Universitat Autònoma de Barcelona; IPS, collections from the ICP.

Fossil sites: ACM, local stratigraphic series of Abocador de Can Mata; BCV, sector of Barranc de Can Vila; BDA, sector of Bassa de Decantació d’Aigües Pluvials; C4-A, sector of Cell 4, subsector A; HI, Hostalets Inferior (Aragonian levels of els Hostalets de Pierola).

Dental nomenclature follows Alberdi et al. (2004: Fig. 2). Standard dental measurements (L, B and L/B) were measured from original specimens and taken from the literature.

The restricted sample of Anchitherium fossil remains described in this paper (including six cheek-teeth and several mandibular fragments from four different individuals) are housed at the ICP. The comparative sample includes material from other Anchitherium species of the Iberian Peninsula housed at the ICP, as well as data from other Iberian and European localities taken from the literature.

The holotype IPS20433 includes several uninformative mandibular fragments and three molar crowns, corresponding to the m2 (with an advanced degree of wear) and the two (right and left) m3 (with a somewhat less marked degree of wear). Two additional isolated teeth (paratypes) are available from ACM: IPS41937, an m3 crown with roots (showing a slight degree of wear); and IPS41989, a P2 crown (only very slightly worn).

All these cheek-teeth are very low-crowned and display a dental size overlapping with A. aurelianense. The P2 displays a subquadrangular occlusal contour, with a very protruding anterostyle on the mesiolabial corner of the crown, a biconvex lingual profile, a straight distal profile, and a W-shaped fold on the labial crown wall. The metacone is more protruding than the paracone, and the protocone is somewhat less peripheral than the similarly developed hypocone. The protoloph, which is unworn, does not connect to the ectoloph, and a considerable additional degree of wear would be required in order to do so. The metaloph displays a similar development and only minimally connects to the ectoloph, because the latter is worn to some degree. The prefossete fold is very small, merely consisting in a small style of enamel. The labial cingulum is not well developed and discontinuous (restricted to a small mesial portion between paracone and anterostyle, without reaching the distalmost portion of the crown, as well as a somewhat longer portion from mesostyle to metastyle). The lingual cingulum is very poorly-developed, almost vestigial, being very restricted between the protocone and hypocone. There are no mesial or distal cingula.

The available lower molars display a subrectangular occlusal profile (being longer than wide), the labial side of the crown being biconvex due to the presence of two distinct lobes, of which the mesial one is wider than the distal one (particularly in the m2). There is no lingual cingulid, whereas the labial cingulid is narrow and poorly developed (particularly between the mesial and distal lobes). The latter extends mesially until disappearing at about crown midline (being more marked in the m2 than in the m3), whereas distally it extends until crown midline in the m2, or until the distal lingual valley in the m3. Otherwise, there are no stylids on the lingual valleys of the lower molars. The m3 further differs from the m2 by being longer and relatively narrower, and further displaying a relatively small and rounded third lobe. The latter bears a distinct hypoconulid, lower than the remaining cuspids, and connected to the entoconid by a short crest.

Finally, the isolated p2 from HI referred to the newly described species preserves the complete crown and partially the fused roots of the distal lobe, whereas those from the mesial lobe are not preserved. This tooth is low-crowned and displays a small dental size, with a bilobulate occlusal profile (although the two lobes are united due to the advanced degree of wear), the mesial lobe being narrower than the distal one. There are no lingual, mesial or distal cingulids, whereas the labial one is only very slightly developed (almost imperceptible). Small and poorly-developed stylids are present on the distal lingual valley.

A. nievei most closely resembles in occlusal morphology and crown height the several brachyodont Anchitherium species that have been distinguished only as subspecies of A. aurelianense by other authors, although with significant proportional differences between them (Fig. 3). Thus, A. hippoides differs from A. aurelianense by the relatively broader P2 (Fig. 3A); the longer (although with some overlap) and relatively broader m2 (Fig. 3B); and the longer (although with substantial overlap) and relatively slightly broader m3 (Fig. 3C). In turn, A. steinheimense displays a P2 clearly broader than A. aurelianense, and even slightly so as compared to A. hippoides (although with substantial overlap) (Fig. 3A); a m2 intermediate in length and relative breadth between A. aurelianense and A. hippoides (Fig. 3B); and a m3 similar in length to A. hippoides but resembling A. aurelianense in relative breadth (Fig. 3C). The P2 of the new ACM species is more comparable to A. hippoides and A. steinheimense, i.e. slightly broader than in A. aurelianense (Fig. 3A), even though the considerable overlap and the small sample size preclude testing the significance of these differences. A. nievei further displays an m2 in the upper size range of A. aurelianense and A. steinheimense, and in the lower one of A. hippoides, although being clearly relatively wider than in all these other taxa (Fig. 3B and Table 2). Finally, the new species displays an m3 similar in length to A. hippoides and A. steinheimense (further overlapping with A. aurelianense), although clearly relatively broader, particularly as compared to A. aurelianense and A. steinheimense (Fig. 3C). When the m2 size and proportions are compared with other Iberian species of Anchitherium (Table 2), it can be concluded that A. nievei clearly differs from them by displaying a shorter but relatively much broader m2 (with a LBI lower than in the remaining European Anchitherium taxa investigated).

A. nievei can be distinguished from other endemic Iberian Anchitherium spp. by the more marked brachyodonty, the smaller dental size but relatively broader lower molars, and other details of the occlusal morphology: from A. corcolense by the lower and discontinuous lingual cingulum of the P2, the lack of hypostyle-hypocone connection on the P2, the lack of lingual stylids on the distal valley of the m3, the rounder distal lobe of the m3, and the lower and shorter labial and lingual cingulids of the m3; from A. ezquerrae (= A. matritense) by the shorter prefossette fold, the lower lingual cingulum, the lack of mesial and distal cingula and the less developed labial cingulum in the P2; from A. procerum by the presence of a prefossette fold (albeit not very developed) in the P2, the presence of a distal cingulid in the m3, and the lack of stylids on the distal valley of the m2; from A. alberdiae by the presence of stylids on the distall valley of the m3; from A. castellanum by the presence of prefossette fold and the lack of lingual cingulum in the P2, and the lack of stylids on the distall valley of the m3; and from A. cursor by the larger prefossette fold and the lack of lingual cingulum in the P2, and the lack of stylids on the distall valley of the m3. Only A. parequinum (MN5 of Zaragoza) cannot be directly compared to the ACM taxon due to the lack of dental material in the former.

With regard to the remaining Anchitherium species, both A. aurelianense and A. hippoides are recorded in the Early and Middle Miocene (respectively) of the Iberian Peninsula. However, given the scarcity of the remains and the lack of published descriptions, the comparisons between A. nievei and these other species must rely on the much more abundant material from France. The dental size of the ACM species is more similar to these European taxa than to the endemic Iberian ones, although it can be clearly distinguished from the former on the basis of dental proportions. The details of the occlusal morphology further enable to distinguish the ACM remains from A. aurelianense and the roughly synchronous A. hippoides and A. steinheimense by a unique combination of features, as enumerated in the differential diagnosis. One of the most diagnostic features of the ACM taxon is that the ectoloph and metaloph of the P2 would become connected to the hypocone only by through very extensive dental wear, whereas other differences refer to the development of cingula and cingulids.

The new species A. nievei, recorded from the Late Aragonian (MN7+MN8), is roughly contemporaneous with A. hippoides from the Calatayud-Daroca Basin (Toril 3A and Nombrevilla 2; Álvarez Sierra et al., 2003, Azanza et al., 2004 and Hernández Fernández et al., 2003), being only postdated in the Iberian Peninsula by the record of Anchitherium sp. from Castell de Barberà (MN8 or MN9) and especially Can Ponsic I (MN9), also from the Vallès-Penedès Basin. The lack of any record of Anchitherium in this basin during the MN5 and MN6, despite being recorded from several MN3 and MN4 localities, strikingly contrasts with the presence of several species throughout the whole Aragonian in the inner Iberian basins as well as in the rest of Europe, and is probably attributable to the scarcity of rich localities from this particular time interval at the Vallès-Penedès Basin. The genus Anchitherium displays a temporal trend towards size increase (Abusch-Siewert, 1983), in parallel to paleonvironmental changes. Climatic deterioration, and the associated increase of seasonability, resulted in more open habitats with a higher proportion of trophic resources with a high fibre content, which might have been the selection pressure that drove the size increase in the genus Anchitherium through the Middle Miocene (Forsten, 1991 and Hernández Fernández et al., 2003). This would explain the prevalence of large-bodied and more hypsodont forms amongst the last-surviving anchitheriines, particularly in the Iberian Peninsula, where trends towards increased occlusal area (Forsten, 1991) and hypsodonty (Eronen et al., 2010) have been detected. According to Sánchez et al. (1998), two different Iberian evolutionary lineages would have experienced this trend, beginning from a morphotype more similar to A. aurelianense, as an adaptation to more open environments with more abrasive vegetation.

This scenario is partly contradicted by the marked brachyodonty of the ACM Anchitherium, resembling the condition displayed by the earliest representatives of this genus in the Iberian Peninsula, together with the dispersal into Iberia during the late MN6 of A. hippoides (Hernández Fernández et al., 2003). Anchitheriines are considered inhabitants of subtropical to warm-temperate forests (Kaiser, 2009), displaying a browsing or mixed feeding niche (Eronen et al., 2010, Forsten, 1991 and Kaiser, 2009), although the paleoecological implications of each Anchitherium species depend on its particular dental and postcranial features. On the basis of the small and low-crowned cheek-teeth of both A. nievei and A. hippoides, a closed forest environment can be inferred (Hernández Fernández et al., 2003). This agrees with the increase in humidity that had been previously detected in the Calatayud-Daroca Basin from the MN6 onwards (Van der Meulen and Daams, 1992), and which contrasts with the more arid conditions inferred from older Iberian localities (Hernández Fernández et al., 2003). In Toril 3A, a humidity peak is confirmed by dental microwear studies (DeMiguel et al., 2011), whereas in ACM this is confirmed by the presence of particular taxa linked to closed forest environments, such as primates (Alba et al., 2010, Moyà-Solà et al., 2004, Moyà-Solà et al., 2009a and Moyà-Solà et al., 2009b). This further agrees with paleoenvironmental inferences based on the whole ACM fauna (Alba et al., 2006b, Alba et al., 2009, Alba et al., 2011, Casanovas-Vilar et al., 2008a and Casanovas-Vilar et al., 2008b), which is more similar to that from contemporaneous French and central European localities than to those from the inner Iberian basins (Casanovas-Vilar et al., 2008a), and which together with regional paleobotanical data agree with the presence of subtoprical to warm-temperate forest environments, although with a large proportion of deciduous taxa (Alba et al., 2009 and Casanovas-Vilar et al., 2008a).

Throughout most of the Miocene, the Vallès-Penedès Basin apparently displayed more humid and forested conditions than most of the rest of the Iberian Peninsula (Casanovas-Vilar and Agustí, 2007, Casanovas-Vilar et al., 2008a and Casanovas-Vilar et al., 2010), which would explain the presence of an Anchitherium species more closely related to French and European taxa than to endemic Iberian species. Given the scarcity of Anchitherium remains in the Catalan basins throughout all the Miocene, it is not possible to determine whether A. nievei represents an endemic evolutionary lineage, evolved in NE Iberian Peninsula from the earliest A. aurelianense that dispersed there by the Early Miocene, or whether it represents an independent later dispersal event from France, as suggested by the presence of A. hippoides in MN6 and other MN7+8 Iberian localities. The evolutionary and paleoecological implications of the presence of an endemic Anchitherium in the Vallès-Penedès Basin, contrasting with the presence of a more widely-distributed species (A. hippoides) in other Iberian basins, should be subject to further research in the future when the available material of the latter taxon from Spanish localities is published in detail.

Finally, although the rarity of Anchitherium in the ACM series could be partly explained by taphonomic reasons – such as the lack of environments preferred by small and brachyodont anchitheriine species –, this explanation seems rather unlikely. Although very humid environments with small ponds and lakes would have been only occasionally developed in the area of els Hostalets – as suggested by the rarity of castorids (Casanovas-Vilar et al., 2008b) and tragulids (Alba et al., 2011) – Anchitherium is also rare (or even absent) at many other MN7+8 localities from the Vallès-Penedès Basin, such as Castell de Barberà, where beavers are particularly abundant (Casanovas-Vilar and Agustí, 2007). Alternatively, the rarity of Anchitherium in the Vallès-Penedès Basin, particularly during the Late Aragonian, might be related to the competence with the small-bodied, browsing rhino Alicornops simorrense, as previously suggested in order to explain the rarity of A. hippoides amongst Iberian faunas by Hernández Fernández et al. (2003). In the Vallès-Penedès Basin, Al. simorrense is more well represented than Anchitherium during the Late Aragonian and Vallesian (Cerdeño and Sánchez, 2000), albeit being more scarce than in other Iberian localities, further being recorded at the ACM series (Alba et al., 2006a and Alba et al., 2009), where it is more frequent (both in quantity of remains and stratigraphic occurrences) than Anchitherium (DMA, pers. obs.). Nevertheless, a quantitative assessment is precluded because the rhinocerotid remains from ACM are still unpublished and to a large extent remain to be prepared. Moreover, paleoecological and morphofunctional studies outside the scope of this paper would be required in order to test to what extent Alicornops, considered an inhabitant of open woodlands with lacustrine or marshy areas (Cerdeño and Sánchez, 2000), could have displayed similar ecological requirements than a brachyodont Anchitherium species such as A. nievei.