Bollettino della Società Paleontologica Italiana, 50 (3), 2011, 175-208. Modena, 30 dicembre 2011175

Systematics and Paleobiology of malpassii n. sp. (, Mammalia) from the latest of Baccinello V3 (Tuscany, Italy)

Ray L. Bernor, Thomas M. Kaiser, Sherry V. Nelson & Lorenzo Rook

R.L. Bernor, College of Medicine, Department of Anatomy, Laboratory of Evolutionary Biology, Howard University, Washington DC, USA; rbernor@ comcast.net T.M. Kaiser, Biozentrum Grindel und Zoologisches Museum, Universität Hamburg, Hamburg, Germany; [email protected] S.V. Nelson, Department of Anthropology, University of , Albuquerque, New Mexico, USA; [email protected] L. Rook, Dipartimento di Scienze della Terra, Università degli Studi di Firenze, Firenze, Italy; [email protected]; corresponding author

KEY WORDS - Hippotherium malpassii n.sp., Systematics, Paleoecology, Baccinello-Cinigiano Basin, V3 fauna, Late Miocene, Messinian, Tuscany, Italy.

ABSTRACT - A new species of hipparionine , Hippotherium malpassii, is described from the MN13 interval from localities of the Baccinello V3 area (Italy). This species exhibits some advanced features of the facial region and cheek tooth dentition while other skull features and the postcranial skeleton are plesiomorphic. Research related to paleodiet suggests that Hippotherium malpassii was a mixed feeder, not as committed to grass feeding as many Central European species of Hippotherium. However, it may be considered as being similar to others such as Hippotherium primigenium from Eppelsheim (Germany) and Hippotherium microdon from Baltavar (Hungary). Studies of carbon and oxygen isotopes suggest that Hippotherium malpassii from the Baccinello V3 area lived in a forested context and it is inferred herein that it likely ate grass of relatively low abrasive character. Regarding the paleogeography of the species, Hippotherium originated in Central Europe early on in the MN9 interval (ca. 11 Ma) and underwent a modest evolutionary radiation there during the Vallesian and early Turolian (MN9-11). It is documented as having then dispersed throughout the Balkans into Greece during MN11/12 where it is represented by Hippotherium brachypus from the locality of Pikermi (Greece). During the interval MN12, the range of Hippotherium brachypus was extended and populations of very large size with massive metapodials evolved throughout this period. Hippotherium is not documented from the Baccinello area in Tuscany until the base of the MN13 interval (Baccinello V3 faunas) and may have immigrated from the Pannonian Basin, Hungary. The genus Hippotherium would appear to have become extinct at the end of MN12, the latter considered as being a result of increased continental drying and reduction of forest and woodland habitat. Its occurrence at Baccinello V3 in the MN 13 interval represents the latest occurrence of the genus in Europe.

RIASSUNTO - [Sistematica e Paleobiologia di Hippotherium malpassii n. sp. (Equidae, Mammalia) del Miocene Superiore di Baccinello V3 (Toscana, Italia)] - Sono analizzati e descritti i resti fossili di provenienti dall’associazione faunistica V3 del bacino miocenico terminale di Baccinello-Cinigiano (provincia di Grosseto). Il campione studiato rivela la presenza di un hipparion di taglia medio grande, caratterizzato da una morfologia con tratti avanzati nella regione facciale e nella dentatura, mentre conserva tratti plesiomorfici in parte della morfologia craniale e nel postcraniale. Il campione studiato rappresenta una specie nuova per la scienza e che viene qui istituita con il nome di Hippotherium malpassii. Grazie alla analisi della mesousura delle superfici occlusali dei denti sono state ricostruite le evidenze della paleodieta. Le analisi eseguite suggeriscono che Hippotherium malpassii aveva una dieta mista (“mixed feeder”), non vincolata ad una dieta esclusivamente basata su sole erbacee come nel caso della maggior parte delle specie centro-europee di Hippotherium, simile piuttosto alle popolazioni di Hippotherium primigenium del sito tedesco di Eppelsheim, o quelle di Hippotherium microdon di Baltavar, in Ungheria. Lo studio è completato dalla analisi della caratterizzazione degli isotopi stabili (C, O) dello smalto dentario nell’esemplare tipo della nuova specie Hippotherium malpassii. Combinando le evidenze di Carbonio e Ossigeno, i valori che emergono per questa specie sono consistenti con l’interpretazione di un taxon legato a un ambiente di foresta mesofitica e quindi probabilmente le piante erbacee di cui si cibava avevano un potenziale abrasivo relativamente basso. Lo studio dei resti di hipparion di Baccinello V3 consente di comprendere meglio la storia evolutiva del genere Hippotherium. Da un punto di vista paleobiogeografico, il genere Hippotherium si differenzia all’inizio della zona MN9 (circa 11 Ma) nelle regioni dell’Europa centrale a seguito della immigrazione dal Nord America del genere . Questo taxon mostra quindi una radiazione evolutiva durante il Vallesiano ed il Turoliano (MN9 - MN11). La regressione delle aree centrali della Paratetide nel Miocene Superiore provoca una continuità geografica tra il Bacino Pannonico e le aree a sud dei Balcani, portando quindi alla espansione del genere Hippotherium in Grecia (dove è presente a Pikermi con la specie Hippotherium brachypus) ed in Turchia. Durante il Turoliano superiore (zona MN12), la specie Hippotherium brachypus allargò il suo areale di distribuzione dando luogo a popolazioni caratterizzate da individui di grande taglia e metapodiali robusti. Hippotherium di Baccinello V3 rappresenta l’arrivo di questo genere nelle regioni dell’Italia centrale alla base della zona MN13, dove è arrivato espandendo il suo areale probabilmente a partire dal Bacino Pannonico. Il genere Hippotherium sembra estinguersi alla fine della zona MN12, possibilmente a causa della crescente riduzione delle aree forestate e boschive, legata alla tendenza generalizzata verso una aridificazione delle aree continentali durante il Miocene superiore. La nuova specie H. malpassii ci testimonia il più recente rappresentante ad oggi noto della linea evolutiva del genere Hippotherium.

INTRODUCTION (De Terra, 1956; Lorenz, 1968; Hürzeler & Engesser, 1976). Since the late 1990s, research undertaken by The geology and paleontology of the Baccinello the Vertebrate Paleontology Research Group of the area are well known thanks to the early research led University of Florence (after the initial contributions of by Johannes Hürzeler (1908-1995) from the Basel the late Claudio De Giuli, 1938-1998) has increased our Naturhistorisches Museum which recovered abundant knowledge of paleontology, geology and sedimentology faunal remains from well documented stratigraphic levels of the area (Benvenuti et al., 1994, 1999, 2001; Rook et

ISSN 0375-7633 doi:10.4435/BSPI.2011.16 176 Bollettino della Società Paleontologica Italiana, 50 (3), 2011 al., 2000, 2006, 2011; Ligios et al., 2008), providing a when there were multiple lineages of hipparion known better understanding of the sedimentary, environmental from Central Europe and the peri-Mediterranean realm, and faunal evolution of the Baccinello-Cinigiano Basin the most species diverse lineages being Hippotherium, (thereafter referred to as BCB). and Hipparion s.s. It would not at all The Baccinello hipparion samples in the Basel be unexpected to discover multiple species of hipparion Naturhistorisches Museum (NHMB) and Museo di Sto- from the Baccinello V3 localities. This study challenges ria Naturale (Sezione Geologia e Paleontologia) of the us to assemble hypodigms in a clear and logical manner. University of Florence (IGF), include a complex series Herein we will first seek corresponding morphologies to of fossil localities distributed over 100 square kilome- the type specimen, a maxillary tooth row, amongst an ters (Figs 1-2). Specimens usually outcrop as either extensive set of localities. We will then infer local refer- individual, or as handful of specimens from small, dis- rals to mandibular dentitions and postcrania from those crete localities. Moreover, the Late Miocene is a time corresponding localities. Since the Baccinello V3 equid fauna is so distinct, we will further undertake a paleodietary study using the me- sowear method of Fortelius & Solounias (2000). We also include a study of habitat reconstruction as expressed by carbon and oxygen isotope studies of the type specimen of Hippotherium malpassii n. sp. from Baccinello V3. We follow Bernor et al. (1997) in interpreting the func- tional anatomy of H. malpassii. This study establishes that Hippotherium malpassii has a biogeographic history connected to Central Europe. Other, smaller, rarer and more poorly represented Baccinello V3 equids may have an apparent evolution- ary relationship to eastern and southern Mediterranean forms of the genus Cremohipparion. Their description is beyond the scope of this paper.

GEOLOGICAL CONTEXT

The BCB is a Late basin located 25 km east of Grosseto in southern Tuscany (Figs 1-2), being one of the largest Tuscan “central basins”, (sensu Martini & Sagri, 1993). The BCB records a continental sedimentation throughout the late Tortonian-Messinian. The basin (Fig. 2) is filled with a minimum of 250 m of Late Miocene continental conglomerates, sands, silty clays bearing lignite seams, and freshwater carbonates (Lorenz, 1968; Benvenuti et al., 1999a, b, 2001). Before the middle of the last century, only scattered areas of the BCB were the object of concerted research, and these investigations were mainly related to lignite exploitation (e.g. Stoppani, 1880). The first stratigraphic investigations in the Basin began in the late 1950s (De Terra, 1956; Lorenz, 1968). Lorenz (1968) described the complex lithostratigraphy, which included the occurrence of numerous vertebrate-bearing fossiliferous levels. In recent times this Late Miocene succession has been referred to two unconformity-bounded stratigraphic units Fig. 1 - Locator map of the study area. A) Podere Firenze (NHMB: (synthems 1-2, Benvenuti et al., 2001; Fig. 2). Synthem 1 Arcidosso Locs. 11 and 3), type locality for Hippotherium malpassii; (late Tortonian-early Messinian), unconformably resting B) I Piscioli (NHMB: Arcidosso Locs 1 and 2); C) Ribaldella (also on the pre-Tortonian substratum made of limestone, referred to as “Ribardella”; cfr. Abbazzi, 2001; Benvenuti et al., 2001; Angelone & Rook, 2011), large badlands complex (NHMB: claystone and sandstone is further subdivided into six Cinigiano Loc 16; see Fig. 3); D) Tesorino very large complex of main units, in some cases with sub-units, deposited outcrops with several fossiliferus spots (NHMB: Cinigiano Locs within different paleoenvironments, ranging from slope- 4, 5, 6, 7); E) Caprarecce, very large complex of outcrops with palustrine settings to peat bogs to shallow lakes, alluvial several fossiliferus spots (NHMB: Cinigiano Locs 12, 13, 14, 15); plains and deltaic-lacustrine settings (Benvenuti et al., F) Poggio Vannini (NHMB: Cinigiano Locs 8, 9); G) Fosso del 2001). Fine grained deposits in the different units bear Taffone (NHMB: Roccalbegna Loc. 89); H) Le Pigne (NMHB: Roccalbegna Loc. 92); I) Poggio Miliotto; J) La Locca (see Fig. relatively abundant fossil remains that are 3); K) La Crocina (NHMB: Roccalbegna Loc. 4); L) Baccinello-1, grouped into distinct assemblages (vertebrate assemblages a complex of badlands outcropping south and south-east of Podere V0, V1, V2, and V3; Lorenz, 1968; Engesser, 1983, Santa Croce (NHMB: Scansano Locs 1-3, 14). 1989; Rook & Rustioni, 1991; Rook et al., 1991, 1996, R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 177

Fig. 2 - Schematic geologic outline of the Baccinello-Cinigiano Basin, and stratigraphic scheme of the BCB stratigraphic succession. Legend: 1) Quaternary talus and alluvial terraced deposits; 2) Early marine silty clays; 3) Unit 2, late Messinian alluvial gravels, sands and mudstones; 4) Unit 1f, Fluvio-lacustrine marls and sands; 5) Unit 1e, Floodplain silty clays, sands and gravels; 6) Unit 1d, Floodplain silty clays and gravels; 7) Unit 1c, Fluvio- lacustrine silty clays and sands; 8) Unit 1b, Palustrine lignites, organic clays and marls; 9) Unit 1a, Mass-flow gravels; 10) Santa Fiora Unit (Late -Early Eocene); 11) Macigno (Oligocene- Early Miocene); 12) normal fault (modified after Benvenuti et al., 2001 and Rook et al., 2011).

1999, 2000, 2008; Rook & Torre, 1995; Rook, 1999, of the depositional history of the sedimentary succession 2009; Abbazzi, 2001; Bernor et al., 2001; Eronen & based on a well resolved magnetostratigraphic section. Rook, 2004; Abbazzi et al.., 2008;Rook & Martinez The Baccinello V3 fauna is magnetostratigraphically Navarro, 2004; Chesi et al., 2009; Angelone & Rook, correlated to Chron C3An, early Messinian, ca. 6.733 to 2011; Casanovas-Vilar et al., 2011a, b; Gallai & Rook, 6.436, and MN13 (Fig. 3). 2011). Mollusc-rich levels characterize specific lacustrine mudstone and limestone, known in literature as F1 and Hippotherium localities F2 (Gillet et al., 1965; Esu & Girotti, 1989; Ligios et al., The sedimentary unit yielding the V3 faunal assemblage 2008). Synthem 2 (late Messinian) rests unconformably (Unit 1e, as defined by Benvenuti at al., 2001) is the one on the deposits of Synthem 1 and is referred to the late within the BCB sedimentary succession that widely Messinian, since Early Pliocene marine deposits ascribed outcrops across the basin’s geographical extension (Figs to the Sphaeroidinellopsis Zone statigraphically overlies 1-2). The giaciture setting of the geological units together it (Bossio et al., 1991; Benvenuti et al., 2001). with the geomorphologic characteristics of the area and Rook et al. (2011), provided a synthesis of the BCB the erosion patterns versus human use of the region has geological setting and support the chronological range the combined effect to concentrate the V3-fossil bearing 178 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

Fig. 3 - Stratigraphic columns magnetostratigraphically calibrated for reference sections in the southern (La Locca) and north-western (Ribaldella) sectors of the BCB respectively (cfr. Fig. 1). The type locality for Hippotherium malpassii (Podere Firenze) is litostratigraphically correlated with Ribaldella section (modified after Rook et al., 2011). R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 179 outcrops on the north-western and southern portion of Country Locality code the Basin, on the right banks of the Trasubbie river (to Inzersdorf i the south) and the Melacce river (to north-west). Most Austria of the recent single-specimen collections are recorded Gaiselberg g Höwenegg h in museum catalogues as single spot localities, while Germany specimens gathered from the earliest field work provide Eppelsheim e Pikermi p only a general indication of local jurisdictions provenance. Greece In terms of local jurisdictions, the BCB areas falls just Samos x on the junction of 5 different municipalities which Rudabánya r cause us to subdivide the BCB into five different sectors Hungary Baltavar t (Fig. 1): Cinigiano, the north and western portion up Polgardi o to the Melacce river; Arcidosso, the eastern central Italy Baccinello V3 B portion, east of Melacce river and north of Melacciole Turkey Esme Acakoy y stream; Roccalbegna, the southeastern portion, south of Sinap s Melacciole stream, including the upstream portion of the Akkasdagi k Trasubbie valley; Scansano, the southwestern portion, Tab. 1 - Late Miocene Old World comparative sample. the downstream portion of the Trasubbie valley (includes Letters refer to locality designations used in bivariate plots of Figs the hamlet of Baccinello); Campagnatico, the western 4-7, 9 and 11. central portion.

1833 (Bernor et al., 1997). This sample has been found METHODS to be particularly useful for statistical comparisons of postcranial elements. Indeed, the possibilities for broad We will compare the Baccinello V3 hipparions to a comparative study of hipparion metapodials are greatly relevant group of Late Miocene Old World hipparions expanded by the extraordinary preservation of, and body that we refer to the Hippotherium lineage (Tab. 1). These of scholarship on, hipparions from Höwenegg (Bernor et include Central European Vallesian Hippotherium from al., 1997; Scott, 2004). the German localities of Höwenegg (indicated by a “h” Two statistical methods are used. In the first, we on the ensuing bivariate plots) and Eppelsheim (e on the generate a series of bivariate plots which include 95% plots). Also included are the Austrian Vallesian localities confidence ellipses using the Eppelsheim sample for of Inzersdorf (i) and Gaiselberg (g), the Hungarian cheek teeth and the Höwenegg sample for postcranial Vallesian locality of Rudabánya (r) and Turolian age elements as prescribed above. These plots allow a localities of Baltavar (t) and Polgardi (o). We also include visual representation of how Baccinello V3 compares statistical comparisons to “Cormohipparion” sinapensis to our comparative sample for some key morphological from the Turkish Vallesian localities of Esme Acakoy (y) features. We also use log10 ratio diagrams strictly for our and Sinap (s). Baccinello V3 is a late Turolian locality MCIII, MTIII and 1PHIII samples. Traditionally, log10 indicated in our plots as a “B”. These taxa and the age of ratio diagrams have been used as a descriptive tool in their localities have been discussed in several previous discussions of hipparion morphology (e.g., Eisenmann, studies (Bernor et al., 1988, 1993, 1996, 1997, 2003a, b; 1995). We apply this approach here to MPIII’s and

Kaiser et al., 2000, 2003; Bernor & Scott, 2003; Scott et 1PHIII’s. Log10 ratio diagrams graphically illustrate al., 2005a, b; Woodburne, 2007). deviations from a comparative standard (in this study We analyse here the following skeletal elements the Höwenegg population sample) and are a descriptive following previous studies cited above: maxillary P2, heuristic. Profiles shown in ratio diagrams represent astragali (hereafter AST), calcanea (CALC), metacarpal absolute deviations from a morphometric standard. III’s (MCIII’s), metatarsal III’s (MTIII’s; MCIII’s and For mesowear analysis we apply here the mesowear MTIII’s collectively will be referred to as MPIII’s) methodology originally formulated by Fortelius & and 1st phalanges III’s (1PHIII’s) using standard equid Solounias (2000). Fortelius & Solounias (2000) and measurements published by Eisenmann et al. (1988) and Kaiser et al. (2000) found that the mesowear signal Bernor et al. (1997). Measurements of the Baccinello V3 allows a reasonable scoring of paleodietary preference hipparions are provided in Appendix 1. when the sample size is 10 dental specimens or greater. For cheek teeth we use the Eppelsheim sample as A statistically stable paleodietary signal can be expected our standard for calculating 95% confidence ellipses when the sample size is 20-30 individuals. The sample (following Bernor & Franzen, 1997; Bernor et al., 2003b). size available to this study is much too small to obtain a The Eppelsheim sample is superior because it is a large statistically sound mesowear classification if the standard sample of a primitive Old World hipparion composed methodology is followed. In order to gain a reasonably almost entirely of isolated cheek teeth from a single quarry. accurate scoring of our sample, we abandon conventional This allows measurements at all stages of ontogeny. practice here as follows: 1) tooth positions other than For postcranial elements we use the Höwenegg sample upper M2s are included, 2) lower teeth are included. as our analytical standard for the postcranial skeleton. Scorings of mandibular teeth are calibrated following This population is composed of 16 articulated skeletons Kaiser & Fortelius (2003) for the Eppelsheim sample of (many of which are complete), along with isolated bones Hippotherium primigenium. and is “biologically uniform”, including only a single Hierarchical cluster analysis with complete linkage primitive species, Hippotherium primigenium Meyer, (furthest neighbours) is applied here following the 180 Bollettino della Società Paleontologica Italiana, 50 (3), 2011 standard hierarchical amalgamation method of Hartigan (1982) was also consulted for morphological identification (1975). The algorithm of Gruvaeus & Weiner (1972) and comparison. was then used to order the cluster tree using the three mesowear variables (% high, % sharp and % blunt). Conventions - The nomen Hipparion has been used As a comparative dataset for dietary classification, we in a variety of ways by different authors. We follow use 27 extant mammalian ungulate species reported as characterizations and definitions for hipparionine “typical” dietary categories by Fortelius & Solounias recently provided in Bernor et al. (1996, 1997). Ana- (2000). As comparative datasets for fossil hipparionine tomical descriptions have been adapted from Nickel et al. horses, we include mesowear signatures of four Old World (1986). Getty (1982) was also consulted for morphological Hippotherium taxa, Hippotherium primigenium from identification and comparison. Hipparion monographs by the MN9 of Eppelsheim (Dinotheriensande, Germany; Kaiser, 2003), Hippotherium primigenium from MN9 of Höwenegg (Germany; Kaiser, 2003), Hippotherium aff. primigenium and Hippotherium kammerschmittae from MN11 of Dorn-Dürkheim (Germany; Kaiser et al., 2003), Hippotherium sumegense (Hungary; Bernor et al., 1999) and Hippotherium intrans (MN 9 of Rudabánya, Hungary; Bernor et al., 2003a). We further apply Principal Components Analysis (PCA) on mesowear parameters of each of the above mentioned datasets. As in the cluster analysis, we include 27 “typical” extant ungulate species after Fortelius & Solounias (2000) as comparative species with known diets. Systat 9.0 software was used to compute cluster statistics, using default settings. For stable isotope analysis a tooth is sampled three times from crown to root. Samples of pure enamel were removed using a Dremel high-speed rotary tool with a diamond-impregnated bit. Samples were run using an automated carbonate microsample device. Each isotopic sample consisted of 0.5-0.7 mg of enamel, and samples were washed with 3% hydrogen peroxide for 15 minutes and rinsed, followed by 0.1M acetic acid for 15 minutes and rinsed. Samples were then reacted with 100% phosphoric acid at 90oC on the automated carbonate device interfaced with a Finnigan MAT 252 stable isotope ratio mass spectrometer. Isotopic ratios are presented in the per mil (‰) notation 13 18 d C (or O) = (Rsample/RPDB-1) x 1000 13 12 18 16 where Rsample and RPDB are the ratios C/ C (or O/ O) in the sample and standard respectively, and the isotope reference standard is PDB. Oxygen values are reported assuming the acid-calcite fractionation factor for calcite.

Abbreviations and Conventions Abbreviations - IGF: Museo di Storia Naturale (Sezione Geologia e Paleontologia), University of Flor- ence; NHMB: Naturhistorisches Museum, Basel; AST: astragali; CALC: calcanea; MCIII: metacarpal III; MTIII: metatarsal III; MPIII: metapodial III; 1PHIII: 1st phalanx III.

Measurements - Measurement numbers (M1, M2, M3, etc.) refer to those published by Eisenmann et al. (1988; and rounded to 0.1 mm) for the skulls and postcrania (also in Bernor et al., 1997), whereas tooth measurement numbers refer to those published by Bernor et al. (1997), Bernor & Franzen (1997) and Bernor & Harris (2003). Fig. 4 - Bivariate plots of upper teeth. a) For a large sample of P2, occlusal width (M3) is plotted versus occlusal length (M1) with a Anatomical Descriptions - The osteological 95% confidence ellipse based on the Eppelsheim (Germany, MN9) nomenclature, the enumeration and/or lettering of the sample. b) P4 protocone width (M11) is plotted versus protocone figures have been adapted from Nickel et al. (1986). Getty length (M10) with no address to an Eppelsheim 95% ellipse. R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 181

Gromova (1952) and Gabunia (1959) are cited after the all fall in the center, lower right and (one) just outside French and English translations. The taxon Hipparion has the right margin of the ellipse. Baltavar is represented been applied in a variety of ways by different authors. We by several specimens, most in the lower portion or just utilize the following definitions in this work: outside the lower margin of the ellipse. Gaiselberg has • Hipparionine - a taxonomic tribe of Equidae with two specimens in the lower portion of the ellipse and an isolated protocone on maxillary premolar and molar a single specimen well below the ellipse. Overall, the teeth and, as far as known, tridactyl feet, including species vast majority of specimens fall within the ellipse and the of the following genera: Cormohipparion, , outliers outside the ellipse are not substantially smaller , , Hippotherium, Cremohip- than Eppelsheim Hippotherium primigenium. Fig. 4b parion, Hipparion, “”, (= plots P4 protocone width (M11) versus protocone length senior synonym of “Stylohipparion”) , (M10) with no address to an Eppelsheim 95% ellipse. A Plesiohipparion. Characterizations of these taxa can be general characteristic of the Baccinello V3 hipparions is found in MacFadden (1984), Bernor & Hussain (1985), that the protocone is short, giving it a rounded aspect. The Webb & Hulbert (1986), Hulbert (1988), Qiu et al. (1988), Vallesian localities of Gaiselberg (g) and Höwenegg (h) Bernor et al. (1988, 1989, 1990, 1996, 1997, 2003a, b, have longer protocones than most of the Baccinello V3 2004a, b, 2005, 2008, 2010), Woodburne (1989), Hulbert sample. Baltavar (t) compares closely with Baccinello V3 & MacFadden (1991), Bernor & Armour-Chelu (1999), falling in the middle of the Baccinello range. Baccinello Bernor & Harris (2003), Bernor & Scott (2003); Kaiser V3 has a total width range (M11) of 3.3 to 5.6 and total et al. (2003); Scott et al. (2005a, b); Armour-Chelu et al. length range (M10) of 4.2 to 7.2 which is less than that (2006); Bernor & Kaiser (2006); Bernor (2007); Gilbert of Höwenegg: 3.1 to 6.6 and 5.8 to 8.8, respectively, and & Bernor, (2008); Bernor & Haile-Selassie (2009); Ber- therefore not indicative of representing more than a single nor & White (2009); Wolf et al. (2010); Armour-Chelu species in the Baccinello V3 sample. Bernor et al. (2003a) & Bernor (2011). noted the small, rounded protocone present in a hipparion • Hipparion s.s. - the name is restricted to a specific from the Hungarian MN 13 locality of Hatvan. lineage of hipparionine horses with the facial fossa posi- Fig. 5 is a bivariate plot of astragalus length (M1) tioned dorsally high on the face (MacFadden, 1980, 1984; versus distal articular width (M5), compared to the Woodburne & Bernor, 1980; Woodburne et al., 1981; Höwenegg standard. The Baccinello V3 (B) astragali MacFadden & Woodburne, 1982; Bernor, 1985; Bernor mostly fall within the lower half of the ellipse with a single & Hussain, 1985; Bernor et al., 1987, 1989; Woodburne, specimen falling to the left of the ellipse. Eppelsheim (e), 1989). The posterior pocket becomes reduced and even- Rudabánya (r), Inzersdorf (i) and Gaiselberg (g) mostly tually lost, and confluent with the adjacent facial surface fall in the lower portion of the ellipse also. Polgardi has (includes Group 3 of Woodburne & Bernor, 1980). Ber- most of its specimens plotting in the lower part, or just nor’s definition departs from some investigators in not below the ellipse. Baltavar (t) is similar to Polgardi in this recognizing North American species of Hipparion s.s. regard, but has several specimens that are even smaller. Bernor (1985) and Bernor (in Bernor et al., 1989) have In the cases of Rudabánya, Baltavar and Polgardi, there argued that any morphologic similarity between North is a smaller hipparion known from these localities (Scott American “Hipparion” and Old World Hipparion s.s. is due to homoplasy. • “Hipparion” - several distinct and separate lineages of Old World hipparionine horses once considered to be referable to the genus Hipparion (Bernor et al., 1980; Woodburne & Bernor, 1980; MacFadden & Woodburne, 1982; Bernor, 1985; Bernor & Hussain, 1985; Bernor et al., 1988, 1989). We emphasize here the need to avoid con- fusion of well defined hipparionine lineages with poorly characterized taxa of “Hipparion” sensu lato. • Hippotherium - a discrete genus of Western Eurasian hipparionine horses known from Central Europe, Italy, Greece and possibly Ukraine. Species belonging to this genus include H. primigenium, H. intrans, H. microdon, H. kammerschmittae, H. malpassii n. sp. (this contribution), H. brachypus and perhaps H. giganteum.

STATISTICAL ANALYSES

Fig. 4 provides a bivariate plot of P2 occlusal width (M3) versus occlusal length (M1) with a 95% confidence ellipse based on the Eppelsheim (Germany, MN9) sample. Fig. 4a includes a large sample of P2’s. Most of our sample falls within the Eppelsheim ellipse, including: Rudabánya (r), Höwenegg (h), Gaiselberg (g), and a particularly large Fig. 5 - Bivariate plot of astragalus length (M1) versus distal articular sample from Baltavar (t). The Baccinello V3 P2’s (B) width (M5), compared to the Höwenegg standard. 182 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

Fig. 6 - Bivariate plot of calcanea for maximum length (M1) versus maximum width (M6).

et al., 2003), but in some of these the smaller size may be due to the juvenile status of the individual sampled: it is very difficult to ascertain juvenile status of astragali. The result of this analysis suggests that there is more than a single sized hipparion that exists in several of these locality samples. Fig. 6 plots the known calcanea in our sample for maximum length (M1) versus maximum width (M6). Specimens plotted within the ellipse include Eppelsheim (e), Rudabánya (r) and Inzersdorf (i). Baccinello V3 (B) plots just outside the lower border of the ellipse and also far before the lower border; this latter, smaller specimen may be of a different species than H. malpassii n. sp. since a rare, smaller horse is already know from the cheek teeth. Rudabánya (r) includes two specimens, one on the lower border of the ellipse and one just outside the lower border of the ellipse. Baltavar (t) has an extremely small specimen very far below the ellipse that may represent a rare (other than Hippotherium intrans and Hippotherium microdon), third species in the fauna. The Sinap specimen (s) is referable to C. sinapensis (Bernor et al., 2003b) Fig. 7 - Bivariate plots of Metacarpal III. a) Plot of MCIII maximum length (M1) versus distal articular width (M11) compared to the which has been shown to be smaller than Central European Höwenegg standard 95% confidence ellipse. b) Plot of MCIII distal Vallesian Hippotherium primigenium. midsagittal keel (M12) versus distal articular width (M11). Fig. 7a plots MCIII maximum length (M1) versus distal articular width (M11) compared to the Höwenegg standard 95% confidence ellipse. Most specimens fall within the Höwenegg ellipse, including 2 specimens from of the ellipse, or below the ellipse. Most of the Pikermi Baccinello V3 (B) in the middle of the ellipse, 4 specimens specimens of Hippotherium brachypus plot within the from Esme Acakoy (y), 4 specimens from Inzersdorf (i), Höwenegg ellipse, with one plotting to the right of the 2 from Sinap (s; C. sinapensis), and 6 specimens from ellipse. Samos has two specimens plotting in the upper Baltavar (t) in the lowermost portion of the ellipse. One part of the ellipse and two well above the ellipse. These specimen from Inzersdorf (i) plots to the right of the larger Samos specimens overlap with several specimens ellipse, while 5 specimens from Baltavar (t) plot below of Hippotherium “brachypus” from Akkasdagi, Turkey the ellipse. Baltavar (t) Hippotherium microdon is once (Vlachou & Koufos, 2009). We believe that these larger again shown to be a small member of the Hippotherium specimens of Hippotherium “brachypus” from Samos clade with several specimens falling in the lower portion and Akkasdagi are of an advanced member of the R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 183

Hippotherium brachypus clade. Fig. 7b plots MCIII distal Vallesian Hippotherium spp. from Inzersdorf, Austria midsagittal keel (M12) versus distal articular width (M11). (MN 9; Bernor et al., 1988), Cormohipparion sinapensis The results are essentially the same as for Fig. 7a and we from Esme Acakoy, Turkey (MN 9; Bernor et al., 2003b), include this plot here to emphasize the consistently larger the Dinotheriensandes Hippotherium primigenium, size of the advanced Hippotherium “brachypus” from and Rudabánya Hippotherium intrans compared to the Samos (x) and Akkasdagi (k). Höwenegg mean Log 10 standard. The Vallesian MCIII Fig. 8 includes Log 10 ratio diagrams of MCIII. Fig. sample is remarkably similar in its dimensions with 8a compares the two Baccinello V3 MCIII’s to relevant Rudabánya Hippotherium intrans being the longest MCIII

Fig. 8 - Log 10 ratio diagrams of MCIII. Fig. 8a compares the two Baccinello MCIII’s to relevant Vallesian Hippotherium spp. compared to the Höwenegg mean Log 10 standard. Fig. 8b replots the Baccinello MCIII’s compared to Turolian members of the Hippotherium clade. 184 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

suggesting that they are of the same species and related to other members of the Hippotherium clade. Together they mostly bracket the rest of the sample in their morphology. Fig. 8b replots the Baccinello V3 MCIII’s compared to Turolian members of the Hippotherium clade. Once again, this plot exhibits the same proportional relationships of maximum length (M1), mid-shaft width (M3) and mid- shaft depth (M4) as found in the Vallesian taxa. Here, Baccinello’s Log 10 ratio would appear to track most closely with the Polgardi mean measurements while Baltavar (Kaiser & Bernor, 2007) and Dorn Dürkheim (Bernor & Franzen, 1997; Kaiser et al., 2003) compare particularly closely to one another except for the former’s shorter maximum length measurement (M1). As per the bivariate plots, Samos Hippotherium “brachypus” (here SAM_Hbra2) plots along with Akkasdagi Hippotherium “brachypus” with much larger dimensions than the rest of the sample. We distinguish further here a smaller form of Samos Hippotherium “brachypus” which has a MCIII that has a very short maximum length (M1) and is similar to Pikermi Hippotherium brachypus in not showing the Esme Acakoy effect: midshaft width (M3) is not proportionally narrower than M4. This smaller member of the Hippotherium brachypus clade (Hbra1 here) is known from Samos Quarry 6, while the larger form (Hbra2 here) is from Samos Quarry 1, 2 and 4. Fig. 9a plots MTIII maximum length (M1) versus distal articular width (M11) compared to the Höwenegg standard 95% confidence ellipse. Again, most of our sample plots within the Höwenegg ellipse, including: 7 specimens from Esme Acakoy (y; C. sinapensis), 4 specimens from Inzersdorf (i), 1 from Gaiselberg (g) and 2 from Polgardi (o). There is one Baccinello V3 specimen that plots on the lower left border of the ellipse (B). Rudabánya (r) has a specimen that plots slightly above the ellipse while there are specimens from Esme Acakoy (y) and Sinap (s) that plot on the left side of the ellipse. While there are 3 Baltavar specimens (t) that plot within the ellipse, most of the Baltavar specimens, referable to Hippotherium microdon plot at the lower boundary or below the Höwenegg ellipse. Once again, all Pikermi specimens of Hippotherium brachypus plot within the upper half of the Höwenegg ellipse, except for one that plots just outside the right border of the ellipse. There is a single Samos specimen of Hippotherium “brachypus” (Hbra1) that plots in the center of the ellipse (x), while others plot Fig. 9 - Bivariate plots of Metatarsal III. a) Plot of MTIII maximum at the top of the ellipse. There are two specimens of the length (M1) versus distal articular width (M11) compared to the larger Samos Hippotherium “brachypus” at the top and Höwenegg standard 95% confidence ellipse. b) Plot of MTIII distal outside the upper border of the ellipse overlapping in their midsagittal length (M12) versus distal articular width (M11). size with Akkasdagi Hippotherium “brachypus”. Fig. 9b plots MTIII distal midsagittal length (M12) versus distal articular width (M11). Interestingly, this plot shows that Pikermi Hippotherium brachypus shows a trend toward increased midsagittal length (M12) that is at the top and and Inzersdorf Hippotherium primigenium being the above the Höwenegg ellipse overlapping with Samos shortest MCIII. The remaining Vallesian Hippotherium advanced, large Hippotherium “brachypus”. A specimen samples are similar in length, most have relatively from Samos (x) is extremely large plotting well above all slender mid-shaft widths (M3) versus mid-shaft depths other specimens. Another specimen from Samos (Hbra1) (M4) referred to as the “Esme Acakoy Effect” (Bernor et plots in the upper right portion of the Höwenegg ellipse al., 2003b; Scott, 2004). Baccinello V3 specimens IGF along with Pikermi specimens. 8192V and IGF 9397V (Plate 1, figs 1,2) show variability, Fig. 10 includes Log 10 ratio diagrams of MTIII. but have very similar Log10 ratio profiles (except M6, Fig. 10a compares Baccinello V3 IGF 8193V (Plate 1, proximal articular depth, with IGF 8192V being smaller) fig. 3) to a suite of Vallesian aged European and Turkish R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 185

Fig. 10 - Log 10 ratio diagrams of MTIII. Fig. 10a compares Baccinello V3 IGF 8193V to a suite of Vallesian aged European and Turkish MTIII. Fig. 10b plots IGF 8193V compared to Turolian members of the Hippotherium clade.

MTIII. The oldest and most primitive of the Central somewhat younger (10.5 Ma; Bernor et al., 2003b) than the European hipparions is the Inzersdorf Hippotherium Inzersdorf hipparion, having a greater maximum length primigenium which has a relatively shorter MTIII (M1) dimension (M1) than Inzersdorf, but exhibiting the same with particularly small proximal articular depth (M6) modest M3 versus M4 contrast as seen in the Inzersdorf dimensions and minimal M3 versus M4 midshaft “Esme Hippotherium. Rudabánya Hippotherium intrans has a Acakoy-like” dimensions Scott et al. (2005a) argued that very long maximum length (M1) and very wide midshaft the Inzersdorf Hippotherium is a primitive member of (M3) compared to its midshaft depth (M4). While being the genus and reflects the primitive morphology. Turkish shorter than H. intrans, Dinotheriensandes Hippotherium Cormohipparion sinapensis from Esme Acakoy is primigenium (mean) measurements reflect the same 186 Bollettino della Società Paleontologica Italiana, 50 (3), 2011 elevated M3, and proportionally lower M4 dimensions of H. intrans and has large proximal articular (M5 and M6) and distal articular (M10-M14) dimensions being a large population of Hippotherium primigenium. Baccinello V3 MTIII most closely tracks the Esme Acakoy Mean measurements, having particularly similar M1 proportions but accentuated narrow M3 versus deep M4 proportions. The length of these two taxa MTIII’s are very similar to the Höwenegg sample mean. Fig. 10b plots IGF 8193V MTIII compared to Turolian members of the Hippotherium clade. Here, Baccinello’s overall profile is most similar to Baltavar Hippotherium microdon, but Baccinello V3 has mostly larger and longer MTIII. In turn, Baccinello V3 most closely resembles the Polgardi mean population differing mostly in Polgardi’s larger dimensions of M3-M12 and M14. Pikermi MTIII is very similar to Höwenegg Hippotherium primigenium. Baltavar Hippotherium intrans is elongate and has robust midshaft proportions. Akkasdagi and Samos Hbra2 (Hippotherium “brachypus”) have longer MTIIIs than Baccinello V3 and Höwenegg but are remarkable for their robust mid- shaft and articular measurements: they are tremendously Fig. 11 - Bivariate plot of first Phalanx III for maximum length (M1) robust. Our analysis suggests that Baccinello V3 versus proximal articular width (M4). Hippotherium malpassii n. sp. is most closely related to Central European Hippotherium microdon and Polgardi, and in turn are derived from Hippotherium primigenium. Our analysis further suggests that Pikermi Hippotherium brachypus is also closely related to Central European smallest of this sample from Dorn Dürkheim and Baltavar Hippotherium primigenium. compare most closely with the smaller specimens from Fig. 11 includes a 1PHIII plot of maximum length Baccinello V3. The larger two from Rudabánya and the (M1) versus proximal articular width (M4). Again, most Dinotheriensandes compare most closely with Höwenegg of our sample plots within the Höwenegg ellipse, including and the single intermediate specimen, NHMBJH158 from several specimens from Rudabánya (r) and Eppelsheim Baccinello V3. The two largest Baccinello V3 specimens (e). There is 1 specimen from Baltavar (t) plotted within IGF 9390V and NHMB JH nonumb1are longer and have the ellipse and 1 below and 1 far to the left of the ellipse. greater proximal articular widths than the entire remainder The single Sinap (s) specimen plots below the ellipse of the sample under consideration. These results are not close to a specimen from Eppelsheim (e) and another strongly conclusive about Baccinello V3 1PHIII, but there from Rudabánya (r). There are 4 Baccinello V3 specimens seems to be a larger form, Hippotherium malpassii n. sp., plotted within the ellipse and two below the ellipse. The likely represented by IGF 9390V, NHMB JH nonumb1 two largest specimens, IGF 9390V (Plate 1, fig. 4) and and NHMB JH 158, with IGF 8194V, NHMB JH 129 and NHMB JH “nonumb” (=not numbered) group amongst IGF 9392V possibly representing a smaller taxon. As such, the largest Rudabánya specimens, while two more plot Hippotherium malpassii would appear to be advanced in lower in this ellipse; all of these 4 specimens are referable the elevated length, proximal (M5 and M6) and distal (M7) to Hippotherium malpassii. There is yet another group articular measurements compared to other members of the of Baccinello V3 specimens (IGF nonumb & NHMB Central European Hippotherium clade (H. primigenium, JH 151) that are amongst the smallest in our sample at H. intrans and H. microdon). the bottom and below the bottom of the ellipse. There The bivariate plots and Log 10 plots support a number are yet specimens that are below the ellipse that may be of conclusions about the Baccinello V3 hipparions. of another taxon (NHMB JH 129, IGF 8194V and IGF The P2 occlusal width (M3) versus occlusal length 9392V; Plate 1, figs 5-6). (M1) plot (Fig. 4a) suggests that the Baccinello V3 Fig. 12a includes Log 10 ratio diagrams of 1PHIII of hipparion is somewhat smaller than German Vallesian Baccinello V3 specimens IGF 8194V, IGF 9390V, IGF Hippotherium primigenium and compares in size to the 9392V, NHMB JH 129, NHMB JH 158 and NHMB JH Baltavar, Hungary hipparions. The P4 protocone width nonumb1. The log10 ratios are generally similar to one (M11) versus length (M10) plot demonstrates that the another with 3 specimens, IGF 8194V, NHMB JH 129 protocones are generally shorter and more rounded and IGF 9392V diverging from the other 4 in their shorter than in the Vallesian hipparions, comparing closely length measurements (M1 and M2) and IGF 9392V being with the Baltavar, and apparently the Hatvan, Hungary smaller yet in its proximal articular measurements (M5) hipparions. The astragalus plot (Fig. 5) demonstrates and distal tuberosity (M6) measurements. Fig. 12b includes that the Baccinello V3 hipparion groups in the lower 1PHIII from Dorn Dürkheim, Germany (MN11), Baltavar, part of the ellipse with Eppelsheim, Rudabánya, Hungary (MN12), the Dinotheriensandes, Germany Polgardi, Inzersdorf and Gaiselberg, with the Baltavar (MN9) and Rudabánya, Hungary (MN9). Interestingly, the sample being smaller yet. Baccinello’s calcaneum R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 187

Fig. 12 - Log 10 ratio diagrams of 1PHIII. Fig. 12a includes Log 10 ratio diagrams of 1PHIII of Baccinello specimens. Fig. 12b includes 1PHIII from Baccinello V3 and Dorn Dürkheim, Germany (MN11), Baltavar, Hungary (MN12), the Dinotheriensandes, Germany (MN9) and Rudabánya, Hungary (MN9).

(Fig. 6) maximum length (M1) versus maximum width again has Baccinello V3 plotting on the lower left border (M6) is small compared to most of this sample, and in of the ellipse close to the Baltavar hipparion. The MTIII particular German Vallesian hipparions. Baccinello V3 Log 10 ratio plot (Fig. 10b) shows a close comparison MCIII (Fig. 7) compares well in its maximum length between the Turolian hipparions from Baltavar and (M1) versus distal articular width (M11) with Esme Polgardi. The 1PHIII bivariate plots (Fig. 11) suggest two Acakoy, Inzersdorf, Sinap and much of Baltavar plotting taxa at Baccinello V3 and this is reinforced by the log10 together in the lower part of the Höwenegg ellipse. ratio plots (Figs 12a,b), with the larger sample being Baltavar has the smallest MCIIIs of the sample. The referable to Hippotherium malpassii. In conclusion, MCIII Log10 ratios (Figs 8a, b) reveal that Baccinello Baccinello V3 Hippotherium malpassii n. sp. compares V3 Hippotherium malpassii compare most closely with most closely with Hungarian (Baltavar and Polgardi) Turolian members of the clade from Polgardi, Baltavar species of Hippotherium, and is readily derived from and Dorn Dürkheim. The MTIII bivariate plot (Fig. 9) Central European Hippotherium primigenium. 188 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

SYSTEMATIC PALEONTOLOGY Diagnosis (autapomorphies in this lineage compared to Hippotherium primigenium are in italics) - A mod- Order Perissodactyla Owen, 1848 erate sized member of the Hippotherium lineage distin- Suborder Hippomorpha Wood, 1937 guished by having the skull with a long preorbital bar Superfamily Equoidea Hay, 1902 and lacrimal extending approximately ¾ of the distance Family Equidae Gray, 1821 from the anterior orbital rim to the posterior rim of the Subfamily Steinmann & Doderlein, 1890 preorbital fossa. The preorbital fossa has moderate me- dial depth, only slight posterior pocketing and a weakly Hippotherium malpassii n. sp. defined anterior rim (all known only in a single juve- nile skull). Adult maxillary cheek teeth are extremely Derivatio nominis - Named in honour of the late Nedo richly plicated, including even the anterior margin of Malpassi (Firenze) as a tribute to his long and enthusiastic the prefossette; pli caballins are richly and complexly survey activity in the Baccinello-Cinigiano area. ornamented; protocones varying from short ovals to small rounded morphology; hypoglyphs often only mod- Holotype - IGF 9400V - Right maxillary I1-I3, left erately deeply incised. Mandibular cheek teeth with maxillary P2-M3 (Fig. 13). substantially larger metaconid than metastylid with an elongate isthmus separating them; pre- and postflexids Hypodigm - See Tab. 2. are very complex. Permanent cheek teeth with a maxi- mum crown height approximately 55 mm. Deciduous Other referred specimens - See Tab. 2. maxillary cheek teeth have small, rounded protocones. Deciduous mandibular cheek teeth have sharply reduced Type Locality - Badlands south of Podere Firenze, to absent ectostylids. Metacarpal III and metatarsal III Baccinello-Cinigiano Basin, Arcidosso, Tuscany, Italy. are elongate-slender in morphology, otherwise close in UMTS coordinates 32TE693975 N4746299 (Fig. 1). size to H. primigenium and larger than H. microdon. The 1st phalanges III are similar to, but more slender than Age - Latest Miocene, Messinian Stage. The sedimen- H. primigenium; astragali and calcanea are generally the tary succession outcropping south of Podere Firenze, is size of H. primigenium and larger than Baltavar (Hun- lithostratigraphically correlated with the Ribaldella log gary) Hippotherium microdon. (Fig. 3), paleomagnetically calibrated to the C3An.2n Chron, bracketed between 6.733 and 6.436 Ma (Rook et Description - The Holotype IGF 9400V includes a al., 2011). Baccinello V3 Faunal Unit. MN13 zone in the right maxillary I1-3 and left maxillary P2-M3 (Fig. 13). The European Mammal Biochronological scale. incisors are worn and unremarkable in their morphology.

Fig. 13 - Hippotherium malpassii n. sp. Holotype from Arcidosso (Podere Firenze). IGF 9400V, left maxillary dentition (P2-M3) in a) labial, and b) occlusal views. Scale bar represents 5 cm. R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 189

Hippotherium malpassii n.sp. - Hypodigm Arcidosso - Podere Firenze / NHMB Loc. 11 2 IGF 9362V, left pyramidale; IGF 9363V, left lunatum; IGF 9365V, left capitatum; IGF 9366V, right 1PHIV; IGF 9398V, left P3-M3; IGF 9405V, left P ; NHMB 3 3 3 3 1 2 4 JH 142A-B, right M , left M ; NHMB JH 143, left P4; NHMB JH 229A-E, right P , left P , left M , right P , left P ; NHMB JH 231, left P2

Hippotherium malpassii n.sp. - Other referred material Unrecorded locality NHMB JH 2, right P3; NHMB JH 41, left M1; NHMB JH 42, left M2; NHMB JH 43, left P2; IGF 10876, right M1 Scansano - NHMB Loc 1-3, Podere Santa Croce 3 2 3 1 2 2 4 1 NHMB JH 124, right P , NHMB JH 126A-G, left P , left P , left M , left M , right P , right P , right M ; NHMB JH 129, left 1PHIII; NHMB JH 132, right P3, 4 NHMB JH 133, left P3, NHMB JH 134, left P Roccalbegna - Podere Le Pigne / NHMB Loc. 89 IGF 5286V, Cranium and mandibles, juvenile; IGF 9391V, left 1PHIII + 2PHIII; NHMB JH 127, left M3 Arcidosso - I Piscioli / NHMB Loc. 1 - 2 NHMB JH 149, right astragalus; NHMB JH 150, left calcaneum; NHMB JH 151, 1PHII; NHMB JH 153, right humerus; NHMB JH 152, right radius; NHMB

JH 158, left 1PHIII; NHMB JH 160A-M, left P2, left P4, left P3, left M1, left M2, left M3, right P2, right P4, right P3, right M1, right M2; NHMB JH 168, left MC III Cinigiano - Tesorino / NHMB Loc 4 - 7 NHMB JH 187, left M1-2; NHMB JH 190, left M1; NHMB JH 198, right calcaneum; NHMB JH nonumb1, left 1PHIII

Cinigiano - Ribaldella / NHMB Loc. 16 IGF 8192V, left MCIII; IGF 8193V, left MTIII; IGF 8195V, left astragalus; IGF 9376V, right MTIII; IGF 9377V, right great cuneiform; IGF 9394V, left mandibular

branch with P3-M3; IGF 9395V, astragalus; IGF 9397V, left MCIII; IGF 9387V, 2PH III; IGF 9388V, 2PHIII; IGF 9389V, 1PHIII; IGF 9390V, left 1PHIII; IGF 9401V, right P2; IGF 9402V, right P3; IGF 9403V, right M1; IGF 9404V, right M3; IGF 9406V, left M1; IGF 9407V, left M2; IGF 9408V, right P4; IGF 9409V, left 3 1 4 M ; IGF 9410V, left I ; IGF 9411V, right P4; IGF 9412V, right M1; IGF 9413V, right P ; NHMB JH 226, right navicolar

Hippotherium sp. Unrecorded locality

NHMB JH 16, left astragalus; NHMB JH 108, P4-M3; NHMB JH nonumb2, left 2PHIII; NHMB JH nonumb3, left P4 Scansano - Podere Santa Croce / NHMB Loc. 1-3 4 IGF 9385 V, left P ; NHMB JH 122A, left M1-M3; NHMB JH 122B, right M2- M3; NHMB JH 125C, left M2; NHMB JH 125C, right M3; NHMB JH 125C, left M1; NHMB JH 128, left astragalus; NHMB JH 130, right P3; NHMB JH 135, left M3; NHMB JH 234, right 2PHIII Roccalbegna - Podere Le Pigne / NHMB Loc. 89 1-2 2 IGF 5958V left P2; IGF 9417V, right M ; NHMB JH 140, right P ; NHMB JH 232, left 2PHIII Roccalbegna - Podere La Locca / Poggio Miliotto /La Crocina 4 3 IGF 9382V, right P ; IGF 9383V, right P ; NHMB JH 136, left M3; NHMB JH 137, left M2; NHMB JH 138, left M3; NHMB JH 233, right 2PHIII Arcidosso - I Piscioli / NHMB Loc. 1 – 2 IGF 9373V, fragmentary M1-2; IGF 9374V, fragmentary M1-2; NHMB JH 144, MC III (diafisis); NHMB JH 146, right 2PHIII; NHMB JH 147, left 2PHIII; NHMB

JH 148, right 2PHIII; NHMB JH 157, right astragalus; NHMB JH 155, right cuboid; NHMB JH 156, right cuboid; NHMB JH 161, right P4; NHMB JH 162, right 1 M1; NHMB JH 163, left P4; NHMB JH 165, right P4; NHMB JH 166, right M ; NHMB JH 167, left P4; NHMB JH 169, right MC III; NHMB JH 170, 1PHIII; NHMB JH 173, right scaphoid, NHMB JH 174, left cuboid; NHMB JH 176A, left M1; NHMB JH 176A, left M2; NHMB JH 176A, left M3; NHMB JH nonumb3, left M1 Cinigiano - Poggio Vannini / NHMB Loc. 8 IGF 2026V, astragalus, scaphoid, right 1PHIII; NHMB JH 212, left M1; NHMB JH 214, left M2 Cinigiano - Tesorino / NHMB Loc. 4 – 7 2 3 4 IGF 9384 V, right P4; NHMB JH 186, left P ; NHMB JH 188, left P ; NHMB JH 189, left P ; NHMB JH 191, left M3; NHMB JH 196, left P4; NHMB JH 197, left 1 3 M ; NHMB JH 202, right M ; NHMB JH 205, right magnum; NHMB JH 207, right M2; NHMB JH 218, right 2PHIII; Cinigiano - Ribaldella / NHMB Loc. 16 IGF 9367V, left scaphoid; IGF 9368V, left 1PHIII; IGF 9369V, left tibia; IGF 9370V, fragmentary 1PHIII; IGF 9371V, fragmentary 2PHIII; IGF 9372V, fragmentary 2PHIII; IGF 9375V, 2PHIII; IGF 9378V, right astragalus; IGF 9380V, right P3; IGF 9381V, right M1; IGF 9386V, 2PHIII; IGF 9396V, astragalus; IGF 9399V,

mandibular symphysis; IGF 9414V, left P4; IGF 9415V, right M1; IGF 9418V, fragmentary 2PHIII; NHMB JH 225, left astragalus; NHMB JH 227, right 2PHIII

Hippotherium sp. (small size) Roccalbegna - Podere Le Pigne / NHMB Loc. 89

IGF 9416V, left dp2-4 Arcidosso - Podere Firenze / NHMB Loc. 11 IGF 9364V, left scafoid Cinigiano - Poggio Vannini / NHMB Loc. 8 – 10

IGF 8194V, left 1PHIII + 2PHIII; NHMB JH 194, left MC III; NHMB JH 215, right M3; NHMB JH 216, left M2 Cinigiano - Tesorino / NHMB Loc. 4 – 7 st NHMB JH 195, left 1 PHIII; NHMB JH 199, left scaphoid; NHMB JH 206, left P4 Cinigiano - Ribaldella / NHMB Loc. 16 IGF 9392V, left 1PHIII, IGF 9393V, 2PHIII

Tab. 2 - List of studied Hippotherium material from Baccinello-Cinigiano Basin “V3” localities. 190 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

P2 has anterostyle elongate, pre- and postfossette borders are extremely complex, including the mesial and lingual borders of the prefossette, which is unusual; pli caballin is richly ornamented having 5 individual plis; protocone has a distinct mesial pli, primitive for Old World hip- parion, and has a short oval shape overall; hypoglyph is moderately deeply incised, an autapomorphy for the Hip- potherium lineage. P3’s morphology is as in P2, lacking anterostyle, lacking a pli on the protocone and having pli caballin with 8 individual plis; P4 is as in P3 except that the pli caballin has 3 individual plis; M1 is the most worn tooth in the row, and as a result the pre- and postfossette plis are shorter, but still show the same level of com- plexity as the premolars, including on the mesial bor- der of the prefossette, the remaining morphology is the same as in the premolars except the pli caballin which has 4 plis; M2 and M3 differ from M1 in having shorter, rounded protocones, both have complexly ornamented pli caballins which defy an accurate count and M3 is the only cheek tooth with a deeply incised hypoglyph that does not quite surround the hypocone. There are a number of localities that preserve maxil- Fig. 14 - Hippotherium malpassii n. sp. from Cinigiano (Ribaldella). lary cheek teeth that are similar to the type specimen. IGF 9406V (a) and IGF 9407V (b) are a left M1 and M2, respectively. Scale bar represents 2 cm. Amongst these, the most important is Ribaldella (also referred as “Ribardella”) which also preserves rare, com-

Fig. 15 - Hippotherium malpassii n. sp. from Scansano (Loc. 1-3). This sample includes a number of closely associated specimens: NHMB JH 126A, a left P2 (a, b); NHMB JH 126B, a left P3 (a, c); NHMB JH 126C, a left M1 (d, e); NHMB JH 126D, a left M2 (d); NHMB JH 126E, a right P2; NHMB JH 126F, a left P4 (f). Scale bar represents 4 cm. R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 191

Fig. 16 - Hippotherium malpassii n. sp. from Scansano (Loc. 1-3). Fig. 17 - Hippotherium malpassii n. sp. from Scansano (Podere 4 NHMB JH 133, left P4 in a) occlusal, and b) lingual views. Scale Santa Croce). NHMB JH 134, left P in occlusal (A) and labial (B) bar represents 2 cm. views. Scale bar represents 2 cm.

plete postcranial material. IGF 9406V and IGF 9407V height of more than 50 mm. This is a primitive feature (Figs 14a and 14b) are a left M1 and M2, respectively, and for Old World hipparions. Other salient morphological both are just beyond their middle stage-of-wear. Both of characters typical of H. malpassii n. sp. include: pro- these specimens are similar to the Type H. malpassii in tocone is a small rounded structure and hypoglyph is several morphologic features: the fossettes are extremely only moderately deeply incised. Due to the early stage- complex, and in the case of the M1, the prefossette even of-wear, the plications are not as developed as in the has a complex lingual margin and both the mesial border type specimen, but one can see that the pli caballin is of the prefossette and the distal border of the postfos- already complex and the mesial border of the prefos- sette are very complex. In both, there are multiple (3) pli sette is beginning to express complexity not found in caballins on both teeth. The protocone is elongate oval most hipparion taxa. in both, being less rounded than in the type specimen. A Roccalbegna (Fosso del Taffone) specimen, Hypoglyph is moderately deeply incised on the M1 and NHMB JH 140, is a right P2 in an advanced stage-of- more deeply incised on the M2. The morphological de- wear (crown height=13.2 mm). As expected this speci- tails of these two specimens secure their inclusion in the men has a larger rounded protocone and plications are hypodigm of Hippotherium malpassii. diminished. Still, while the pre- and postfossettes have The locality of Scansano “1-3” has dental material become labio-lingually more restricted, there remains that shares a number of similarities to the type specimen. complex plications on the mesial and distal borders of NHMB JH 126 includes a number of closely associat- the prefossette and the mesial border of the postfossette. ed specimens: NHMB JH 126A, a right P2; NHMB JH Taking into account this specimen’s advanced stage-of- 126B, a left P3; NHMB JH 126C, a left M1; NHMB JH wear, it has a completely expected occlusal morphology 126D, a left M2; NHMB JH 126E, a right P2; NHMB for H. malpassii. JH 126F, a left P4. These teeth are similar to the type Arcidosso “11” (type locality) has specimens that are specimen in all regards except the protocone presents as included in the hypodigm of Hippotherium malpassii. mostly a slightly smaller, round structure (Fig. 15a-f). NHMB JH 229 includes 5 associated maxillary cheek Another Scansano specimen, NHMB JH 126E, a right P2, is identical to the Type specimen P2 in all regards including the primitive (and often rare) occurrence of a mesial pli. Scansano “1-3” also has some lower cheek teeth that we refer to H. malpassii. NHMB JH 133 (Fig.

16a-b) is a left P4 characterized as having: metaconid and metastylid large and oblong-shaped, pre- and postflexids with complex margins, pli caballinid double, protostylid absent on the occlusal surface, and very small rounded structure on the mesiolabial wall of the tooth. The locality of Scansano Pod. S. Croce has a left P4 in early wear (NHMB JH 134; Fig. 17a-b) that also ex- hibits the morphologic hallmarks of Hippotherium mal- passii. This specimen has a mesostyle crown height of Fig. 18 - Hippotherium malpassii n. sp. from Arcidosso (Podere 44.3 mm, and given that it is in early wear, one would Firenze, type locality). NHMB JH 229D, right P2 in a) occlusal, not expect the crown to have had an original unworn and b) labial views. Scale bar represents 2 cm. 192 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

NHMB JH 229C is the most worn of these teeth (crown height = 41.9 mm), and it exhibits extremely complexly ornamented plications of the pre- and postfossettes and the pli caballin (Fig. 19a-b). The pli caballin nearly de- fies description in that there are two main trunks with several small plis emanating from the borders. As in the previously described NHMB JH 229 specimens, the hy- poglyph is deeply incised and the protocone is elongate oval. The left P4, NHMB JH 229E has all of the salient features described above for the two P3’s. Arcidosso “1-2”, a site located 1.5 Km NE of the type locality, has yielded a complete right mandibular denti- tion, NHMB JH 160A-F (note these are serially labeled

from P2-M3, but P3 and P4 have been reversed so that P3 is NHMB JH 160C and P4 is NHMB JH 160B; Fig. 20). The mandibular P4 (NHMB JH 160B) has the highest crown, 49.6 mm., which given the full expression of its occlusal morphology, and for that matter the full expres-

sion of M3’s occlusal morphology, suggests that it had a maximum unworn crown height was in excess of 50 mm. Fig. 19 - Hippotherium malpassii n. sp. from Arcidosso (Podere In itself, this suggests that the Arcidosso “1-2” hippa- Firenze, type locality). NHMB JH 229C, right M1 in a) labial, and rion was slightly advanced in its crown height compared b) occlusal views. Bar scale represents 2 cm. to the Type specimen of Hippotherium malpassii. The crowns of all these specimens are worn, but obviously not in a middle stage-of-wear where occlusal features are characteristically expressed. The metaconids and teeth: NHMB JH 229A, a right P3; NHMB JH 229B, a metastylids are essentially as in the Type specimen and left P3; NHMB JH 229C, a left M1; NHMB JH 229D, a specimens referred to H. malpassii above, but the pre- right P2; NHMB JH 229E, a left P4. The right P2 (NHMB and postflexids do not have complex margins and the pli JH 229D, Fig. 18a-b) is similar to the type Hippotherium caballinid is not expressed on the occlusal surface. This malpassii in its extremely complexly plicated fossettes leads us to refer the Arcidosso dental sample to Hippoth- and presence of 5 pli caballins, three of which are bifid erium aff. malpassii. at their apices. Likewise, the anterostyle is elongate, the There is a juvenile skull and mandible, IGF 5286V in hypoglyph is moderately deeply incised and the proto- the Florence collection that is derived from the Roccal- cone is oval-shaped, but lacks a mesial pli. NHMB JH begna “Podere le Pigne” locality. The skull (Fig. 21) is 229A and NHMB JH 229B, right and left P3s (respec- complete except for the snout and posterior cranium and tively), would appear to be somewhat higher crowned has right and left dP2-4 with M1s apparent in its crypts. than the type for its stage-of-wear, but retains charac- The labial view of the skull shows a clear, triangular teristic morphologic features of H. malpassii including shaped POF with a long preorbital bar. There is a sepa- complex plications of the pre- and postfossettes and ration of the skull along the lacrimal suture that is well multiple (4) pli caballins. They both differ however from posterior to the POF. The POF itself has medium depth, the type specimen in having an elongate-oval protocone only slight posterior pocketing, moderate medial depth and very deeply incised hypoglyph, this last feature per- and a prominent ventral and less prominent anterior rim haps being due to their early stage-of-wear. The left M1, and is placed well above the facial-maxillary crest. The

Fig. 20 - Hippotherium aff. malpassii from Arcidosso (Loc. 1-2). NHMB JH 160A-F, right mandibular dentition in lingual view. Scale bar represents 4 cm. R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 193

IOF is found above the limit between dP2 and dP3. The border of the prefossette, being less complex on the me- dP2 is in early wear with only the opposing borders of sial borders of the postfossettes and simple on the mesial the pre- and postfossettes showing some plications; the border of the prefossettess and distal border of the post- anterostyle is elongate, pli caballins are weakly double, fossette; protocones are as in the dP2. The hypoglyphs protocone is small and rounded. The dP3 and dP4 are rect- are very deeply incised, encircling the hypocones on all angular shaped, plications are only complex on the distal deciduous cheek teeth. The juvenile mandible includes

Fig. 21 - Hippotherium malpassii n. sp. from Roccalbegna (Podere le Pigne). IGF 5286V, juvenile skull in left lateral (a) and basal (b) views. Scale bar represents 5 cm. 194 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

dP2-4 with M1 emerging from its crypt. The teeth are as in the maxillary dentition being in early wear. Salient morphological features include: metaconids are gener- ally rounded while metastylids are more square-shaped; pre- and postflexids are poorly developed; ectoflexids are deep separating metaconid and metastylid.

Remarks - Baccinello V3 Hippotherium malpassii n. sp. is a late Turolian (MN 13) species of the Hippotherium lineage. The facial morphology is similar to Vallesian members of the clade except for some reduction in the height and depth of the POF as judged from a single juvenile specimen. The dentition exhibits even greater complexity of the pre- and postfossettes and number of Fig. 22 - Extension of the Central European Hippotherium clade pli caballins than Central European Vallesian hipparions from the area in Central Europe (A) where we recognize its origin (Bernor et al., 1988, 1997; Woodburne, 2007). Protocone during MN 9 (ca. 11 Ma), to Greece and Turkey (B) in MN 11/12 morphology differs from Vallesian members and compares (circa 8 Ma) and to the expansion of its range into central Italy (C) during MN13 (around 6.7 Ma). closely with Turolian members of Central European Hippotherium in its general short, round morphology. The postcranium of Hippotherium malpassii n. sp. is likewise similar to Central European Hippotherium, and C and already shows complex plication amplitudes of the in particular Turolian members of the clade. Koufos & fossettes and pli caballins. Maximum crown height would Vlachou (2005) Vlachou & Koufos (2009) have shown appear to have been near 55 mm and MPIIIs were not as that “Hipparion” (= Hippotherium to us) brachypus large as in later Vallesian hipparions from Eppelsheim and from Akkasdagi, Turkey as well as Pikermi and Samos, Höwenegg, Germany (Scott et al., 2005a). Hippotherium Greece has many of the skull features of Hippotherium had a demonstrated Vallesian age distribution that included but is larger, has a longer snout and has more massive Central Europe: Germany and the greater Pannonian metapodials than the Central European members of Basin of Austria and Hungary. By the medial Turolian, the clade and Pikermi H. brachypus s.s. We accept the the clade had extended its range as far as Greece and comparisons of these Greek species and include Pikermi Turkey to the South and East as well as Tuscany to the H. brachypus and Samos and Akkasdagi H. “brachypus” south and west. The origin of the Italian Hippotherium within the Hippotherium clade. malpassii n. sp. would best be posited as being the western Our current understanding of the Hippotherium clade’s Pannonian Basin with Baltavar, Polgardi and Hatvan all historical biogeography is that it is derived from North having hipparions that are morphologically very similar American Cormohipparion, its oldest known occurrence to the Baccinello V3 hipparion. Pikermi Hippotherium being in the Vienna Basin (Bernor et al., 1988, 1996; brachypus is found herein to compare closely with Woodburne, 2007). Central European Hippotherium is German Vallesian Hippotherium primigenium, particularly the sister taxon of Turkish Cormohipparion sinapensis. in metapodial proportions. Samos and Akkasdagi Earliest Hippotherium sp. is known from the Pannonian Hippotherium “brachypus” (our Hbra2) are related to

EXPLANATION OF PLATE 1

Hippotherium malpassii n. sp. and Hippotherium sp. from the paleomagnetically calibrated section of Ribaldella (Cinigiano).

Figs 1-4, 8-15 - Hippotherium malpassii n. sp. 1 - Left MCIII IGF 8192V in dorsal (a) and ventral (b) views; 2 - Left MCIII IGF 9397V in dorsal (a) and ventral (b) views; 3 - Right MTIII IGF 8193V in dorsal (a) and ventral (b) views; 4 - Left 1PHIII IGF 9390V in dorsal (a) and ventral (b) views; 8 - 2PHIII IGF 9387V in dorsal (a) and ventral (b) views; 9 - Left astragalus IGF 9395V in ventral (a) and dorsal (b) views; 10 - Left astragalus IGF 8195V in ventral (a) and dorsal (b) views;

11 - Left fragmentary mandible with P3-M3 IGF 9394V in labial (a) and occlusal (b) views; 12 - Right M3 IGF 9404V in labial (a) and occlusal (b) views; 13 - Right M1 IGF 9403V in labial (a) and occlusal (b) views; 14 - Right P3 IGF 9402V in labial (a) and occlusal (b) views; 15 - Right P2 IGF 9401V in labial (a) and occlusal (b) views.

Figs 5-7 - Hippotherium sp. (small size). 5 - Left 1PHIII IGF 9392V in dorsal (a) and ventral (b) views; 6 - Left 1PHIII IGF 8194V in dorsal (a) and ventral (b) views; 7 - 2PHIII IGF 9393V in dorsal (a) and ventral (b) views;

Scale bar = 10 cm. R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 Pl.195 1 196 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

Pikermi Hippotherium brachypus, being derived in its is an upper cheek tooth row, and 4 teeth (P4, M1, M2 and increased size and greater massivity of the postcranials. M3) are analyzed (DS 6). The Greek and Turkish members of the Hippotherium brachypus clade were likely vicariant from Central Results and dietary interpretation - Mesowear European and Italian Hippotherium by the early Turolian. frequencies and calibration factors are given in Table Fig. 22 provides a biogeographic summary of the Central 3 following the methodology of Kaiser & Solounias European MN 9 origin of the Hippotherium clade, its (2003) and Kaiser & Fortelius (2003). The cluster extension into Greece and Turkey by MN12 and the diagram (Fig. 23) illustrates the relationships between the extension of MN13 Hippotherium from the Pannonian datasets. The closer the data, the smaller the Euclidean Basin into Tuscany. distance (ED) at the branching point. The dendrogram shows four main clusters. Cluster 1 represents the most abrasion dominated feeding traits among grazers; PALEODIETARY RECONSTRUCTION cluster 2 comprises several grazers with less abrasive diets. Cluster 3 contains the intermediate-feeders, and Mesowear Analysis cluster 4 corresponds to the attrition-dominated end of Introductory Considerations - The mesowear the dietary spectrum and contains browsers only. The method, initially developed by Fortelius & Solounias two datasets of Hippotherium malpassii (Himal-BA) are (2000) has proven to be a powerful tool for reconstructing positioned in the cluster 2 together with the Eppelsheim paleodietary adaptations (Kaiser et al., 2000). The population of H. primigenium and appear closest to the mesowear method is based on facet development of cheek extant African grazing antelopes Kobus ellypsiprymunus, tooth occlusal surfaces. The degree of facet development Redunca redunca, Hippotragus equinus and Hippotragus reflects the relative proportions of tooth-to-tooth contact niger. The waterbuck, Kobus ellipsiprymnus, clusters (attrition) and food to tooth contact (abrasion). Attrition closest with Hippotherium primigenium (Eppelsheim) creates facets while abrasion obliterates them. The and Hippotherium malpassii. The waterbuck frequents entire exposed surface of a tooth is affected by wear, reedbeds and shrubby growth and also forage into but mesowear analysis so far has focused on the cutting woodlands. The , however, usually live in proximity edges of cheek tooth enamel surfaces where the buccal to water (Nowak, 1999). According to Kingdon (1982) the wall (ectoloph in perissodactyls) meets the opposing major sources of food are medium and short grasses like occlusal plane. Kaiser & Solounias (2003) extended the Andropgogon, Brachiaria, Cenchrus, Chloris, Cynodon, mesowear method to four upper cheek teeth (P4-M3), and Cyperaceae, Cymbopogon, Dichrostachys, Digitaria, thus expanded the methodology to intrinsically larger Heteropogon, Hyparrhenia and Panicum species, reeds and sample sizes. The ”extended” mesowear method (Kaiser rushes (Phragmites and Typha). To a lesser extent browse & Solounias 2003) was applied to maxillary premolar and like leaves and fruit are eaten (Kingdon, 1982). Based molar teeth. Kaiser & Fortelius (2003) further extended on cluster analysis, the prevailing dietary regime of the the mesowear method to lower equid cheek teeth. We Baccinello V3 hipparion assemblage is therefore suggested investigate the upper and lower cheek tooth sample from to be equivalent to that of the waterbuck. Baccinello V3. After exclusion of unworn specimens Principal components analysis (PCA; Fig. 24) plots and specimens in early wear, as well as specimens in the Baccinello V3 bulk sample (Himal-BA-txtm) of advanced wear, 16 upper premolars and molars (P4, M1, Hippotherium malpassii in the transitional factor space M2 and M3) (DS 1) and 3 lower premolars and molars between the intermediate-feeders in kernel 3 and the (DS 2) were available for analysis (Tab. 3). We compute less abrasion dominated grazers (kernel 2). We realize calibrated scores (DS 3) for lower cheek teeth (after Kaiser that within the factor space H. malpassii is slightly & Fortelius, 2003). Combined lower and upper cheek closer to the intermediate-feeder kernel (k3), where in tooth scores are represented by DS 4 (absolute) and DS 5 fact the African impala (Aepyceros melampus) would be (percentages). We independently investigate the holotype the closest analogue species in terms of the mesowear of Hippotherium malpassii (IGF 9400V). This specimen signature. The impala is an ecotone species living in

dataset (DS) l h s r b n 1 tx 4 16 4 12 0 16 2 tm 1 2 0 3 0 3

CfHp 0.11 1.18 2.82 0.96 0.00 3 tm (CfHp) 0.11 2.36 0.00 2.88 0 4 tx+tm (CfHp) 4.11 18.36 4.00 14.88 0 %l %h %s %r %b 5 Tx+tm (CfHp) [%] 18.29 81.71 21.19 78.81 0 6 tx (IGF9400V-BA only) 0 100 25 75 0 4 Tab. 3 - Absolute and calibrated mesowear scorings (datasets 1-6). DS = #of dataset, tx = maxillary tooth, tm = mandibular tooth, l = low occlusal relief, h = high occlusal relief, s = sharp cusp, r = round cusp, b = blunt cusp. DS 1 = absolute mesowear counts maxillary teeth. DS

2 = absolute mesowear counts mandibular teeth. CFHp = Calibration factor for lower teeth of hipparionines as established for Hippotherium primigenium by Kaiser & Fortelius (2003). DS 3 = absolute mesowear counts of mandibular teeth calibrated by CFHp. DS 4 = sum of calibrated counts in upper and lower teeth. DS 5 = calibrated frequency (percentage) used for analysis. DS 6 calibrated frequency (percentage) of mesowear variables of the type specimen of Hippotherium malpassii (IGF9400V-BA). R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 197 light woodland with little undergrowth and grassland of diet. The status of the impala as an intermediate-feeder low to medium height. The impala is found frequently is uniformly acknowledged by all workers who have in grass dominated open environments like bushland evaluated ungulate dietary classifications (Janis, 1988; and Acacia savannahs but also in Acacia forests and Janis & Erhadt, 1988; Hofmann, 1989; Van Wieren, other deciduous woodlands. Impalas live near water and 1996; Fortelius & Solounias, 2000; Gagnon & Chew, shift feeding from grazing to browsing depending on the 2000; Sponheimer et al., 2003). VanWieren (1996) gives season (Estes, 1991). Aepyceros melampus has a diet a grass/browse ratio of 60/40% based on Breymeyer & that may be considered to be a perfectly balanced mixed Van Dyne (1980) and Skinner & Smithers (1990). Based

Fig. 23 - Hierarchical cluster diagram comparing a set of 27 ”typical” extant species from Fortelius & Solounias (2000) with the Baccinello V3 samples of Hippotherium malpassii “IGF9400V-BA” and “Himal-txtm” (set bold). Distance = normalized Euclidean distance (root- mean-squared difference). Circle = extant browser, rectangle = extant intermediate-feeder, triangle = extant grazer. Cluster 1 represents the most abrasion dominated feeding traits among grazers; cluster 2 comprises several grazers with less abrasive diets. Cluster 3 contains the intermediate-feeders, and cluster 4 corresponds to the attrition-dominated end of the dietary spectrum and represents browsers only. Comparative datasets for fossil Old World Hippotherium are: Hippotherium primigenium from Eppelsheim (Hipri-EP; Kaiser, 2003), Hippotherium primigenium from Höwengegg (Hipri-HO; Kaiser, 2003), Hippotherium aff. primigenium from MN11 of Dorn-Dürkheim (Hipri-DD; Kaiser et al., 2003), Hippotherium kammerschmittae from Dorn-Dürkheim (Hikam-DD; Kaiser et al., 2003, Hippotherium sumegense (Hisum-SU; Bernor et al., 1999), Hippotherium “microdon” from Baltavar (MN 12; Himic-BA; Kaiser & Bernor, 2007) and Hippotherium intrans from Rudabánya (Hiint-RU; Bernor et al., 2003a). 198 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

Fig. 24 - Principal components factor plot based on mesowear parameters %s, %r and %b and a set of 27 “typical” species from Fortelius & Solounias (2000). Dataset from Baccinello V3 Hippotherium malpassii are set bold. Hipri-EP = Hippotherium primigenium from Eppelsheim (MN 9), Hipri-HO = Hippotherium primigenium from Höwenegg (MN 9), Hiint-RU = Hippotherium intrans from Rudabánya (MN 9), Hisum-SU = Hippotherium sumegense from Sümeg (MN 10), Hipri-DD = Hippotherium aff. primigenium from Dorn-Dürkheim (MN 11), Hicam-DD = Hippotherium kammerschmittae from Dorn-Dürkheim, Himic-BA = Hippotherium “microdon” from Baltavar (MN 12). Capital initials = browser, lower case initials = grazer, mixed capital and lower case initials = intermediate-feeder. Grazers (k1): bb = Bison bison, cs = Ceratotherium simum, dl = Damaliscus lunatus, eb = burchellii, eg = Equus grevyi. Grazers (k2): ab = Alcelaphus buselaphus, ct = Connochaetes taurinus, he = Hippotragus equinus, hn = Hippotragus niger, ke = Kobus ellipsiprymnus, rr = Redunca redunca. Intermediate-feeders (k3): Ca = Capricornis sumatraensis, Cc = Cervus canadensis, Gg = Gazella granti, Gt = Gazella thomsoni, Me = Aepyceros melampus, Om = Ovibos moschatus, To = Taurotragus oryx, Ts = Tragelaphus scriptus. Browsers (k4): AA = Alces alces, DB = Diceros bicornis, DS = Dicerorhinus sumatrensis, GC = Giraffa camelopardalis, OH = Odocoileus hemionus, OJ = Okapia johnstoni, OV = Odocoileus virginianus, RS = Rhinoceros sondaicus.

on stable isotopes Sponheimer et al. (2003) found a C4 The fossil hipparion population closest to Hippotherium grass ration of 51% in the Southern African population malpassii is Hippotherium “microdon” from Baltavar (MN of A. melampus and 54% in the East African population 12; Himic-BA; Kaiser & Bernor, 2007). The Eppelsheim of the species. Based on faeces, Codron et al. (2005) population of H. primigenium from the MN 11; Hipri- suggest a grass/browse ratio of 50/50% for A. melampus. EP; Kaiser, 2003) classifies closer to the grazer kernel Among the extant comparison species the impala is the k2, than to the intermediate-feeder kernel k3. From the only that consistently eats substantial amounts of grass. PCA it also becomes evident, that the hipparions from The remaining intermediate-feeders in this comparison Sümeg (Hisum-SU), Dorn-Dürkheim (Hiprim-DD) and have far less extensive grass components in their diets. Rudabánya (Hiint-RU) are very similar in their mesowear The impala has therefore the most abrasive diet among all signatures, as indicated by their crowded classification in known intermediate-feeders and is consistently classified the center of the intermediate-feeder factor kernel (k3). next to the grazers in all mesowear evaluations. As in Fig. 22, the Höwenegg H. primigenium (Hipri-HO) R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 199 and H. kammerschmittae from Dorn-Dürkheim (Hikam- from the Miocene Siwaliks of Pakistan exhibit oxygen DD) are the only populations that classify in the browser ranges of approximately 2.5‰ per tooth, reflecting a kernel (k4). seasonal environment comparable to a monsoon forest We thus consider it likely, that Baccinello V3 (Nelson, 2005). The small range for the Baccinello Hippotherium malpassii had a very similar diet as V3 Hippotherium malpassii tooth suggests either a Baltavar Hippotherium “microdon” which was not considerably less seasonal rainfall and temperature very different in the proportion of abrasive agents regime, or alternatively more permanent, buffered water as compared to the impala’s diet. Eppelsheim H. sources that reflect the average annual precipitation primigenium also had a more abrasive dominated diet. isotopic values rather than the full range. Combined, the It is remarkable, that all other Old World hipparion carbon and oxygen values of this equid tooth are consistent assemblages [Hippotherium sumegense (Bernor et with a forest-dweller in a mesic environment. al., 1999), Hippotherium kammerschmittae and H. primigenium from Dorn-Dürkheim (Kaiser et al., 2003), H. primigenium from Höwenegg (Kaiser, 2003), and H. CONCLUSIONS intrans from Rudabánya (Bernor et al., 2003a)] had diets corresponding to the intermediate feeding and browsing The Baccinello V3 faunas include a dominant, part of the dietary spectrum. Our work on paleodiet moderately large hipparion which we refer to Hippotherium suggests that Hippotherium malpassii was a mixed feeder malpassii n. sp. not as committed to browsing as many Central European Our study of the Baccinello V3 hipparion provides species of Hippotherium, but similar to others such as us new insights into the Hippotherium clade and its Eppelsheim, Germany Hippotherium primigenium and members. We agree with previous researchers that Baltavar, Hungary Hippotherium microdon. Hippotherium originated in Central Europe following the immigration of North American Cormohipparion Isotopic analysis (Bernor et al., 1996; Woodburne, 2007). Hippotherium Carbon and oxygen stable isotopic analyses of tooth underwent a modest evolutionary radiation in Central enamel can be informative for reconstructing paleodiets Europe between 11.2 and approximately 8 m.y. Species and paleohabitats. Carbon isotopes distinguish between of Central European Hippotherium were adapted to diets of graze (C4 grasses and sedges) and browse (C3 living in rich mesophytic forests and grassy woodlands plants such as trees and shrubs) due to differences in with subtropical to warm temperate climate (Bernor et photosynthetic pathways between C4 and C3 plants. al., 1986) and adapted to mixed locomotor strategies that Additionally, plants yield a wide range of carbon and included running and leaping and springing (Bernor et oxygen isotopic values in response to differences in al., 1997). irradiance and water stress and can therefore be useful The Vienna and Pannonian Basins included the in distinguishing closed canopy forests, woodlands, richest diversity of Hippotherium species including: H. and open habitats. Increasingly open habitats, where primigenium, H. intrans and H. microdon (Bernor et plants are under high stress, yield the highest carbon al., 1993a,b; Bernor et al., 2003a; Scott et al., 2005a,b). and oxygen values. Isotopic values of tooth enamel The regional regression of the Central Paratethys in the reflect the values of plants eaten and water drunk by Late Miocene established geographic connections from an individual during tooth development. Thus isotopic the Pannonian Basin southward through the Balkans to analyses of fossil teeth are useful in reconstructing an Greece. The Late Miocene faunas of Pikermi and Samos, extinct species’ diet and habitat. Additionally, oxygen Greece and Akkasdagi, Turkey included immigrant isotopes can be used in reconstructing paleoclimates, members of the Hippotherium clade. We believe that the for water sources such as rivers and lakes respond to oldest member of this eastern Mediterranean - southwest seasonal fluctuations in precipitation and temperature. Asian clade is Hippotherium brachypus from Pikermi, Hypsodont equid teeth that take over a year to develop Greece (Koufos, 1987a, b), which Bernor believes is circa can be sampled multiple times along the length of the 8 Ma. (Kostopolous & Bernor, in press). The Samos and tooth to capture annual seasonal variability in oxygen Akkasdagi Hippotherium is larger and has more massively isotopes, and hence reflect the degree of seasonality in built MPIIIs and we believe likely to be a species derived that paleoenvironment. δ from Pikermi Hippotherium brachypus. This species We sampled a single Hippotherium malpassii tooth, would be younger than Hippotherium brachypus being P3 IGF 9400V, three times from crown to root. After between 8 and 7 Ma. standard treatment, isotopic ratios are presented in the Our paleoecological reconstructions demonstrate that per mil (‰). In modern habitats, carbon values below Baccinello V3 Hippotherium malpassii n. sp. remained -13‰ represent closed forests (Quade et al., 1995). tied to warm mesophytic subtropical woodland habitats Miocene carbon values are enriched by 1.5‰ relative into latest Miocene times. By the latest Miocene, to modern values due to changes in the atmosphere hipparions had dispersed across all of Eurasia and Africa since the industrial revolution (Marino & McElroy, and included numerous clades, many of them being 1991). Therefore, carbon values of -12‰ or lower in adapted to dry, seasonal open country habitats. Some of the Miocene record are indicative of forest habitats. these in Africa and South Asia had shifted their dietary The δ13C values for the three Baccinello V3 samples all adaptation to eating coarse, C4 grasses and had adaptations fall below this cutoff, ranging from -13.2‰ to -12.6‰. strongly divergent from Central European Hippotherium. The δ18O values ranged from -3.6‰ to -3.1‰, for an Baccinello V3 represents the last known stand of the intra-tooth range of 0.5‰. As a comparison, equid teeth Hippotherium lineage. 200 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

ACKNOWLEDGMENTS Bernor R.L. & Armour-Chelu M. (1999). Towards an evolutionary history of African hipparionine horses. In Bromage T. & For the access to collections, we thank B. Engesser and L. Schrenk F. (eds), African Biogeography, Climate Change and Costeur, curators of the Basel Naturhistorisches Museum and E. Early Hominid Evolution, Oxford University Press, New York: Cioppi and S. Dominici, curators of the Florence Natural History 189-215. Museum (Geology and Paleontology Section). Bernor R. L., Armour-Chelu M., Gilbert H., Kaiser T.M. & Schulz The Baccinello collection housed at the Basel Naturhistorisches E. (2010). Equidae. In Werdelin L. & Sanders W.L. (eds), Museum has been increased since late 1950’s thanks to the precious Cenozoic of Africa. University of Press, and passionate work of the eminent paleontologist Prof. Johannes Berkeley: 685–721. Hürzeler. Bernor R.L., Armour-Chelu M., Kaiser T. M. & Scott R.S. (2003a). We deeply acknowledge M. Benvenuti and M. Papini An Evaluation of the Late MN 9 (Late Miocene, Vallesian Age), (University of Florence), for their contribution to the geological Hipparion Assemblage from Rudabánya (Hungary): Systematic mapping and understanding the sedimentary evolution of the Background, Functional Anatomy and Paleoecology. Coloquios Baccinello-Cinigiano basin, and Valerio Malpassi (Florence) for de Paleontología, Vol. Ext., 1: 35-45. his painstaking field surveys in the Cinigiano area. Early phases Bernor R.L., Fahlbusch V., Andrews P., De Bruijn H., Fortelius M., of the preparation of this paper started in summer 2002, when L. Rögl F., Steininger F.F. & Werdelin L. (1996). The evolution Rook visited the Laboratory of Evolutionary Biology at Howard of Western Eurasian Neogene mammal faunas: a chronologic, University (Washington DC, USA) holding a CNR-NATO Outreach systematic, biogeographic, and paleoenvironmental synthesis. In Fellowship. This paper is based upon work supported by grants Bernor R.L., Fahlbusch V. & Mittman H.W. (eds), The Evolution from the U.S. National Science Foundation under NSF Award of Western Eurasian Neogene Mammal Faunas, Columbia (#BCS-0321893) to F.C. Howell and T.D. White (RHOI; Berkeley University Press, New York: 449-469. University), EAR0125009 to R.L. Bernor, and the University of Bernor R.L., Fortelius M. & Rook L. (2001). Evolutionary Florence (Fondi di Ateneo) to L.Rook. biogeography and paleoecology of the “Oreopithecus bambolii Faunal Zone” (Late Miocene, Tusco-Sardinian Province). Bollettino della Società Paleontologica Italiana, 40: 139-148. REFERENCES Bernor R.L. & Franzen J. (1997). The hipparionine horses from the Turolian Age (Late Miocene) locality of Dorn Dürkheim, Abbazzi L. (2001). Cervidae and Moschidae (Mammalia, Germany. Courier Forschungsinstitut Senckenberg, 197: 117- Artiodactyla) from the Baccinello V3 faunal assemblage 185. (Late Miocene, Late Turolian, Grosseto, Central Italy). Rivista Bernor R.L. & Haile-Selassie Y. (2009). Equidae. In Haile-Selassie Italiana di Paleontologia e Stratigrafia, 107: 10-123. Y. & WoldeGabriel G. (eds), Ardipithecus kadabba: Late Abbazzi L., Delfino M., Gallai G., Trebini L. & Rook L. (2008). Miocene Evidence from the Middle Awash, Ethiopia, University New data on the vertebrate assemblage of Fiume Santo of California Press, Berkeley and Los Angeles: 397-428. (North-western Sardinia, Italy), and overview on the Late Bernor R. L. & Harris J.M. (2003). Systematics and evolutionary Miocene Tusco-sardinian paleobioprovince. Palaeontology, biology of the Late Miocene and Early Pliocene hipparionine 51: 425-451. horses from Lothagam, Kenya. In Harris J.M. & Leakey M.G. Agustí J., Carbrera L., Garcés M., Krijgsman W., Oms O. & Parés (eds), Lothagam: Th e Dawn of Humanity in Eastern Africa, J.M. (2001). A calibrated mammal scale for the Neogene of Columbia University Press, New York: 387-438. Western Europe. State of the Art. Earth Sciences Review, 52: Bernor R.L., Heissig K. & Tobien H. (1987). Early Pliocene 247-260. Perissodactyla from Sahabi, Libya. In Boaz N.T., El-Arnauti Angelone C. & Rook L. (2011). Alilepus meini nov. sp. (Leporidae, A., Gaziry A., de Heinzelin J. & Boaz D.D. (eds), Neogene Lagomorpha) from the Early Messinian of Tuscany (central- Paleontology and Geology of Sahabi, Alan R. Liss, New York: western Italy). Géobios, 44: 151-156. 233-254. Armour-Chelu M. & Bernor R.L. (2011). Equidae. In Harrison T. Bernor R.L. & Hussain S. T. (1985). An assessment of the (ed.), Paleontology and geology of Laetoli: Human evolution systematic, phylogenetic and biogeographic relationships in context (Fossil hominins and the associated fauna, Vol. 2), of Siwalik hipparionine horses. Journal of Vertebrate Springer Dordrecht: 295-326. Paleontology, 5: 32-87. Armour-Chelu M., Bernor R.L. & Mittmann H.-W. (2006). Bernor R.L. & Kaiser T.M. (2006). Systematics and paleoecology Hooijer’s hypodigm for “Hipparion” cf. ethiopicum (Equidae, of the earliest Pliocene equid Eurygnathohippus hooijeri n. Hipparioninae) from Olduvai, Tanzania and comparative sp. from Langebaanweg, South Africa. Mitteilungen aus dem material from the East African Plio-. In Nagel Hamburgischen Zoologischen Museum und Institut, 103: D. & van den Hoek Ostende L.W. (eds), Abhandlungen der 147-183. Bayerischen Akademie der Wissenschaften, 30:15-24. Bernor R.L., Kaiser T.M., Kordos L. & Scott R.S. (1999). Benvenuti M., Bertini A. & Rook L. (1994). Facies analysis, Stratigraphic Context, Systematic Position and Paleoecology vertebrate paleontology and palynology in the Late Miocene of Hippotherium sumegense Kretzoi, 1984 from MN 10 (Late Baccinello–Cinigiano basin (southern Tuscany). Memorie della Vallesian of the Pannonian Basin). Mitteilungen der Bayrischen Società Geologica Italiana, 48: 415-423. Staatssammlung für Paläontologie und historische Geologie, Benvenuti M., Papini M. & Rook L. (2001). Mammal biochronology, 39: 1-35. UBSU and paleoenvironment evolution in a post-collisional Bernor R.L., Kaiser T.M. & Nelson S.V. (2004a). The oldest Ethiopian basin: evidence from the Late Miocene Baccinello-Cinigiano hipparion (Equinae, Perissodactyla) from Chorora: Systematics, basin in southern Tuscany. Italy. Bollettino della Società paleodiet and paleoclimate. Courier Forschunginstitut Geologica Italiana, 120 (1995): 97-118. Senckenberg, 246: 213-226. Benvenuti M., Papini M. & Testa G. (1999). Sedimentary facies Bernor R.L., Kaiser T.M. & Wolf D. (2008). Revisiting Sahabi equid analysis in paleoclimatic reconstructions. Examples from the species diversity, biogeographic patterns and diet preferences. upper Miocene–Pliocene successions of southcentral Tuscany In El-Arnauti A., Salem M., Pavlakis P. & Boaz N. (eds), (Italy). In Agustí J., Rook L. & Andrews P. (eds.), The Evolution Circum- Mediterranean Geology and Biotic Evolution During of Neogene Terrestrial Ecosystems in Europe, Cambridge the Neogene Period: The Perspective from Libya. Garyounis University Press, Cambridge: 355-377. Scientific Bulletin, Special Issue, 5: 159-167. Bernor R.L. (1985). Systematic and evolutionary relationships of Bernor R.L., Kordos L., Rook L., Agustí J., Andrews P., Armour- the Hipparionine horses from Maragheh, Iran (Late Miocene, Chelu M., Begun D.R., Cameron D.W., Damuth J., Daxner- Turolian Age). Paleovertebrata, 15: 173-269. Höck G., de Bonis L., Fejfar O., Fessaha N., Fortelius M., Bernor R.L. (2007). The latest Miocene Hipparionine (Equidae) Franzen J., Gasparik M., Gentry A., Heissig K., Hernyak N., from Lemudong’o, Kenya. Kirtlandia, 56: 148-151. Kaiser T., Koufos G.D., Krolopp E., Jánossy D., Llenas M., R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 201

Meszáros L., Muller P., Renne P., Rocek Z., Sen S., Scott R., Eisenmann V., Alberdi M.T., De Giuli C. & Staesche U. (1988). Szyndlar Z., Topál Gy., Ungar P.S., Utescher T., Van Dam Studying Fossil Horses. Volume I - Methodology. In Woodburne J.A., Werdelin L. & Ziegler R. (2004b). Recent advances on M.O. & Sondaar P.Y. (eds), Collected papers after the “New multidisciplinary research at Rudabánya, Late Miocene (MN9), York International Hipparion Conference, 1981”, Brill, Leiden: Hungary: A compendium. Palaeontographica Italica, 89: 3-36. 1-71. Bernor R.L., Kovar J., Lipscomb D., Rögl F. & Tobien H. (1988). Engesser B. (1983). Die jungtertiären Klinsäuger des Gebietes Systematic, stratigraphic and paleoenvironmental contexts of der Maremma (Toskana, Italien. 1.Teil: Gliridae (Rodentia, first appearingHipparion in the Vienna Basin, Austria. Journal Mammalia). Eclogae Geologicae Helvetiae, 76: 763-780. of Vertebrate Paleontology, 8: 427-452. Engesser B. (1989). The Late Tertiary small mammals of the Bernor R.L., Mittmann H.-W., Kretzoi M. & Tobien H. (1993). Maremma region (Tuscany, Italy): II part. Muridae and A preliminary systematic assessment of the Rudabánya Cricetidae (Rodentia, Mammalia). Bollettino della Società hipparions. Mitteilungen der Bayerischen Staatssammlung fur Paleontologica Italiana, 29: 227-252. Paläontologie und historische Geologie, 33: 1-20. Eronen J. & Rook L. (2004). The Mio-Pliocene European primate Bernor R.L., Qiu Z, & Hayek L. (1990). Systematic revision of fossil record: dynamics and habitat tracking. Journal of Human Chinese hipparion species described by Sefve, 1927. American Evolution, 4: 323-341. Museum Novitates, 2984: 1-60. Estes R. (1991). The Behavior Guide to African Mammals, including Bernor R.L. & Scott R.S. (2003). New interpretations of the Hoofed Mammals, Carnivores, Primates. Los Angeles, systematics, biogeography and paleoecology of the Sahabi University of California Press: 1-611. hipparions (latest Miocene), Libya. Geodiversitas, 25: 297-319. Esu D. & Girotti O. (1989). Late Miocene and Early Pliocene Bernor R.L., Scott R. S., Fortelius M., Kappelman J. & Sen S. continental and oligohaline molluscan faunas of Italy. Bollettino (2003b). Systematics and evolution of the Late Miocene della Società Paleontologica Italiana, 28: 253-263. Hipparions from Sinap, Turkey. In Fortelius M., Kappelman Fortelius M. & Solounias N. (2000). Functional characterization J., Sen S. & Bernor R.L. (eds), The Geology and Paleontology of ungulate molars using the abrasion-attrition wear gradient: of the Miocene Sinap Formation, Turkey, Columbia University a new method for reconstructing paleodiets. American Museum Press, New York: 220-281. Novitates, 3301: 1-36. Bernor R.L., Scott R.S. & Haile-Selassie Y. (2005). A contribution Gabunia L. (1959). The Hipparion history (according to SSSR to the evolutionary history of Ethiopian hipparionine horses: Neogene material). Izdatelstvo Akademii Nauk SSSR, Moscou: Morphometric evidence from the postcranial skeleton. 5-565 [in Russian]. Geodiversitas, 27: 133-158. Gagnon M. & Chew A.E. (2000). Dietary preferences in extant Bernor R.L., Tobien H., Hayek L.-A. C. & Mittmann H.-W. (1997). African Bovidae. Journal of Mammology, 81: 490-511. Hippotherium primigenium (Equidae, Mammalia) from the late Gallai G. & Rook L. (2011). Propotamochoerus provincialis Miocene of Höwenegg (Hegau, Germany). Andrias, 10: 1-230. (Gervais, 1859) (Suidae, Mammalia) from the latest Miocene Bernor R.L., Tobien H. & Woodburne M.O. (1989). Patterns of Old (late Messinian; MN13 ) of Monticino Quarry (Brisighella, World hipparionine evolutionary diversification.In Lindsay E., Emilia-Romagna, Italy). Bollettino della Società Paleontologica Fahlbusch V. & Mein P. (eds), European Neogene Mammal Italiana, 50: 29-34. Chronology. Plenum Press, New York: 263-319. Getty R. (1982). The Anatomy of Domestic Animals. Saunders Bernor R.L. & White T.D. (2009). Systematics and biogeography Company, Philadelphia: 1-1211. of “Cormohipparion” africanum, early Vallesian (MN9, ca. Gilbert H.W. & Bernor R.L. (2008). Equidae. In: Gilbert H.W. & 10.5 MA) of Bou Hanifa, Algeria. Museum of Northern Arizona Asfaw B. (eds), Homo erectus - Pleistocene Evidence from Bulletin, 65: 635-653. the Middle Awash, Ethiopia. University of California Press, Bernor R.L., Woodburne M.O. & Van Couvering J.A. (1980). A Berkeley: 133-166. contribution to the chronology of some Old World Miocene Gillet S., Lorenz H.G. & Woltersdorf F. (1965). Introduction à faunas based on hipparionine horses. Géobios, 13: 705-739. l’étude du Miocène supérieur de Baccinello (environs de Bossio A., Costantini A., Foresi L., Mazzei R., Monterforti B., Grosseto, Italie). Bullettin Service Carte Géologique de Alsace Salvatorini G. & Sandrelli F. (1991). Notizie preliminari et Lorraine, 18: 31-42. sul Pliocene del bacino del medio Ombrone e della zona di Gray J.E. (1821). On the natural arrangement of vertebrose animals. Roccastrada. Atti della Società Toscana di Scienze Naturali, London Medical Repository, 15(1): 296-310. Memorie A, 98: 259-269. Gromova V. (1952). Le genre Hipparion (transalded from Russian Breymeyer A.I. & Van Dyne G.M. (1980). Grasslands, systems by St. Aubin). Bureau de Recherches Geologiques et Minieres, analysis and man. Cambridge: Cambridge University Press. Orléans, CEDP, 12: 1-288. Casanovas-Vilar I., van Dam J.A. Moyá-Solá S. & Rook L. Gruvaeus G. & Wainer H. (1972). Two additions to hierarchical (2011). Late Miocene insular mice from the Tusco-Sardinian cluster analysis. British Journal of Mathematical and palaeobioprovince provide new insights on the palaeoecology Statatistical Psychology, 25: 200-206. of the Oreopithecus faunas. Journal of Human Evolution, 61: Hartigan J.A. (1975). Clustering algorithms, (John Wiley & Sons), 42-49. New York. Casanovas-Vilar I., van Dam J.A., Trebini L. & Rook L. (2011). Hay O.P. (1902). Bibliography and catalogue of the fossil The rodents from the Late Miocene Oreopithecus-bearing site Vertebrata of . U.S. Geological Survey Bulletin, of Fiume Santo (Sardinia, Italy). Géobios, 44: 173-187. 179: 1-868. Chesi F., Delfino M. & Rook L. (2009). Late Miocene Mauremys Hofmann R.R. (1989). Evolutionary steps of ecophysiological (Testudines, Geoemydidae) from Tuscany (Italy): evidence adaptation and diversification of ruminants: a comparative view of terrapin persistence after a mammal turnover. Journal of of their digestive system. Oecologia, 78: 443-457. Paleontology, 83: 379-388. Hulbert R.C., Jr. (1988). Cormohipparion and Hipparion Codron D.J.,Sponheimer M., Lee-Thorp J.A., Robinson T., Grant (Mammalis, Perissodactyla, Equidae) from the Late Neogene C.C. & deRuiter D. (2005). Assessing diet in savanna herbivores of . Bulletin of the Florida State Museum, Biolgical using stable carbon isotope rations of faeces. Koedoe, 48: Sciences, 33: 229-338. 115-124. Hulbert R.C., Jr. & MacFadden B.J. (1991). Morphological De Terra H. (1956). New approaches to the problem of man’s origin. transformation and cladogenesis at the base of the adaptive Science, 124: 1282-1285. radiation of Miocene hypsodont horses. American Museum Eisenmann V. (1995). What metapodial morphometry has to Novitates, 3000: 1-61. say about some Miocene hipparions. In Vrba E.S., Denton Hürzeler J. & Engesser B. (1976). Les faunes des mammifères G.H., Partridge T.C. & Burckle L.H. (eds), Paleoclimate and néogènes du Bassin de Baccinello (Grosseto, Italie). Comptes Evolution, with Emphasis on Human Origins, Yale University Rendus de l’Academie de Sciences de Paris, sér. 2D, 283: Press, New Haven: 148-163. 333-336. 202 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

Janis C.M. (1988). An estimation of tooth volume and hypsodonty MacFadden B.J. & Woodburne M.O. (1982). Systematics of the indices in ungulate mammals, and the correlation of these Neogene Siwalik hipparions (Mammalia, Equidae) based factors with dietary prefernce. In Russell D.E., Santoro J.P. & on cranial and dental morphology. Journal of Vertebrate Sigogneau-Russell D. (eds), Teeth Revisited: Proceedings of the Paleontology, 2: 185-218. 7th International Symposium Dental Morphology, Paris 1986, Marino B. & McElroy M. (1991). Isotopic composition of

367-387, Mémoires du Museum national d’Histoire Naturelle, atmospheric CO2 inferred from carbon in plant cellulose. Nature, (Sér. C), Paris. 349: 127-131. Janis C.M. & Ehrhard D. (1988). Correlation of relative muzzle Martini P.I. & Sagri M. (1993). Tectono-sedimentary characteristic width and relative incisor width with dietary preference in of LateMiocene–Quaternary extensional basins of the Northern ungulates. Zoological Journal of the Linnean Society, 92: Appenines. Italy. Earth Science Review, 34: 197-233. 267-284. Meyer H.v. (1833). Fossile Säugethiere. Nova Acta Caesaris Kaiser T.M. (2003). The dietary regimes of two contemporaneous Leopoldi, 16 (2): 1-423. populations of Hippotherium primigenium (Perissodactyla, Nelson S. (2005). Paleoseasonality inferred from equid teeth Equidae) from the Vallesian (upper Miocene) of Southern and intra-tooth isotopic variability. Palaeogeography, Germany. Palaeogeography, Palaeoclimatology, Palaeoecology, Palaeoclimatology, Palaeoecology, 222: 122-144. 198: 381-402. Nickel R., Schummer A., Seiferle E., Wilkens H., Wille K.-H. & Kaiser T.M. & Bernor R.L. (2007). The Baltavar Hippotherium: Frewein J. (1986). Th e Anatomy of the Domestic Animals, A mixed feeding Upper Miocene hipparion (Equidae, Volume I: The Locomotor System. Verlag Paul Parey-Springer Perissodactyla) from Hungary (East-Central Europe); In Nagel Verlag, Berlin. D. & van den Hoek Ostende L.W. (eds), Abhandlungen der Nowak R.M. (1999). Walker’s Mammals of the World, 6th ed. John Bayerischen Akademie der Wissenschaften, 30: 241-267. Hopkins University Press, Baltimore, London. Kaiser T. M., Bernor R.L., Scott R.S., Franzen J.L. & Solounias Owen R. (1848). Description of teeth and portions of jaws of two N. (2003). New Interpretations of the Systematics and extinct anthracotherioid quadrupeds discovered in the Eocene Palaeoecology of the Dorn-Dürkheim 1 Hipparions (Late deposits on the N. W. coast of the Isle of Wight. Quarterly Miocene, Turolian Age [MN 11]) Rheinhessen, Germany. Journal of the Geological Society of London, 4: 103-141. Senckenbergiana lethaea, 83: 103-133. Qiu Z., Weilong H. & Zhihui G. (1988). Chinese hipparionines from Kaiser T.M. & Fortelius M. (2003). Differential mesowear in the Yushe Basin. Palaeontologica Sinica, Series C, 175: 1-250. occluding upper and lower molars - opening mesowear analysis Quade J., Cerling T., Andrews P. & Alpagut B. (1995). Paleodietary for lower molars and premolars in hypsodont equids. Journal reconstruction of Miocene faunas from Pasalar, Turkey using of Morphology, 258: 67-83. stable carbon and oxygen isotopes of fossil tooth enamel. Kaiser T.M. & Solounias N. (2003). Extending the tooth mesowear Journal of Human Evolution, 28: 373-384. method to extinct and extant equids. Geodiversitas, 25: 321-345. Rook L. (1999). Late Turolian Mesopithecus (Mammalia, Primates, Kaiser T.M., Solounias N., Fortelius M., Bernor R.L. & Schrenk F. Colobinae) from Italy. Journal of Human Evolution, 36: 535- (2000). Tooth mesowear analysis on Hippotherium primigenium 547. from the Vallesian Dinotheriensande (Germany) – A blind test Rook L. (2009). The Italian fossil primate record: an update study. Carolinea, 58: 103-114. and perspective for future research. Bollettino della Società Kingdon J. (1982). East African Mammals (Bovids). Vol. III Part Paleontologica Italiana, 48: 67-77. C. Academic Press, London. Rook L., Abbazzi L., Chesi F., Delfino M., Ferretti M.P. & Gallai Kostopolous D. & Bernor R.L. (in press). The Maragheh G. (2008). The Italian record of latest Miocene continental bovids (Mammalia, Artiodactyla): systematic revision vertebrates. Bollettino della Società Paleontologica Italiana, and biostratigraphic-biozoogeographic interpretation. 47: 191-194. Geodiversitas. Rook L., Ficcarelli G. & Torre D. (1991). Messinian carnivores from Koufos G.D. (1987a). Study of Pikermi hipparions Part I: Italy. Bollettino della Società Paleontologica Italiana, 30: 7-22. Generalities and . Bulletin du Museum Nationale Rook L., Gallai G. & Torre D. (2006). Lands and endemic d’Histoire Naturelle, Paris, ser. 4, 9, sec. C, 2: 197-252. mammals in the Late Miocene of Italy: Constraints for Koufos G.D. (1987b). Study of Pikermi hipparions Part II: paleogeographic outlines of Tyrrhenian area. Palaeogeography Comparisons and odontograms. Bulletin du Museum Nationale Palaeoclimatology Palaeoecology, 238: 263-269. d’Histoire Naturelle, Paris, ser. 4, 9, sec. C, 2: 327-363. Rook L., Harrison T. & Engesser B. (1996), The taxonomic status Koufos G.D. & Vlachou T. (2005). Equidae (Mammalia, and biochronological implications of new finds ofOreopithecus Perissodactyla) from the late Miocene of Akkasdagi, Turkey. from Baccinello. Journal of Human Evolution, 30: 3-27. Geodiversitas, 27 (4): 633-705. Rook L. & Martínez-Navarro B. (2004). Viverra howelli n. sp., a new Ligios S., Benvenuti M., Gliozzi E., Papini M. & Rook L. viverrid (Carnivora, Mammalia) from the Baccinello-Cinigiano (2008). Late Miocene palaeoenvironmental evolution of the basin (latest Miocene, Italy). Rivista Italiana di Paleontologia Baccinello–Cinigiano Basin (Tuscany, central Italy) and new e Stratigrafia, 110: 719-723. autoecological data on rare fossil fresh- to brackish-water Rook L., Oms O., Benvenuti M. & Papini M. (2011). ostracods. Palaeogeography Palaeoclimatology Palaeoecology, Magnetostratigraphy of the Late Miocene Baccinello-Cinigiano 264: 277-287. basin (Tuscany, Italy) and the age of Oreopithecus bambolii Lorenz H.G. (1968). Stratigraphisches und mikropaläontologisches faunal assemblages. Palaeogeography Palaeoclimatology Untersuchungen des Braunkohlengebietes von Baccinello Palaeoecology, 305: 286-294. (Grosseto, Italien). Rivista Italiana di Paleontologia e Rook L., Renne P., Benvenuti M. & Papini M. (2000). Geochronology Stratigrafia, 74: 147-270. of Oreopithecus bearing succession at Baccinello (Italy) and the Lourens L., Hilgen F., Shackleton N.J., Laskar J. &Wilson J. (2004). pattern of European Miocene hominoids. Journal of Orbital tuning calibrations and conversions for the Neogene Human Evolution, 39: 577-582. Period. In Gradstein F., Ogg J. & Smith A. (eds), A Geologic Rook L. & Rustioni M. (1991). Tapirus cf. arvernensis from the Late Time Scale. Cambridge University Press, Cambridge: 469-471. Turolian Baccinello V3 Faunal Assemblage (Grosseto, Italy). MacFadden B. (1980). The Miocene horse Hipparion from North Bollettino della Società Paleontologica Italiana, 30: 325-327. America and from the type locality in Soputhern France. Rook L. & Torre D. (1995). Celadensia grossetana nov. sp., Palaeontology 23(3): 617-635. (Cricetidae, Rodentia) from the late Turolian Baccinello- MacFadden B. (1984). Systematics and phylogeny of Hipparion, Cinigiano basin (Southern Tuscany, Italy). Géobios, 28: Neohipparion, Nanhippus, and Cormohipparion (Mammalia, 379-382. Equidae) from the Miocene and Pliocene of the New World. Scott R.S. (2004). The Comparative Paleoecology of Late Miocene Bulletin of the American Museum of Natural History, 179: Eurasian Hominoids.Unpublished Ph.D. Dissertation thesis, The 1-196. University of at Austin. R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 203

Scott R.S., Armour-Chelu M. & Bernor R.L. (2005b). Evidence for Webb S.D. & Hulbert R.C. Jr. (1986). Systematics and evolution of Two Hipparion Species at Rudabánya II. In Bernor R.L., Kordos Pseudohipparion (Mammalia, Equidae) from the late Neogene L. & Rook L. (eds), Multidisciplinary Research at Rudabánya. of the Gulf Coastal Plain and the Great Plains. In Flanagan Palaeontographia Italica, 90: 213-216. K.M. & Lillegraven J.A. (eds), Vertebrates, Phylogeny, and Scott R.S., Bernor R.L. & Raba W. (2005a). Hipparionine Horses Philosophy. University of Wyoming Contributions in Geology, of the Greater Pannonian Basin: Morphometric Evidence spec. paper, 3: 237-272. from the Postcranial Skeleton. In Bernor R.L., Kordos L. & Wolf D., Kaiser T.M., Nelson S.M., Nelson S.V., Semprebon Rook L. (eds), Multidisciplinary Research at Rudabánya. G., Schwartz H. & Bernor R.L. (2010). The systematics and Palaeontographia Italica, 90: 195-210. paleodiet of a diverse equid assemblage from Makuyuni, Scott R.S., Fortelius M., Huttunen K. & Armour-Chelu M.J. (2003). Northern Tanzania. Palaeodiversity, 3: 249-269. Abundance of Hipparion. In Fortelius M., Kappelman J., Sen Wood H.E. (1937). Perissodactyl suborders. Journal of Mammalogy, S. & Bernor R.L. (eds), The Geology and Paleontology of the 18(1): 1-106 Miocene Sinap Formation, Turkey. Columbia University Press, Woodburne M.O. (1989). Hipparion horses: a pattern of endemic New York: 380-397. evolution and intercontinental dispersal. In Prothero D.R. & Skinner J. D. & Smithers R.H.N. (1990). The mammals of the Schoch R.M. (eds), The Evolution of the Perissodactyls. Oxford southern African subregion. Pretoria. University of Pretoria. Monographs on Geology and Geophysics, 15: 197-233. Sponheimer M., Lee-Thorp J.A., DeRuither J.D.J., Smith J.M., Woodburne M.O. (2007). Phyletic diversification of the Van Der Merwe N.J., Reed K., Grant C.C., Ayliffe L.K., Cormohipparion occidentale Complex (Mammalia, Robinson T.F., Heidelberger C. & Marcus W. (2003). Diets of Perissodactyla, Equidae), Late Miocene, North America, and Southern African bovidae: Stable isotope evidence. Journal of the origin of the Old World Hippotherium Datum. Bulletin of Mammalogy, 84: 471-479. the American Museum of Natural History, 306: 1-138. Steinmann G. & Döderlein L. (1890). Elemente der Paläontologie. Woodburne M.O. & Bernor L.R. (1980). On superspecific groups of Wilhelm Engelmann, Leipzig: 1-848. some Old World hipparionine horses. Journal of Paleontology, Stoppani A. (1880). Cenni sulle nuove miniere di lignite in territorio 54: 1319-1348. di Cana (Toscana).Tipografia degli Ingegneri, Milano: 1-8. Woodburne M.O., MacFadden B.J. & Skinner M.F. (1981). The Van Wieren S.E. (1996). Digestive strategies in ruminants and North American “Hipparion” datum, and implications for the nonruminants. Landbouwuniversiteit Wageningen, Den Haag: Neogene Old World. Géobios, 14: 1-32. 1-191. Vlachou T.D. & Koufos G.D. (2009). Equidae. In Koufos G.D. Manuscript received 11 September 2011 & Nagel D. (eds), The Late Miocene Mammal Faunas of Revised manuscript accepted 18 November 2011 the Mytilinii Basin, Samos Island, Greece: New Collection. Published online 12 December 2011 Beiträge zur Paläontologie, 31: 207-281. Editor Raffaele Sardella

Appendix – Baccinello V3 Hippotherium studied sample: measurement tables

Table A1 - Upper Incisors M1 = length at occlusal level; M2 = length 10mm above the tooth’s base; M3 = width at occlusal level; M4 = width 10mm above the tooth’s base; M5 = crown height.

Specimen # Species element side M1 M2 M3 M4 M5 IGF 9400V H. malpassii (Type) I1 R 15.6 - 10.6 - - IGF 9410V H. malpassii I1 L 17.4 10.4 9.6 - 22.0 IGF 9400V H. malpassii (Type) I2 R 14.1 - 8.7 - - IGF 9400V H. malpassii (Type) I3 R 16.5 - 6.9 - -

Table A2 - Upper cheek teeth M1 = length at occlusal level; M2 = length 10mm above the tooth’s base; M3 = width at occlusal level; M4 = width 10mm above the tooth’s base; M5 = crown height measured along the mesostyle; M6 = number of plications on the anterior face of the prefossette; M7 = number of plications on the posterior face of the prefossette; M8 = number of plications on the anterior face of the postfossette; M9 = number of plications on the posterior face of the postfossette; M10 = protocone length; M11 = protocone width.

Specimen # Species element side M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 IGF 9400V H. malpassii (Type) P2 L 32.5 28.8 23.8 23.3 28.7 7 8 9 2 7.6 4.5 IGF 9401V H. malpassii P2 R 33.4 32.3 23.0 22.7 29.3 7 6 9 3 6.4 4.5 IGF 9405V H. malpassii P2 L - - 22.2 22.6 17.1 4 7 4 1 6.9 5.1 IGF 5286V H. malpassii P2 L 35.9 - 22.5 - - 2 3 3 - 6.3 2.9 NHMB JH 43 H. malpassii P2 L 32.5 31.0 21.7 22.7 41.6 ------NHMB JH 126A H. malpassii P2 L 34.6 31.9 22.9 24.2 37.9 - 4 6 3 6.3 5.0 NHMB JH 126E H. malpassii P2 R 34.4 31.9 22.5 23.6 35.9 5 5 6 1 6.9 4.9 NHMB JH 229D H. malpassii P2 R 32.9 32.9 23.1 22.1 38.8 6 10 7 3 5.8 3.7 NHMB JH 140 H. sp. P2 R 33.2 30.7 23.9 24.0 13.2 3 7 4 2 7.2 4.9 NHMB JH 186 H. sp. P2 L 33.6 33.8 - - 23.6 5 5 6 1 - - IGF 9400V H. malpassii (Type) P3 L 24.3 20.4 23.7 24.4 28.3 4 13 10 2 7.1 4.4 IGF 9401V H. malpassii P3 R 25.0 22.2 22.5 23.8 37.9 7 7 5 5 6.1 4.4 IGF 5286V H. malpassii P3 L 26.5 - 22.0 - - 2 4 1 - 6.3 2.9 204 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

NHMB JH 2 H. malpassii P3 R 26.4 22.6 24.1 23.7 39.2 5 5 6 1 4.6 3.9 NHMB JH 124 H. malpassii P3 R 24.9 23.5 24.2 25.8 25.1 6 16 10 - 5.2 4.9 NHMB JH 126B H. malpassii P3 L 27.8 23.6 23.2 24.3 40.9 4 7 2 1 4.6 4.4 NHMB JH 229A H. malpassii P3 R 26.1 24.3 23.7 23.7 45.8 7 9 6 1 7.6 3.2 NHMB JH 229B H. malpassii P3 L 26.8 23.6 23.7 23.9 45.6 6 11 9 4 7.6 3.3 IGF 9383V H. sp. P3 R 22.6 20.5 22.2 22.0 26.0 1 4 1 1 6.4 6.9 IGF 9380V H. sp. P3 R 22.1 20.5 20.2 19.9 26.7 - - - - 5.6 3.3 NHMB JH 188 H. sp. P3 L 25.2 22.5 24.0 25.6 22.7 4 7 6 3 6.7 5.5 IGF 9400V H. malpassii (Type) P4 L 24.1 20.5 23.7 24.4 29.8 3 10 7 1 5.3 4.2 IGF 9406V H. malpassii P4 L 21.7 19.7 22.2 22.9 23.3 7 11 9 4 6.9 5.3 IGF 9408V H. malpassii P4 R 23.5 19.4 22.7 24.5 31.5 6 8 9 5 6.6 4.0 IGF 9413V H. malpassii P4 R 25.3 23.1 25.1 26.2 27.4 10 13 14 1 7.2 5.6 IGF 5286V H. malpassii P4 L 27.8 - 20.9 - - 2 2 3 - 4.5 4.0 NHMB JH 126F H. malpassii P4 R 27.6 23.4 21.9 - 42.4 1 6 4 - 4.2 4.8 NHMB JH 134 H. malpassii P4 L 27.7 23.6 22.1 25.3 44.3 1 6 4 - 4.0 3.6 NHMB JH 229E H. malpassii P4 L 25.1 22.2 21.9 22.7 48.1 7 10 2 1 7.0 3.4 IGF 9382V H. sp. P4 R 20.5 18.3 23.9 24.0 26.9 1 4 4 1 5.6 3.3 IGF 9385V H. sp. P4 L 23.8 21.5 23.7 23.8 25.5 - - - - 6.5 5.0 NHMB JH 189 H. sp. P4 L 26.2 24.3 24.7 25.2 21.0 3 8 8 2 7.0 5.5 IGF 9400V H. malpassii (Type) M1 L 21.1 17.7 21.5 23.5 27.7 6 8 10 5 6.7 4.2 IGF 9401V H. malpassii M1 R 22.8 20.8 20.3 21.3 37.5 5 6 4 3 5.9 4.3 IGF 9406V H. malpassii M1 L 21.6 19.3 22.4 23.1 22.8 6 10 8 3 7.2 4.8 NHMB JH 41 H. malpassii M1 L 23.9 20.5 20.8 22.0 42.9 5 12 6 3 6.3 3.9 NHMB JH 126C H. malpassii M1 L 23.7 20.3 21.8 23.7 38.8 7 9 7 2 5.4 4.1 NHMB JH 126G H. malpassii M1 R 24.1 20.6 22.1 24.1 39.1 7 7 7 2 5.4 4.6 NHMB JH 190 H. malpassii M1 L 21.7 20.0 - - 20.8 3 7 7 4 - - NHMB JH 229C H. malpassii M1 L 22.9 19.7 21.5 22.5 41.9 7 14 8 3 6.9 3.7 IGF 9382V H. sp. M1 R 19.9 18.9 18.4 19.1 28.7 1 3 4 - 5.7 3.6 NHMB JH 197 H. sp. M1 L 22.4 21.1 - - 32.6 6 9 8 3 - - NHMB JH 212 H. sp. M1 L - - - - 32.9 7 8 6 - 6.2 4.2 IGF 9400V H. malpassii (Type) M2 L 20.7 18.1 20.8 21.6 30.8 6 10 8 4 6.3 4.2 IGF 9407V H. malpassii M2 L 23.2 23.1 19.8 20.9 23.8 5 10 7 5 7.0 4.1 NHMB JH 42 H. malpassii M2 L 23.6 19.6 18.5 20.8 45.1 5 7 3 - 5.0 3.5 NHMB JH 126D H. malpassii M2 L 24.3 21.3 21.0 21.5 41.1 5 5 6 1 5.4 4.1 NHMB JH 166 H. sp. M2 L 22.0 20.7 20.5 20.7 22.5 5 8 8 7 5.7 4.9 NHMB JH 214 H. sp. M2 L 22.4 20.6 21.5 21.7 37.2 4 6 4 1 6.0 4.4 IGF 9400V H. malpassii (Type) M3 L 22.1 20.5 17.2 19.1 31.9 3 2 5 2 6.4 4.2 IGF 9401V H. malpassii M3 R 23.0 23.5 18.7 19.7 38.7 5 3 2 3 7.0 3.7 IGF 9409V H. malpassii M3 R 22.2 23.4 20.3 22.0 32.9 6 9 5 3 7.4 3.7 NHMB JH 127 H. malpassii M3 L 20.7 21.8 13.9 17.8 36.2 - 3 - - 5.2 2.7 NHMB JH 142A H. malpassii M3 R 19.8 21.2 15.5 18.6 46.3 - 4 3 - 6.7 2.8 NHMB JH 142B H. malpassii M3 L 20.7 21.0 14.4 17.5 46.7 - 7 3 - 6.2 2.3 NHMB JH 202 H. sp. M3 R 23.1 22.8 - - 26.3 4 6 5 4 8.3 3.5 NHMB JH 187 H. malpassii M1-2 L 22.1 20.7 21.7 22.2 24.5 4 6 5 2 6.5 4.3 IGF 9373V H. sp. M1-2 - - - - 24.6 - - 6 1 - - IGF 9374V H. sp. M1-2 - - - - 10.8 - 8 8 1 7.8 76.0

Table A3 - Lower cheek teeth M1 = length at occlusal level; M2 = length 10mm above the tooth’s base; M3 = length of metaconid-metastylid; M4 = length of the prefossette; M5 = length of the postfossette; M6 = width across plane of ectoflexid/linguaflexid ; M7 = width 10mm above the tooth’s base; M8 = width across plane of metaconid and enamel band labial to protoconid; M9 = width across plane of metstylid and enamel band labial to hypoconid; M10 = crown height as measured from base to occlusal level on mesial face of the tooth.

Specimen # Species element side M1 M2 M3 M4 M5 M6 M7 M8 M9 M10

IGF 5286V H. malpassii P2 L 33.3 - 11.6 10.8 12.7 13.0 - 9.1 11.2 -

NHMB JH 231 H. malpassii P2 L 29.3 27.7 10.6 7.5 12.0 11.0 13.8 10.2 11.4 28.7

NHMB JH 160A H. aff. malpassii P2 L 32.7 30.9 13.9 9.8 12.1 - - - - 32.7

NHMB JH 160G H. aff. malpassii P2 R 31.1 30.9 12.2 9.3 12.6 - - - - 32.7

NHMB JH 125C H. sp. P2 L 29.5 30.2 11.1 8.5 11.8 13.0 14.8 11.3 13.1 27.4

IGF 5286V H. malpassii P3 L 27.4 - 13.0 8.7 10.7 13.8 - 11.5 10.4 -

IGF 9398V H. malpassii P3 L 23.9 22.5 14.2 7.7 11.6 - 14.8 15.0 - -

NHMB JH 132 H. malpassii P3 R 25.8 22.8 13.9 9.7 11.7 14.4 14.4 13.2 14.1 38.2 R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 205

NHMB JH 133 H. malpassii P3 L 28.0 24.0 13.6 9.8 12.9 15.5 15.3 12.2 14.1 38.8

NHMB JH 160C H. aff. malpassii P3 L 28.0 24.4 15.5 9.1 11.7 14.2 14.8 12.6 13.6 43.0

NHMB JH 160I H. aff. malpassii P3 R 27.2 24.8 15.3 9.2 12.0 14.4 15.2 13.1 13.7 46.6

NHMB nonumb3 H. sp. P3 L 25.9 25.7 14.6 8.3 11.5 15.0 15.6 15.7 14.4 17.6

NHMB JH 130 H. sp. P3 R 25.8 23.4 13.0 8.5 10.8 13.9 15.3 12.2 10.3 40.8

IGF 5286V H. malpassii P4 L 28.2 - 11.1 7.5 10.7 13.6 - 11.7 9.6 -

IGF 9398V H. malpassii P4 L 23.8 - 14.2 7.4 11.5 - 12.1 14.2 - -

IGF 9411V H. malpassii P4 R 26.8 24.7 15.4 7.7 12.5 15.9 15.2 15.3 15.4 28.2

IGF 9414V H. malpassii P4 L 25.5 23.1 13.7 7.8 12.2 13.3 14.1 13.7 13.3 32.1

NHMB JH 143 H. malpassii P4 L 26.1 23.8 13.8 9.5 11.9 14.7 14.8 12.7 13.0 39.3

NHMB nonumb3 H. malpassii P4 L 24.9 23.3 13.9 7.7 11.3 14.5 15.2 14.3 13.3 22.6

NHMB JH 160B H. aff. malpassii P4 L 26.9 24.2 13.1 9.3 13.6 15.1 15.6 11.4 11.7 49.6

NHMB JH 160H H. aff. malpassii P4 R 26.3 23.5 14.3 8.0 12.5 14.4 16.1 11.4 12.2 49.6

IGF 9384V H. sp. P4 R 24.3 23.4 15.1 7.3 12.1 12.7 15.8 15.2 13.9 23.0

NHMB JH 161 H. sp. P4 R 23.8 24.1 14.5 7.9 12.1 14.6 17.1 14.3 13.8 10.7

NHMB JH 163 H. sp. P4 L 24.2 22.3 9.4 6.5 11.3 14.1 13.3 9.2 10.3 49.8

NHMB JH 165 H. sp. P4 R 24.6 23.0 13.5 7.1 - 11.6 12.6 12.1 13.7 28.6

NHMB JH 167 H. sp. P4 L 25.3 24.5 14.7 9.3 10.5 17.5 18.5 13.5 12.8 14.4

NHMB JH 196 H. sp. P4 L 25.6 24.3 14.2 8.3 11.4 10.9 16.6 13.8 12.0 36.0

NHMB JH 206 H. sp. (small size) P4 L 26.8 23.6 11.5 7.4 11.3 13.3 14.4 10.6 10.4 43.7

IGF 9412V H. malpassii M1 R 24.2 21.1 12.9 6.7 12.2 12.8 12.9 11.7 11.0 33.6

IGF 9398V H. malpassii M1 L 20.7 - 12.3 6.0 8.0 - 11.9 - - -

NHMB nonumb3 H. malpassii M1 L 22.1 20.0 12.9 7.0 7.8 12.7 13.9 12.4 11.5 20.7

NHMB JH 160D H. aff. malpassii M1 L 25.2 22.5 14.2 8.2 8.7 13.1 14.9 10.8 10.8 44.2

NHMB JH 160L H. aff. malpassii M1 R 25.8 22.6 13.9 8.1 8.9 12.3 12.6 15.5 10.6 42.1

IGF 9415V H. sp. M1 R 24.0 21.2 14.3 7.0 11.2 14.9 15.0 15.7 13.5 17.7

NHMB JH 122A H. sp. M1 L 22.5 - 12.1 6.1 8.4 14.4 - 12.1 11.0 -

NHMB JH 125C H. sp. M1 L 23.5 23.0 13.7 8.4 9.8 13.0 15.1 11.4 11.2 31.7

NHMB JH 162 H. sp. M1 R 21.5 20.3 12.1 6.3 9.1 13.1 13.4 10.1 9.3 22.3

NHMB JH 176A H. sp. M1 L 21.6 20.7 11.5 5.9 7.4 12.2 13.2 11.8 10.2 19.8

NHMB JH 235 H. sp. M1 L 22.9 19.7 12.8 6.6 8.2 16.0 14.1 13.0 11.7 29.6

IGF 9398V H. malpassii M2 L 20.7 - 12.0 6.8 6.6 - 9.3 10.0 - -

NHMB JH 160E H. aff. malpassii M2 L 24.3 21.9 12.8 8.4 9.0 8.8 12.8 9.6 8.8 45.7

NHMB JH 160M H. aff. malpassii M2 R 25.1 21.8 12.9 7.4 8.7 8.2 13.2 9.5 9.6 44.6

NHMB JH 122A H. sp. M2 L 21.1 - 12.0 5.6 8.5 10.8 - 11.1 10.7 -

NHMB JH 122B H. sp. M2 R 21.0 - 12.0 4.7 6.4 11.3 - 11.2 9.8 -

NHMB JH 125C H. sp. M2 L 22.9 22.4 11.3 7.1 7.8 13.2 12.9 11.9 10.2 38.7

NHMB JH 137 H. sp. M2 L 24.4 19.7 13.4 8.4 10.6 12.8 13.3 10.7 10.5 36.6

NHMB JH 176A H. sp. M2 L 21.9 20.2 12.4 6.3 9.3 14.9 13.1 12.1 11.7 25.6

NHMB JH 207 H. sp. M2 R - - 11.8 6.1 7.5 - - 11.1 - 27.3

NHMB JH 216 H. sp. (small sized) M2 L 21.3 20.9 11.4 6.7 9.3 11.0 12.5 9.7 9.0 32.4

IGF 9398V H. malpassii M3 L 26.0 30.7 11.6 7.2 9.2 - 9.6 9.7 - -

NHMB JH 160F H. aff. malpassii M3 L 25.5 28.8 10.6 7.8 9.9 5.5 11.8 8.0 6.9 45.1

NHMB JH 122A H. sp. M3 L 24.0 - 10.2 7.1 7.7 - - 10.5 7.7 -

NHMB JH 122B H. sp. M3 R 23.5 - 10.8 5.7 7.6 10.8 - 10.4 8.7 -

NHMB JH 125C H. sp. M3 R 22.7 28.4 9.1 6.1 - 9.3 12.2 6.8 6.2 43.1

NHMB JH 135 H. sp. M3 L 20.3 25.5 9.2 5.9 6.9 9.1 9.6 6.8 6.5 41.9

NHMB JH 136 H. sp. M3 L 25.0 - 10.6 5.5 7.8 10.1 10.7 8.6 8.1 33.0

NHMB JH 138 H. sp. M3 L 25.4 28.0 11.1 7.2 8.5 10.3 11.5 9.8 8.9 38.2

NHMB JH 176A H. sp. M3 L 23.1 26.2 10.2 4.5 7.3 11.5 11.6 9.1 8.8 32.9

NHMB JH 191 H. sp. M3 L 26.2 28.7 11.7 7.0 8.2 11.9 11.8 10.6 9.9 26.6

NHMB JH 215 H. sp. M3 R 25.5 26.3 9.9 6.7 9.1 7.4 11.7 8.5 6.6 43.0

Table A4 - Mandible M1 = maximum jaw length; M2 = muzzle length; M3 = premolar tooth row length; M4 = molar tooth row length; M5 = cheek tooth row length; M6 = length of ascending ramus; M7 = muzzle breadth; M8 = eight of mandible at condyle; M9 = eight of ascending ramus; M10

= eight of jaw posterior to M3; M11 = eight of jaw between P4 and M1; M12 = eight of jaw anterior to P2; M13 = length of the symphysis; M14 = minimum breadth of the symphysis.

Specimen # Species side M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12 M13 M14 IGF 5286V H. malpassii L - - 89.6 ------58.7 39.7 - - 206 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

Table A5 – Forelimb elements

Humerus M1 = maximal length; M2 = maximal length from caput; M3 = minimal breadth; M4 = diameter perpendicular to, and at level of M3; M5 = proximal maximal breadth; M6 = proximal depth at the level of the median tubercule; M7 = maximal breadth at the trochlea; M8 = distal maximal depth; M9 = maximal trochlear height (medial); M10 = minimal trochlear height (in the middle); M11 = trochlear height at the sagittal crest (near the condyle)

Specimen # Species side M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 NHMB JH 153 H. malpassii R ------51.0 49.9 34.9 25.0 32.1

Radius M1 = maximal length; M2 = medial length; M3 = minimal breadth; M4 = depth of diaphysis at level of M3; M5 = proximal articular breadth; M6 = proximal articular depth; M7 = proximal maximal breadth; M8 = distal articular breadth; M9 = distal articular depth; M10 = distal maximal breadth; M11 = breadth of the radial condyle; M12 = breadth of the ulnar condyle

Specimen # Species side M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12 NHMB JH 152 H. malpassii R - - - - 55.2 29.6 ------Scaphoideum M1 = greatest craniocaudal diameter; M2 = greatest proximodistal diameter; M3 = greatest mediolateral diameter.

Specimen # Species side M1 M2 M3 IGF 9367V H. sp. L 37.3 27.1 23.4 NHMB JH 173 H. sp. R 37.6 27.1 21.0 NHMB JH 199 H. sp. (small sized) L 34.5 24.5 19.9

Lunatum M1 = greatest craniocaudal diameter; M2 = greatest proximodistal diameter; M3 = proximodistal diameter on lateral aspect; M4= greatest mediolateral diameter.

Specimen # Species side M1 M2 M3 M4 IGF 9363V H. malpassii L 29.9 28.1 24.3 29.6

Pyramidale M1 = maximal proximodistal diameter; M2 = maximum craniocaudal diameter; M3 = maximum mediolateral diameter; M4 = craniocaudal diameter at distal heel.

Specimen # Species side M1 M2 M3 M4 IGF 9362V H. malpassii L 33.5 15.9 16.6 13.7

Magnum M1 = proximodistal diameter; M2 = mediolateral diameter; M3 = craniocaudal diameter.

Specimen # Species side M1 M2 M3 NHMB JH 205 H. sp. R 18.4 36.6 27.8

Metacarpale III M1 = maximal length; M2 = medial length; M3 = minimal breadth; M4 = depth of diaphysis at level of M3; M5 = proximal articular breadth; M6 = proximal articular depth; M7 = maximal diameter of the articular facet of the third carpal; M8 = diameter of the articular facet for the fourth carpal; M9 = diameter of the articular facet for the second carpal; M10 = distal maximal supra-articular breadth; M11 = distal maximal articular width; M12 = distal maximal depth of the keel; M13 = Distal minimal depth of the lateral condyle; M14 = distal maximal depth of the medial condyle.

Specimen # Species side M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12 M13 M14 IGF 8192V H. malpassii L 214.3 206.2 27.2 21.5 39.3 25.1 30.6 14.4 6.1 37.5 35.0 27.9 24.2 25.6 IGF 9397V H. malpassii L 216.2 209.2 27.8 22.5 42.1 29.9 36.6 12.9 7.4 40.0 38.2 30.3 25.7 26.6 NHMBJH 168 H. malpassii L ------39.0 36.8 30.9 23.8 25.9 NHMBJH 144 H. sp. - - 29.9 23.4 ------NHMBJH 169 H. sp. R ------37.2 36.6 28.5 23.0 28.4 NHMBJH 194 H. sp. (small size) L ------31.1 32.5 26.3 22.2 23.6 R.L. Bernor et alii - Hippotherium malpassii n.sp. from Baccinello V3 207

Table A6 – Hindlimb elements

Tibia M1 = maximal length; M2 = medial length; M3 = minimal breadth; M4 = depth of diaphysis at level of M3; M5 = proximal maximal breadth; M6 = proximal maximal depth; M7 = distal maximal breadth; M8 = distal maximal depth.

Specimen number Species side M1 M2 M3 M4 M5 M6 M7 M8 IGF 9369V H. sp. L ------69.4 -

Astragalus M1 = maximal length; M2 = maximal diameter of the medial condyle; M3 = breadth of the trochlea (at the apex of each condyle); M4 = maximal breadth; M5 = distal articular width; M6 = distal articular depth; M7 = maximal medial depth.

Specimen number Species side M1 M2 M3 M4 M5 M6 M7 IGF 2026V H. malpassii R 56.7 55.1 25.8 52.5 42.1 31.9 42.4 IGF 8195V H. malpassii L 56.9 55.1 24.6 53.0 42.1 32.3 41.9 IGF 9395V H. malpassii 54.7 55.7 25.8 55.8 42.2 31.4 42.1 NHMB JH 149 H. malpassii R 55.0 52.4 25.0 56.5 42.6 30.8 - IGF 9396V H. sp. 50.8 53.4 21.6 48.4 - - - NHMB JH 16 H. sp. L 56.8 57.6 25.8 57.9 44.3 32.5 - NHMB JH 128 H. sp. L 57.5 52.3 30.0 42.0 38.0 30.6 - NHMB JH 157 H. sp. R - - - - 41.0 31.0 - NHMB JH 225 H. sp. L - - - - 43.6 33.8 -

Calcaneum M1 = maximal length; M2 = length of the proximal part; M3 = minimal breadth; M4 = proximal maximal breadth; M5 = proximal maximal depth; M6 = distal maximal width; M7 = distal maximal depth.

Specimen number Species side M1 M2 M3 M4 M5 M6 M7 NHMB JH 150 H. malpassii L 98.0 60.9 19.1 30.6 46.1 46.8 44.8 NHMB JH 198 H. malpassii R 105.8 66.2 19.7 27.3 28.8 44.7 46.8

Naviculare M1 = proximodistal diameter; M2 = mediolateral diameter; M3 = craniocaudal diameter.

Specimen number Species side M1 M2 M3 NHMB JH 226 H. malpassii R 13.2 39.1 33.1

Cuboid M1 = maximum craniocaudal diameter; M2 = maximum proximodistal diameter; M3 = maximum mediolateral diameter.

Specimen number Species side M1 M2 M3 NHMB JH 155 H. sp. R 26.9 36.8 25.7 NHMB JH 156 H. sp. R 28.2 35.9 24.3 NHMB JH 174 H. sp. L 27.1 37.2 24.0

Metatarsale III M1 = maximal length; M2 = medial length; M3 = minimal breadth; M4 = depth of diaphysis at level of M3; M5 = proximal articular breadth; M6 = proximal articular depth; M7 = maximal diameter of the articular facet of the third tarsal; M8 = diameter of the articular facet for the fourth tarsal; M9 = diameter of the articular facet for the second tarsal; M10 = distal maximal supra-articular breadth; M11 = distal maximal articular breadth; M12 = distal maximal depth of the keel; M13 = Distal minimal depth of the lateral condyle; M14 = distal maximal depth of the medial condyle.

Specimen number Species side M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12 M13 M14 IGF 8193V H. malpassii L 240.0 236.7 26.0 26.9 40.4 34.3 37.8 11.6 7.5 38.1 33.0 29.7 25.6 27.6 208 Bollettino della Società Paleontologica Italiana, 50 (3), 2011

Table A7 – Phalanges

1st Phalanx central digit (III) M1 = maximal length; M2 = anterior length; M3 = minimal breadth; M4 = proximal articular width; M5 = proximal depth; M6 = distal breadth at the tuberosities; M7 = distal articular breadth; M8 = distal articular depth; M9 = minimal length of the trigonum phalangis; M10 = medial supratuberosital length; M11 = lateral supratuberosital length; M12 = medial infratuberosital length; M13 = lateral infratuberosital length.

Specimen # Species side M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12 M13 IGF 9390V H. malpassii L 67.0 63.2 28.3 41.3 31.9 34.5 34.9 20.7 29.6 43.7 42.5 19.7 20.8 IGF 9391V H. malpassii L 60.4 55.6 30.0 - - 32.1 34.2 20.7 23.3 43.0 38.7 14.0 17.3 NHMB JH 129 H. malpassii L 58.4 52.2 28.7 40.3 30.7 31.7 32.7 18.9 14.4 37.4 34.2 14.4 18.5 NHMB JH 158 H. malpassii L 62.7 56.2 27.9 39.7 29.7 32.6 33.2 19.2 21.2 38.4 40.1 17.4 15.9 NHMB nonumb1 H. malpassii L 66.7 59.5 29.1 40.8 33.4 35.2 34.4 22.2 14.5 38.8 41.4 19.7 18.0 IGF 9368V H. sp. L 56.7 55.1 25.8 52.5 42.1 31.9 42.4 ------NHMB JH 170 H. sp. 58.0 ------IGF 8194V H. sp. (small sized) L 57.0 52.8 27.1 36.5 29.8 31.1 30.4 18.8 19.4 34.6 37.2 16.2 14.5 IGF 9392V H. sp. (small sized) L 56.8 52.4 25.7 36.9 25.5 27.6 29.1 18.0 23.2 36.4 37.1 15.1 13.3

1st Phalanx lateral digits (II and IV) M1 = maximal length; M2 = proximal maximal depth; M3 = proximal maximal breadth; M4 = minimal breadth of the diaphysis; distal maximal breadth; M5 = distal maximal depth; M6 = distal maximal breadth.

Specimen # Species element side M1 M2 M3 M4 M5 M6 NHMB JH 151 H. malpassii 1ph2 35.9 19.7 12.3 10.1 14.8 11.7 IGF 9366V H. malpassii 1ph4 R 35.0 22.9 11.9 9.1 15.9 13.1

2nd Phalanx central digit (III) M1 = maximal length; M2 = anterior length; M3 = minimal breadth; M4 = proximal maximal breadth; M5 = proximal maximal depth; M6 = distal articular maximal depth.

Specimen # Species side M1 M2 M3 M4 M5 M6 IGF 9387V H. malpassii 40.1 33.0 32.8 38.3 27.3 37.1 IGF 9391V H. malpassii L 38.7 28.6 33.0 37.9 27.8 33.6 NHMB JH 232 H. malpassii L 42.7 32.8 32.8 40.2 27.9 39.0 IGF 9379V H. sp. 45.9 36.3 35.9 45.8 30.1 39.2 IGF 9375V H. sp. 41.9 31.3 31.9 - 27.4 36.4 NHMB JH 146 H. sp. R 38.2 - 33.8 38.9 - 36.3 NHMB JH 147 H. sp. L 40.7 33.5 33.5 39.6 27.7 36.2 NHMB JH 148 H. sp. R 38.8 31.4 34.4 40.6 27.3 35.2 NHMB JH 218 H. sp. R 43.1 33.5 31.6 38.2 28.2 33.7 NHMB JH 227 H. sp. R 38.9 30.4 30.9 36.8 25.3 33.4 NHMB JH 233 H. sp. R 45.4 34.8 33.2 - 28.5 37.5 NHMB JH 234 H. sp. R - - - 39.4 27.4 - NHMB nonumb2 H. sp. L 44.6 33.5 29.7 - - - IGF 8194V H. sp. (small sized) L 42.1 31.4 31.8 - 27.4 35.6 IGF 9393V H. sp. (small sized) 40.1 30.4 29.8 36.3 26.5 32.0