Geobios 61 (2020) 61–81 Available online at ScienceDirect www.sciencedirect.com Original article Late Miocene Erinaceinae from the Teruel Basin (Spain) Jan A. van Dam a,b,*, Pierre Mein c, Luis Alcalá d a Faculty of Geosciences, Utrecht University, Vening Meinesz Building A, Princetonlaan 8a, 3584 CB Utrecht, The Netherlands b Institut Català de Paleontologia Miquel Crusafont, c/ Columnes s/n Campus de la UAB, 08193, Cerdanyola del Vallès, Barcelona, Spain c A.R.P.A., Université Lyon 1, Bâtiment Géode, 69622, Villeurbanne Cedex, France d Fundación Conjunto Paleontológico de Teruel-Dinópolis, Av. de Sagunto, 44002 Teruel, Spain A R T I C L E I N F O A B S T R A C T Article history: Classifying fossil teeth of Erinaceinae (spiny hedgehogs) is a challenging task, because of their scanty Received 26 November 2019 record and systematic treatment that heavily relies on skull characteristics. In this paper we describe the Accepted 8 June 2020 complete set of isolated dental elements of Erinaceinae from the upper Miocene sediments of the Teruel Available online 7 July 2020 Basin (eastern Central Spain). Four different species were recognized: Postpalerinaceus cf. vireti, Atelerix Corresponding editor: Gilles Escarguel aff. depereti, Atelerix steensmai nov. sp., and a form classified as Erinaceinae genus and species indet. All four are relatively derived in showing multi-purpose dentitions, not showing only adaptations to Keywords: insectivory, but also to carnivory, herbivory and possibly durophagy/malacophagy. The temporal Mammalia occurrence of spiny hedgehogs during the middle to late Miocene in the Teruel Basin and neighboring Erinaceidae Calatayud-Montalbán Basin peaks within periods of relative aridity, a correlation consistent with Dentition modern geographic distribution. Messinian cooling is the best candidate for explaining a remarkable Function morphology Spain demise of Erinaceinae at 7 Ma. C 2020 The Authors. Published by Elsevier Masson SAS. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). 1. Introduction Amphechinus species were small-sized, contained relatively rounded skulls with narrow snouts and shallow, elongated Due to their fragmented fossil record, the phylogenetic history mandibles. Their dentition is typically insectivorous, characterized of European spiny hedgehogs (Erinaceinae) remains poorly known by relatively small teeth that contain many small sharp cusps and (Ziegler, 1999; Engesser, 2009). Although well-preserved cranial ridges, and with rectangularly shaped posterior upper teeth (P4- material is known from some localities, most of the known M2). In many respects, these early European forms resemble their mammal sites have yielded either no remains or a few isolated older (early Oligocene to early Miocene) Asian relatives (Ziegler dental elements only. This scarcity is probably due to a et al., 2007), which were even more primitive in having a well- combination of their relative rarity within living communities of developed P3, and their third molars not extremely reduced. After a small mammals and their relative absence in the diet of birds of period of apparent absence (including at least the MN3 time prey (Reeve, 1994; but see Corbet, 1988), the main agents in interval), new forms entered the European record from the late accumulating small mammal remains that could finally fossilize early Miocene onwards (MN4). Their cranial and dental morphol- (Andrews and Cook, 1990). ogy reflects a broadening of the dietary spectrum: larger size, The currently accepted view holds that the main evolutionary broader snouts and posterior parts of the skull, deeper mandibles, developments in spiny hedgehogs took place in the Old World and upper molars that are more squared. (Butler, 1948; Ziegler et al., 2007). The oldest European spiny The generic assignment of these more advanced Miocene forms hedgehog fossils are represented by Amphechinus arvernensis is still a matter of debate. Confusion stems from the fact that Blainville, 1839, a form that occurred in Central/Western Europe systematic classification in fossil erinaceines depends for a during the late Oligocene to earliest Miocene. It soon developed significant part on characters of the skull, which is usually not into A. edwardsi, Filhol, 1897, as well as into Dimylechinus bernoullii represented in the fossil material. Apart from Amphechinus, species Huerzeler, 1944, a side branch restricted to France. The early have been generally assigned to three Miocene genera: Mioechinus, Postpalerinaceus, and Atelerix. This number can probably be * Corresponding author. reduced to two, when more diagnostically relevant material E-mail address: j.a.vandam@uu.nl (J.A. van Dam). becomes available (see also Engesser, 2009; Ziegler, 2005). https://doi.org/10.1016/j.geobios.2020.06.002 0016-6995/ C 2020 The Authors. Published by Elsevier Masson SAS. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). 62 J.A. van Dam et al. / Geobios 61 (2020) 61–81 The genus Mioechinus was erected by Butler (1948) to in the same basin, represents a clearly advanced form with regard distinguish skull material of an intermediate- to large-sized form to Amphechinus. Recently, this large-sized form has also been from Öhningen (M. oeningensis, MN7-8) from that of the primitive recognized (with complete hemimandibles) outside its type area genus Amphechinus as well as that from modern Erinaceinae. Butler with a record in the Madrid Basin (Cerro de Batallones, MN10; based himself mainly on the morphology of the palate (presence of Álvarez-Sierra et al., 2017). A possible occurrence in eastern a basisphenoid groove), on the condylar morphology, and on dental Central Spain even during times a young as MN17 was previously features (relative size to M1 of P4, M2 and I1). Another skull hypothesized (Crochet and Heintz, 1971). feature, the shape of the orbital region (with the lacrimal foramen In this paper we summarize the erinaceine remains as and the P4 placed anteriorly), was used together with various recovered from the dense late Miocene micromammal sequence dental features by Engesser (1980) to separate similarly-aged from the Teruel Basin (Fig. 1). The record consists of 123 dental Anatolian material (the small-sized M. tobieni from Yeni-Eskihisar, elements originating from 28 localities. For basin stratigraphy and MN7-8) from Amphechinus. It was provisionally assigned to chronology, we refer to Mein et al. (1990); Opdyke et al. (1997), Mioechinus based on dental characters. Subsequently, the name and Van Dam et al. (2001, 2006, ongoing work). We also discuss Mioechinus was also used for similarly-aged material from China some material from the neighboring Calatayud-Montalbán Basin, (Qiu, 1996). part of which is published (Cañada 6 and 12; López-Guerrero et al., The position of the lacrimal foramen (outside the orbit) was 2011) and part of which has remained unpublished (Pedregueras partly the reason for Mein and Ginsburg (2002) to place also 2C; for geographic and updated stratigraphic context, see Van Dam intermediate- to large-sized material from La Grive (previously et al., 2014). We have assigned the studied material to the genera assigned to ‘Amphechinus’ intermedius) outside Amphechinus, and Postpalerinaceus and Atelerix, although these attributions should be include it into Postpalerinaceus, a genus initially erected as a considered as provisional given the need for a thorough revision of subgenus of ‘Palerinaceus’ (junior synonym of Amphechinus) by European Mio-Pliocene Erinaceinae (in which cranial characte- Crusafont and De Villalta, 1947, but later upgraded to specific rank ristics may or may not play a role). by Butler (1956) in order to accommodate large-sized material We will end our paper with a discussion on the temporal from the early late Miocene (MN10) of Spain. Postpalerinaceus was distribution of Iberian and European spiny hedgehogs. Their defined on a mix of both cranial and dental characters, and was considered (like Mioechinus) to occupy a morphologically inter- mediate position between a late Oligocene-early Miocene evolu- tionary stage (Amphechinus) and the modern stage (Erinaceus, Atelerix). Two other forms from la Grive, one of intermediate size and one of small size and both also with their lacrimal foramen outside the orbit, were placed by Mein and Ginsburg (2002) into the modern genus Atelerix as A. depereti and A. rhodanicus, respectively. Moreover, based on a comparison between palatal morphology in Mioechinus oeningensis and Atelerix, these authors proposed to include the genus Mioechinus into Atelerix. Other workers (e.g., Ziegler, 2005), however, consider palatal shape different enough to retain Mioechinus as a separate genus. The Spanish Miocene mammal record is very rich in sites, of which a significant part has yielded Erinaceinae. However, as most of these sites contain only a few dental elements, in most cases generic assignments cannot be certain. The genus name Amphe- chinus was used by Gibert (1975) to accommodate teeth and mandibular parts from the middle and early late Miocene of the northeastern and eastern-central part of Spain. Whereas the oldest (MN5) material described by Gibert (‘A.’ baudeloti) includes several small-sized, rectangularly shaped upper molars that resemble those considered to represent ‘true’ Amphechinus (e.g., as in Ziegler et al., 2007), the dentition of the middle Miocene (MN6) large-sized ‘A.’ robinsoni is clearly more advanced, rather resembling that of forms such as Mioechinus oeningensis. Despite its small size, ‘A’. golpae (earliest MN9 species represented by a mandibular fragment with p4-m3) seems also relatively advanced, lacking the slenderness of molars and mandible that characterize the late Oligocene and early Miocene Amphechinus. Consistently, in their summary work on the middle Miocene small mammal record from the Calatayud-Montalbán Basin, Van der Meulen et al. (2012) prefer to refer to the erinaceine material of this period as ?Amphechinus. Despite its old age, a small-sized maxillary fragment with P3- M3 from the Vallès-Penedès Basin (Sant Mamet, MN4, late early Miocene; Crusafont et al., 1955) was not assigned to Amphechinus, but to Mioechinus (M. butleri). This choice seems justified given its Fig. 1. MN10-13 sites in the Teruel Basin with Erinaceinae. 1: Ması́a de La Roma 3- surprisingly modern (‘molarized’) dental habitus with similarly- 4B-5-7-11; 2: Peralejos D; 3: Concud Cerro de La Garita, Concud 3; 4: Masada del Valle 2, Tortajada A, Masada Ruea 2-3; 5: Puente Minero, Puente Minero 2, Ması́a del sized and square-shaped P4, M1 and M2, and with a P3 with a well- Barbo 2A-2B; 6: Vivero de Pinos; 7: Los Aguanaces, Los Aguanaces 1-3, La Gloria 5- developed lingual part (even containing a hypocone). As referred to 10-11; 8: Los Mansuetos; 9: Aljezar B; 10: Prado 3, 15C; 11: Cascante-Cubla 1; 12: above, there is also little doubt that Postpalerinaceus vireti, defined Cascante 7A; 13: Patrimonio Forestal 5A. J.A. van Dam et al. / Geobios 61 (2020) 61–81 63 middle to late Miocene radiation will be explained in terms of diet- populations (notably Crusafont and De Villalta, 1947; Crusafont related morphological change that was ultimately rooted in a basic and Gibert, 1974; Engesser, 2009; Gibert, 1975; Mein and Oligocene masticatory apparatus aimed at consuming small Ginsburg, 2002; Viret, 1938). Finally, tooth rows of modern invertebrates. Finally, we invoke climatic cooling as a main cause European and African taxa (Erinaceus, Atelerix) were used for for the remarkably low erinaceine abundance that characterizes comparison (both museum material and published photographs, the latest Miocene and Pliocene of both Spain and Europe as a e.g., Gould, 2001). whole. The measurement method for molars and p4/d4 follows De Jong (1988). For the incisors, canine, premolars (except for P4) and 2. Material and methods deciduous teeth, length was measured as the maximum longitu- dinal distance, and width by taking it perpendicularly to it. In their The locations of the studied sites are shown in Fig. 1. The description of the Erinaceinae from La Grive (included in our stratigraphic context and general fauna have been described discussion), Mein and Ginsburg (2002) use the measurement elsewhere (e.g., Van de Weerd, 1976; Adrover, 1986; Mein et al., method of Mein and Martı́n-Suárez (1994) as used for Galericini 1990; Van Dam et al., 2001) or their description is in progress (sites (the other main subfamily of Erinaceidae). Because of the Concud Cerro de la Garita, Cascante 7A, Patrimonio Forestal 5A). symmetrical shape of the P3 of Erinaceinae, both measurement The bulk of the studied material is housed in the collections of the methods will yield almost equal results for this element. Department of Earth Sciences of Utrecht University, Netherlands, Comparisons between P4s, however, require additional care, due the Museo Aragonés de Paleontologı́a (Teruel, Spain), and the to the use of a different method. Although the Teruel Basin contains Collections de Géologie of the Centre de Ressources pour les very few P4 (only two fragments), the effect of the method is still Sciences de l’Evolution, Université Claude Bernard (Lyon, France). relevant here, as also material from other regions will be discussed. Tables 1–4 contain the complete list of the studied material. Differences in length and width measurement of the P4 may stem Summaries with element types and left/right assignments are from the fact that Mein and Martı́n-Suárez (1994) base their included in the systematic part below (Section 4). This study enclosing L-W rectangle on an anteriorly placed reference line essentially contains all upper Miocene erinaceine material from (touching the outline at the level of both protocone and metacone), the Teruel Basin known so far, except for material from Aljezar B whereas De Jong (1988) places his reference line along the buccal (MN12). This material, described as ‘Erinaceus’ sp. (Adrover, 1986; border, which is obliquely positioned with regard to the anterior Van den Hoek Ostende and Furió, 2005) could not be recovered and border in most Erinaceinae. Unfortunately, several of the older could therefore not be included in this study. relevant publications of Spanish material (Crusafont and de During our investigation, we have compared our material with Villalta, 1974; Crusafont and Gibert, 1974; Gibert, 1975) lack a the type material of key species from the Iberian Peninsula, clear description of how length and width are measured in the including Mioechinus butleri, Amphechinus robinsoni (Manchones) upper premolars. On the other hand, Crusafont et al. (1955) and Postpalerinaceus vireti (Viladecaballs, Can Llobateres, and Can mention that P4 length was measured along the buccal border Ponsic) stored in the Institut Català de Paleontologia Miquel (Mioechinus butleri). We assume that length was measured Crusafont (ICP) and in the Utrecht and Lyon collections. Part of the similarly in the later papers by Crusafont, Gibert and colleagues, specimens encountered in the ICP collection were re-measured. and that width was measured perpendicular to length (essentially Furthermore, we compared our specimens to the literature De Jong’s method). The scatter for P4 as included in this paper descriptions and figures from relevant Spanish and French should nevertheless be judged with care, as L and W for French (La Table 1 Studied material of Postpalerinaceus cf. vireti and measurements. Specimens from the Lyon collection are without individual catalogue numbers. Measurement method for molars, p4 and P3 after De Jong (1988), and for P4 after Mein and Martı́n Suárez (1994). Length for incisors, canines, P2 and p3 taken maximally, and width taken perpendicular to length. Locality Code MN Local Age Sup./Inf. Element Element Dex./Sin. Storage Catalogue Length Width Zone (Ma) type nb. nb. (mm) (mm) La Gloria 10 GLO10 11 K 8.4 sup. i 1 dex. UU(MAP) 381 Los Aguanaces 1 AG1 11 K 8.4 sup. di 2 sin. UU(MAP) 320 2.00 1.46 Ması́a del Barbo 2A MB2A 10 J1 9.8 sup. di 2 dex. UU 1129 1.36 1.37 Vivero de Pinos VIP 11 K 8.2 sup. di 2 dex. LY 1.40 1.44 La Gloria 10 GLO10 11 K 8.4 sup. di 2 dex. UU 382 1.70 1.53 La Gloria 11 GLO11 11 J4 8.8 sup. i 2 sin. UU 56 2.03 1.70 Ması́a de la Roma 3 ROM3 9 I 10.1 sup. i 2 dex. UU(MAP) 304 2.08 1.60 Ması́a del Barbo 2A MB2A 10 K 9.8 sup. c 1 sin. UU 1130 1.30 Los Aguanaces 3 AG3 11 K 8.2 sup. p 2 sin. UU(MAP) 2109 2.16 1.90 Vivero de Pinos VIP 11 K 8.2 sup. p 2 dex. LY 2.39 1.92 Ması́a del Barbo 2B MBB 10 J2 9.7 sup. p 2 sin. UU 2058 1.98 1.85 Los Aguanaces AG 11 K 8.2 sup. m 2 dex. LY 4.77 5.42 Peralejos D PERD 11 J4 8.8 inf. i 3 sin. UU(MAP) 630 1.70 Ması́a del Barbo 2B MB2B 10 J2 9.7 inf. i 3 sin. UU 2055 2.26 1.71 La Gloria 10 GLO10 11 K 8.4 inf. c 1 sin. UU(MAP) 384 3.29 2.05 La Gloria 10 GLO10 11 K 8.4 inf. c 1 sin. UU(MAP) 385 3.22 2.11 La Gloria 10 GLO10 11 K 8.4 inf. p 3 sin. UU(MAP) 386 2.15 1.83 La Gloria 10 GLO10 11 K 8.4 inf. p 3 sin. UU(MAP) 387 Vivero de Pinos VIP 11 K 8.2 inf. p 3 dex. LY 1.75 1.53 Vivero de Pinos VIP 11 K 8.2 inf. p 3 dex. LY 1.93 1.90 Vivero de Pinos VIP 11 K 8.2 inf. p 3 dex. LY 2.06 1.77 Cascante-Cubla 1 CCUB1 11 J4 8.8 inf. p 4 sin. LY Ması́a del Barbo 2B MB2B 10 J2 9.7 inf. m 1 sin. UU 2056 Ması́a del Barbo 2B MB2B 10 J2 9.7 inf. m 1 sin. UU 2057 Ması́a del Barbo 2B MB2B 10 J2 9.7 inf. m 1 dex. UU 2059 5.86 3.54 64 J.A. van Dam et al. / Geobios 61 (2020) 61–81 Table 2 Studied material of Atelerix aff. depereti and measurements. See Table 1 for measuring details. Locality Code MN Local Age Sup./Inf. Element Element Dex./Sin. Storage Catalogue Length Width Zone (Ma) type nb. nb. (mm) (mm) Los Aguanaces 3 AG3 11 K 8.2 sup. i 1 sin. UU(MAP) 2101 1.80 Vivero de Pinos VIP 11 K 8.2 sup. i 1 dex. LY 1.96 1.48 Masada Ruea 2 MRU2 11 K 8.5 sup. i 1 dex. MAP 189 1.28 Ması́a de la Roma 5 ROM5 10 J1 9.9 sup. i 1 sin. UU(MAP) 21 1.88 Ması́a de la Roma 7 ROM7 10 J1 9.8 sup. i 2 dex. UU(MAP) 259 1.40 1.04 Concud Cerro de la Garita CG 12 L 7.5 sup. i 3 dex. MAP 754 1.73 1.47 Los Aguanaces 3 AG3 11 K 8.2 sup. i 3 dex. UU(MAP) 2114 Los Aguanaces 3 AG3 11 K 8.2 sup. i 3 dex. UU(MAP) 2108 2.30 1.75 Vivero de Pinos VIP 11 K 8.2 sup. i 3 dex. LY 2.24 1.85 Ması́a de la Roma 11 ROM11 10 J2 9.5 sup. i 3 sin. UU(MAP) 310 Los Mansuetos LM 12 L 7.3 sup. c 1 sin. MAP 4005 2.42 1.67 Vivero de Pinos VIP 11 K 8.2 sup. c 1 dex. LY 1.77 1.62 Vivero de Pinos VIP 11 K 8.2 sup. c 1 sin. LY 1.99 1.67 Prado 15C PRD15C 10 J1 9.8 sup. c 1 sin. MAP 186 2.67 1.70 Ması́a de la Roma 3 ROM3 9 I 10.1 sup. c 1 sin. UU(MAP) 306 Ması́a de la Roma 3 ROM3 9 I 10.1 sup. c 1 sin. UU(MAP) 307 Los Aguanaces 3 AG3 11 K 8.2 sup. p 2 dex. UU(MAP) 2112 2.21 1.72 Los Aguanaces 3 AG3 11 K 8.2 sup. p 2 dex. UU(MAP) 2113 2.07 1.59 La Gloria 10 GLO10 11 K 8.4 sup. p 2 dex. UU(MAP) 383 2.41 1.75 La Gloria 10 GLO10 11 K 8.4 sup. p 2 sin. UU(MAP) 388 La Gloria 10 GLO10 11 K 8.4 sup. p 2 dex. UU(MAP) 389 Prado 3 PRD3 10 J1 9.7 sup. p 2 sin. MAP 189 1.87 1.54 Vivero de Pinos VIP 11 K 8.2 sup. dp 3 sin. LY 2.10 1.58 Vivero de Pinos VIP 11 K 8.2 sup. dp 3 sin. LY 1.97 1.50 Vivero de Pinos VIP 11 K 8.2 sup. dp 3 sin. LY 1.80 1.21 Los Aguanaces AG 11 K 8.2 sup. p 3 dex. LY 2.08 2.32 Vivero de Pinos VIP 11 K 8.2 sup. p 3 sin. LY 2.38 3.05 Pedregueras 2C PED2C 9 I 10.4 sup. p 3 dex. UU 1311 1.98 2.27 Los Aguanaces AG 11 K 8.2 sup. p 4 dex. LY 4.41 Ması́a de la Roma 11 ROM11 10 K 9.5 sup. p 4 dex. UU(MAP) 309 Los Aguanaces AG 11 K 8.2 sup. m 1 dex. LY 4.91 5.57 Vivero de Pinos VIP 11 K 8.2 sup. m 1 dex. LY 4.90 5.10 Puente Minero PM 11 K 8.4 sup. m 1 dex. MA 1753 5,5* 5,6* Los Mansuetos LM 12 L 7.3 sup. m 2 dex. UU 4001 Los Mansuetos LM 12 L 7.3 sup. m 2 sin. UU 4002 3.90 4.62 Los Aguanaces AG 11 K 8.2 sup. m 2 dex. LY 4.27 4.90 Vivero de Pinos VIP 11 K 8.2 sup. m 2 sin. LY 4.45 4.90 Prado 15C PRD15C 10 J1 9.8 sup. m 2 dex. MAP 190 Pedregueras 2C PED2C 9 I 10.4 sup. m 2 dex. UU 1312 Tortajada A TOA 11 K 8.2 sup. m 3 dex. UU 2001 1.69 3.17 Vivero de Pinos VIP 11 K 8.2 sup. m 3 sin. LY 1.61 2.95 Vivero de Pinos VIP 11 K 8.2 sup. m 3 sin. LY 1.69 3.06 Ması́a de la Roma 7 ROM7 10 J1 9.8 sup. m 3 dex. UU(MAP) 260 1.67 2.86 Ması́a de la Roma 3 ROM3 10 I 10.1 sup. m 3 dex. UU(MAP) 301 1.60 2.87 Concud 3 CC3 12 L 7.2 inf. i 3 sin. UU 3016 1.81 1.36 Concud 3 CC3 12 L 7.2 inf. i 3 dex UU 3017 2.04 1.59 Concud 3 CC3 12 L 7.2 inf. i 3 sin. UU 3018 2.02 1.62 Los Aguanaces 3 AG3 11 L 8.2 inf. i 3 sin. UU(MAP) 2110 2.46 Vivero de Pinos VIP 11 K 8.2 inf. i 3 sin. LY 1.77 1.62 Ması́a del Barbo 2A MB2A 10 J1 9.8 inf. i 3 dex. UU 2084 2.01 1.56 Ması́a de la Roma 9 ROM9 10 J1 9.8 inf. i 3 sin. UU(MAP) 158 1.98 Cascante 7A CASC7A 10 J1 9.5 inf. c 1 sin. MAP 9 Prado 15C PRD15C 10 J1 9.8 inf. c 1 sin. MAP 185 1.54 Prado 15C PRD15C 10 J1 9.8 inf. c 1 sin. MAP 187 2.46 1.71 Ması́a de la Roma 3 ROM3 9 I 10.1 inf. c 1 dex. UU(MAP) 302 2.37 1.55 Concud 3 CC3 12 L 7.2 inf. p 3 dex. UU 3030 2.02 1.39 Concud Cerro de la Garita CG 12 L 7.5 inf. p 3 sin. MAP 755 2.04 1.63 Concud Cerro de la Garita CG 12 L 7.5 inf. p 3 dex. MAP 756 1.63 1.63 Ması́a del Barbo 2B MB2B 10 J2 9.7 inf. p 3 dex. UU 2052 2.04 1.64 Ması́a del Barbo 2A MB2A 10 J1 9.8 inf. p 3 dex. UU 2083 1.92 1.85 Los Aguanaces 3 AG3 11 K 8.2 inf. p 4 dex. UU(MAP) 2111 Vivero de Pinos VIP 11 K 8.2 inf. p 4 sin. UU(MAP) 803 Vivero de Pinos VIP 11 K 8.2 inf. p 4 dex. LY 2.49 1.81 Prado 3 PRD3 10 J1 9.7 inf. p 4 sin. MAP 188 Puente Minero 2 PM2 10 J2 9.7 inf. p 4 dex. UU(MAP) 202 Vivero de Pinos VIP 11 K 8.2 inf. m 1 sin. LY Concud Cerro de la Garita CG 12 L 7.5 inf. m 2 sin. MAP 751 Concud Cerro de la Garita CG 12 L 7.5 inf. m 2 sin. MAP 752 5.00 2.99 Vivero de Pinos VIP 11 K 8.2 inf. m 2 sin. UU(MAP) 802 Vivero de Pinos VIP 11 K 8.2 inf. m 2 sin. LY 4.84 3.43 La Gloria 5 GLO5 13 M2 6.4 inf. m 3 dex. LY 1.98 1.62 Concud Cerro de la Garita CG 12 L 7.5 inf. m 3 dex. MAP 753 2.01 1.60 J.A. van Dam et al. / Geobios 61 (2020) 61–81 65 Table 2 (Continued ) Locality Code MN Local Age Sup./Inf. Element Element Dex./Sin. Storage Catalogue Length Width Zone (Ma) type nb. nb. (mm) (mm) Vivero de Pinos VIP 11 K 8.2 inf. m 3 dex. LY 2.31 1.84 Vivero de Pinos VIP 11 K 8.2 inf. m 3 dex. LY 2.43 2.17 Prado 15C PRD15C 10 J1 9.8 inf. m 3 dex. MAP 3 2.26 1.64 Pedregueras 2C PED2C 9 I 10.4 inf. m 123 dex. UU 1314 * Size extrapolation based on Alcalá et al. (1991: pl. 1, fig. 38). Table 3 Studied material of Atelerix steensmai nov. sp. and measurements. See Table 1 for measuring details. Locality Code MN Local Age Sup./Inf. Element Element Dex./Sin. Storage Catalogue Length Width Zone (Ma) type nb. nb. (mm) (mm) Masada del Valle 2 MDV2 12 L 7.6 sup. i 2 sin. UU 4051 2.40 1.59 Masada del Valle 2 MDV2 12 L 7.6 sup. c 1 dex. UU 4067 3.64 2.03 Masada Ruea 3 MRU3 12 L 7.5 sup. p 2 sin. MAP 364 2.71 2.02 Masada del Valle 2 MDV2 12 L 7.6 sup. p 2 dex. UU 4068 2.70 2.06 Masada del Valle 2 MDV2 12 L 7.6 sup. m 1 dex. UU 4074 Masada del Valle 2 MDV2 12 L 7.6 sup. m 1 dex. UU 4075 Masada del Valle 2 MDV2 12 L 7.6 sup. m 2 dex. UU 4072 4.59 5.27 Masada del Valle 2 MDV2 12 L 7.6 inf. i 3 dex. UU 4054 2.86 2.01 Masada del Valle 2 MDV2 12 L 7.6 inf. i 3 dex. UU 4055 1.97 Masada del Valle 2 MDV2 12 L 7.6 inf. c 1 sin. UU 4066 3.63 2.00 Masada Ruea 3 MRU3 12 L 7.5 inf. p 4 sin. MAP 363 Masada del Valle 2 MDV2 12 L 7.6 inf. p 4 sin. UU 4064 Masada del Valle 2 MDV2 12 L 7.6 inf. m 1 sin. UU 4071 Masada Ruea 3 MRU3 12 L 7.5 inf. m 2 dex. MAP 231 5.04 3.82 Masada del Valle 2 MDV2 12 L 7.6 inf. m 2 dex. UU 4061 5.70 3.80 Masada del Valle 2 MDV2 12 L 7.6 inf. m 2 sin. UU 4076 Masada del Valle 2 MDV2 12 L 7.6 inf. m 2 sin. UU 4077 Table 4 Studied material of Erinaceinae indet. and measurements. See Table 1 for measuring details. Locality Code MN Local Age Sup./Inf. Element Element Dex./Sin. Storage Catalogue Length Width Zone (Ma) type nb. nb. (mm) (mm) Los Aguanaces 3 AG3 11 K 8.2 sup. i 2 sin. UU(MAP) 2102 1.73 1.13 Los Aguanaces 3 AG3 11 K 8.2 sup. i 3 sin. UU(MAP) 2103 2.22 1.72 Ması́a de la Roma 3 ROM3 9 I 10.1 sup. m 2 dex. UU(MAP) 308 Los Aguanaces 3 AG3 11 K 8.2 sup. m 3 dex. UU(MAP) 2107 1.54 2.92 Patrimonio Forestal 5A PF5A 11 J4 8.8 sup. p 2 dex. MAP 52 Puente Minero 2 PM2 10 J2 9.7 sup. p 4 sin. UU(MAP) 201 Grive) specimens will be somewhat under- or over-estimated, et al., 2010). Recently, the remarkable occurrence of a M4 was respectively, with regard to the Spanish specimens. discovered in the living Asian species Mesechinus wangi (Ai et al., 2018). 3. Dental variation 3.1. Upper dentition The classification of isolated fossil Erinaceinae teeth is partly Upper erinaceine incisors have a convex external (buccal or hampered by the lack of sufficient comparative material. Especially antero-buccal) and concave internal (lingual or postero-lingual) variation in the anterior elements remains poorly known. side. I1 is the largest upper incisor. It is elongated, relatively Morphological convergence of posterior incisors, canine and curved, and usually has a sharp posterior edge. The single, long root anterior premolars is an additional complicating factor. The of the I1 extends below the posterior incisors and may reach the following paragraphs (which particularly focus on Miocene level of C. I2 is the smallest of the three upper incisors. Its crown is Erinaceinae from Europe) summarize our criteria for recognizing relatively pointed and it has one root. Its crown base is curving the different elements and for assigning these to the different upward in buccal direction. I3 is much larger and tends to be genera. caniniform, with both faces separated by a relatively sharp, The tooth formula of Erinaceinae is I 3/2 C 1/1 P 3/2 M 3/3, buccally placed longitudinal ridge. I3 tends to have a characteristic although an additional lower premolar is still present in primitive root positioning, with two parallel roots pointing backward with representatives (e.g., Paleoscapter from the Oligocene and lower regard to an anteriorly up-curving crown base (e.g., Mioechinus Miocene of Asia; Ziegler et al., 2007). Anomalies in tooth number tobieni in Engesser, 1980: figs. 22, 23). In some forms (e.g., modern seem to be common. For instance, variability in tooth numbers is Atelerix) either one or two roots may occur (Gould, 2001). In rare recorded in modern Erinaceus europaeus for the lower incisors and instances where fossil DI2-3 where found, they generally resem- upper premolars. P3 may be missing entirely in this species (Reeve, bled I2-3 (e.g., P. intermedius from La Grive; Mein and Ginsburg, 1994) and Atelerix (e.g., in the type species A. albiventris; Santana 2002). 66 J.A. van Dam et al. / Geobios 61 (2020) 61–81 Compared to I and P, C is relatively symmetric bucco-lingually. typically present in M1, although it is presumably absent in the It is dominated by a centrally positioned main cusp, from which a most primitive (latest Oligocene and earliest Miocene) Amphe- posterior crest and usually also an anterior crest are running chinus forms (Gibert, 1975). The protoconule is often absent. If downward. In the most primitive forms, the upper canine is larger present, its expression is variable: e.g., relatively distinct in P. and higher than I3 and P2. In the late Miocene Postpalerinaceus intermedius from La Grive (Mein and Ginsburg, 2002), but vireti, C is very large compared to the neighboring elements, with represented by a faint swelling only in the same species from the main cusp positioned relatively anteriorly and the anterior part Sansan (Engesser, 2009). The M1 has three roots. occupied by a high anterior shelf instead of a crest (observation on The M2 has a posteriorly tapering outline, because of a much the holotype specimen). In contrast to the configuration in I3, the stronger parastyle compared to the metastyle. A metaconule is two roots tend to diverge. DC resembles C, but its main cusp tends present in Mioechinus tobieni, Amphechinus baudeloti, A. robinsoni, to be situated at a more anterior position (for example Post- poorly developed or absent in P. intermedius, and absent (i.e. palerinaceus intermedius; Mein and Ginsburg, 2002: Fig. 5). integrated in the metaloph) in P. cingulatus and Atelerix depereti. If Generally, the P2 is small. Its occlusal outline varies between present, the metaconule is weaker than in M1. Crusafont and De rounded and sub-rectangular. The P2 contains one, antero-buccally Villalta, 1947 mention an isolated M2 with a metaconule in P. vireti placed main cusp, which is lower than its homologous cusp in C from Viladecaballs. The M2 has three roots. and I3, and from where a buccal cutting edge is running in The M3 is an antero-posteriorly reduced element with an posterior direction. Normally, the P2 contains two slightly elongated outline, which is oval- or drop-shaped. The protocone is diverging roots, which may be fused (as in A. edwardsi, Viret, the largest cusp. The smaller paracone may take the shape of a 1938) or incipiently fused (as in P. intermedius; Mein and Ginsburg, ridge (e.g., Atelerix depereti from Sansan, Engesser, 2009). One root 2002). The stronger, posterior root is flattened. In various forms is present in M. tobieni and various species of Amphechinus (Butler, (e.g., Mioechinus oeningensis, M. tobieni; Engesser, 1980: fig. 24a,c) 1948), whereas two, fused roots are present in Atelerix depereti the postero-lingual part of the crown extends downward over the (Mein and Ginsburg, 2002; Engesser, 2009). A variable root pattern posterior root, creating a sub-rectangular outline in occlusal view has been observed in P. intermedius from La Grive with either one and an apparently oblique position of the posterior root with or two (separate or fused) roots. regard to the posterior outline. A similar shape was described for P. intermedius from La Grive (site L3), whereas a more rounded 3.2. Lower dentition shape was pictured for Atelerix depereti from the same site (Engesser, 1980: fig. 24b; ‘M. sansaniensis’, later reassigned to A. The i2 is the lower counterpart of I1. It is chisel-like and less depereti by Mein and Ginsburg, 2002). According to Mein and curved than the latter. The flattened internal side may reveal an Ginsburg (2002), DP2 (P. intermedius from La Grive) more or less oblique (anterior) occlusal/wear surface that reminds that of resembles P2. rodent lower incisors (Viret, 1938). Its strong root extends far P3 has a sub-triangular outline with a strong buccal part and a backwards and may reach the level of the p4 (e.g., Amphechinus; poorly developed lingual part. Its shape and size can be interpreted Butler, 2010). as strongly reduced with respect to P4. The lingual part may The i3 is small and premolariform. It may be very low-crowned sometimes be well developed as in modern Erinaceus, in which its (as in P. intermedius) or somewhat more rounded and higher- shape nevertheless tends to be highly variable (Reeve, 1994), but crowned (as in P. vireti). It contains a longitudinally elongated, also in old forms such as Mioechinus butleri, where it even contains antero-buccally placed main cusp, which is integrated into the a hypocone (Crusafont et al., 1955). Length exceeds width in M. main longitudinal ridge. One posteriorly pointing root is present. tobieni and Atelerix depereti (Engesser, 1980; Mein and Ginsburg, The shape of the slightly larger c resembles that of i3, but its 2002), but the reverse is true for P. vireti (type material; Crusafont outline is more oval. It is also asymmetric with the main cusp and De Villalta, 1947). Mostly three roots have been observed. placed antero-buccally. It may show a tendency for bulbousness. Where encountered (e.g., P. intermedius, La Grive), DP3 are highly One strong, posteriorly pointing root is present that runs less variable in shape, with the lingual parts more poorly developed obliquely than in i3. The size of the lower canine approaches that of than in P3. p4 in Amphechinus/Postpalerinaceus, whereas it is much smaller P4 is a genuine carnassial tooth and easy recognizable by its than p4 in Erinaceus (Viret, 1938). high and sharp buccal cutting edge. This edge may be quite long, The first premolar is generally considered to be a p3 (but see causing the outline to become trapezium-shaped (e.g., in the Butler, 1948, 2010). It has the shape of a rounded triangle with a primitive A. edwardsi; Viret, 1938). The degree of molarization in straight posterior border. A single posterior root is present, the the P4 is still low in M. tobieni (small-sized, lingual length reduced), direction of which is less oblique than in c and i3. The p4 is a but is larger in M. oeningensis and A. depereti. By contrast, slightly molarized tooth with a short talonid. The protoconid is molarization is well advanced in forms such as M. butleri, P. large and sharp. The metaconid tends to be present in older forms intermedius (e.g., Engesser, 2009) and Erinaceus, in which the P4 is (e.g., Amphechinus arvernensis), whereas it is either absent (as in P. almost square-shaped. The P4 has three roots. Compared to P4, DP4 intermedius from Sansan) or distinctly present in younger forms has its lingual part reduced. (for instance, it is observed on the holotype of Postpalerinaceus M1 has a rectangular outline in most of the older early-middle vireti). The protolophid is positioned perpendicularly to the lingual Miocene forms (Amphechinus, Postpalerinaceus intermedius), border (e.g., in P. intermedius and P. vireti) or directed posteriorly whereas a square shape is common in younger forms (P. vireti, (for instance in Amphechinus robinsoni and Atelerix depereti). The d4 Atelerix, Erinaceus). An exception to this trend is formed by the late is very different from p4 in early forms such as A. edwardsi. In this early Miocene Mioechinus butleri, which shows square-shaped M1, species it is only slightly molarized and basically consists of a M2 and P4 that are similarly sized. Generally, however, M1 size massive posteriorly flattened cone with a small, incipient clearly exceeds that of P4 and M2 in most older forms, whereas the paraconid (Viret, 1938). The d4 of P. intermedius lacks a talonid. three teeth have a similar size in modern forms (e.g., Erinaceus). Its anterior part contains a small but distinct paraconid and The metastyle in M1 is usually stronger than the parastyle contains a tiny, incipient metaconid as well (Mein and Ginsburg, (exception: M. oeningensis). The metastyle may sometimes be very 2002). In Erinaceus, the shape of d4 resembles that of p4, but the pronounced, for instance in the rectangular shaped, ‘premolarized’ talonid is shorter, the cusps are less sharp, the paraconid is much M1 of ’Amphechinus’ baudeloti (Gibert, 1975). A metaconule is lower than the protoconid, and the metaconid is not distinct. J.A. van Dam et al. / Geobios 61 (2020) 61–81 67 As in the upper dentition, the second lower molar is relatively reduced with regard to the first molar in Amphechinus, whereas the size of both elements tends to converge in more advanced erinaceines. The trigonid is longer than the talonid in the m1 in all species of Postpalerinaceus. In this genus, the protolophid tends to run perpendicular to the lingual border, whereas in other forms the metaconid tends to be positioned more forward than the protoconid. In contrast to m1, the trigonid in m2 is not elongated. The paralophid, which may fully incorporate the paraconid, extends along the lingual border, in extreme cases reaching the metaconid (A. baudeloti; Gibert, 1975), thus closing the trigonid. Both m1 and m2 have two roots. The m3 basically consists of a trigonid, and has one root. 4. Systematic paleontology The systematic allocation of the studied teeth presented below is sometimes tentative. This is because elements are represented by one or several specimens only and because the variation in some elements is poorly known in the first place. For these reasons we preferred to have the description of each element immediately followed by a short comparison/discussion where appropriate, instead of presenting a complete description per species followed by general discussion. Specimens smaller than half a tooth are referred to as ‘fragm.’ (fragment) in the material lists. Abbreviations: LY: Collections de Géologie, Centre de Res- sources pour les Sciences de l’Evolution, Université Claude Bernard, Lyon, France; MAP: Museo Aragonés de Paleontologı́a, Spain; UU: Fig. 2. A-C. Postpalerinaceus cf. vireti from various localities. A: right I2, GLO10-382; B: left i3, MBB-2055; C1: left c, GLO10-384, postero-lingual view; C2: left c, GLO10- Department of Earth Sciences, Utrecht University, Netherlands; 384, postero-buccal view. D-H. Atelerix aff. depereti from various localities. D: right UU(MAP): Department of Earth Sciences, Utrecht University, to be I3, AG3-2108; E: left C, LM-4005; F: left p3, CG-755; G: left P2, PRD3-189; H: right deposited in Museo Aragonés de Paleontologı́a. m3, CG-753. Scale bar: 1 mm. Order Eulipotyphla Waddell, Okada et Hasegawa, 1999 Family Erinaceidae Fischer, 1814 of the two ridges that separate the convex from the concave face, Subfamily Erinaceinae, Fischer, 1814 whereas in GLO11-56 these additional cuspules seem to be lacking. Genus Postpalerinaceus Crusafont et Villalta, 1947 MBA-1129 contains a relatively small and pointed main cusp and a Type species: Palerinaceus (Postpalerinaceus) vireti Crusafont et broad basal rim. In this specimen, anterior and posterior ridges Villalta, 1947 extend down from the pointed main cusp, with the posterior one being the sharpest. Very tiny additional cusps occupy the ends of Postpalerinaceus cf. vireti Crusafont et Villalta, 1947 both ridges. A distinct, anteriorly climbing cingulum is present. Fig. 2(A, B, C1-2) Also GLO10-382 has a strong basal rim at the base of its concave 2001. Erinaceinae indet. - Van Dam et al., pro parte (specimens face, although it is less pronounced. The La Gloria specimen has its from Ması́a de la Roma 3, Ması́a del Barbo 2B). root preserved, which is rounded, but somewhat flattened 2011. cf. Postpalerinaceus - López-Guerrera et al., pro parte buccally. AG1-320 has a much weaker basal rim and an overall (specimen from Cañada 12). simpler shape. Similarly, the small specimen from VIP shows a Material (see Table 1 for locality abbreviations): AG: 1 RM2; simple, symmetric shape. Because of their smaller size, we AG1: 1 L(D?)I2; AG3: 1 LP2; CCUB1: 1 Lp4; GLO10: 1 RI1, 1 RI2, tentatively assign the four latter specimens to DI2. In the type 1 LP2 (fragm.), 1 RP2, 2 Lc, 1 Lp3, 1 Lp3 (fragm.), 1 fragm.; GLO11: population of P. intermedius from La Grive a weaker cingulum was 1 LI2; MB2A: 1 R(D?)I2, 1 LC; MB2B: 1 LP2, 1 Li3, 1 Lm1, 1 Lm1 used as one the features for separating I2 from DI2 (Mein and (fragm.), 1Rm1; PERD: 1 Li3; ROM3: 1 RI2; VIP: 1 RI2, 1 RP2, 3 Rp3. Ginsburg, 2002). Applied to our material, this criterion fits the Measurements: see Table 1 and Figs. 3–6. identification of AG1-320 as a DI2, but would not fit our Description and comparison: An external part of a large identification of the other two smaller-sized specimens. incisor from La Gloria 10 (GLO10-381, MN11) compares well to the The size of an anterior part of an upper canine (MBA-1130, external face I1 on the holotype skull of Postpalerinaceus vireti from MN10) implies a tooth size that is larger than that of A. aff. depereti Viladecaballs (MN10). Both specimens show the same moderately (see below), fitting that of P. vireti. Its central cusp is rounded degree of curvature. anteriorly and contains a faint and thin cingulum. Six I2/DI2 have rounded circumferences, are relatively sym- A complete P2 from La Gloria 10 (GLO10-383, MN11) has its metric, have a concave postero-lingual face, and one big, rounded main cusp worn flat. It lacks postero-central and lingual cuspules, root. The specimens from Ması́a de la Roma 3 (ROM3-304, MN9) but contains a very poorly developed postero-buccal cuspule. The and La Gloria 11 (GLO11-56, MN10) are relatively large, whereas lingual cingulum is restricted to the anterior part of the tooth. the specimens from Ması́a del Barbo 2A (MBA-1129, MN10), La Another antero-buccal half of a P2 from the same site (GLO10-388) Gloria 10 (GLO10-382, MN11, Fig. 2(A)), Los Aguanaces 1 (AG1- fits the morphology of the complete specimen. 320, MN11), and Vivero de Pinos (MN11) are smaller in size. A second premolar from Los Aguanaces 3 (AG3-2109, MN11) is ROM3-304 has a small, laterally placed cuspule at the base of each larger than the other P2 from this site (identified below as A. aff. 68 J.A. van Dam et al. / Geobios 61 (2020) 61–81 Fig. 3. Length-width scatters for I1, I2, I3, and C. See Tables 1–4 for L and W values. depereti). Its shape is rounded and it is relatively broad. Its main Two M2 from Los Aguanaces (AG, MN11) clearly differ in size cusp is pointing anteriorly, while its anterior face is sub-vertical. A and morphology. In size, the largest of the two approaches the postero-buccal shelf is present as well. The tooth contains at least largest known specimen of P. vireti from its type area (Can two cuspules, which are placed postero-centrally and lingually on Llobateres; Fig. 4). Because also the morphology of this specimen, a weak cingulum. (The additional presence of a postero-lingual with relatively straight borders and a lingual length that is cuspule cannot be checked because of damage). A thin anterior distinctly smaller than its buccal length, is similar to that of the Can cingulum is present. The broad posterior root is pointing Llobateres specimen (as figured in Crusafont and Gibert, 1974), we downward (in contrast to I3, on which it would be directed more are confident in assigning the larger of the two specimens from Los posteriorly). A specimen from Ması́a del Barbo 2B (MBB-2058, Aguanaces to P. cf. vireti. MN10) has a similar crown size and shape and posterior root The crown of i3 from Ması́a del Barbo 2B (MBB-2055, MN10; shape. This specimen, which lacks the anterior part of its main Fig. 2(B)) is not flat (as in A. depereti), but somewhat inflated. Its cusp, has its posterior shelf less well developed than the specimen main cusp is part of a longitudinal ridge of which the height from Los Aguanaces 3. gradually decreases towards the posterior border, and which ends J.A. van Dam et al. / Geobios 61 (2020) 61–81 69 Fig. 4. Length-width scatters for P2, P3, P4, M1, M2, and M3. See Tables 1–4 for L and W values; see caption Fig. 3 for symbols. in a posterior cuspule. Departing from this cuspule, a cingulum is Two lower canines from La Gloria 10 (GLO10-384 and 385, running along the lingual border towards the anterior apex, from MN11; Fig. 2(C1,2)) are relatively bulbous. The anterior part of where it is climbing back towards the main cusp. The specimen their occlusal surface is worn flat (as in P2). A small postero-buccal resembles the i3 on the holotype of P. vireti from Viladecaballs, cuspule is present. The postero-lingual border is characterized by a although it is somewhat smaller and lower-crowned. Unfortu- low and broad cingulum ridge. Both teeth have very strong, nately, the morphology of the lingual cingulum could not be backwards pointing roots. compared to that of the holotype tooth, because the latter is both A p3 from the same site (GLO10-386) has a rounded outline. It is worn and damaged. The posteriorly damaged specimen from wider anteriorly than posteriorly. The postero-lingual corner is the Peralejos D (PERD-630, MN11) is slightly less wide. It has a strong most reduced corner. The buccal border shows a little re-entrant main cusp, which is included in a strong, antero-posteriorly between the larger anterior and smaller posterior part. A robust running ridge. main cusp and a smaller posterior cusp are present, with no 70 J.A. van Dam et al. / Geobios 61 (2020) 61–81 Fig. 5. Length-width scatters for i3, i3, c, and p3. See Tables 1–4 for L and W values; see caption Fig. 3 for symbols. connecting ridge in between. The strong, posteriorly placed root is 1992. Amphechinus cf. robinsoni - Alcalá et al. slightly curved posteriorly. An isolated posterior part from the 2001. Erinaceinae indet. - Van Dam et al. (specimens from Ması́a same site (GLO-387) fits the morphology of the complete de la Roma 7, Concud 3 and Los Mansuetos). specimen. Three p3 from Vivero de Pinos have a similar rounded 2001. Erinaceinae indet. - Van Dam et al., pro parte (specimens outline. Their root is straight, pointing slightly posteriorly. from Ması́a de la Roma 3, Ması́a del Barbo 2B). The m1 from Ması́a del Barbo 2B (MBB-2059, MN10) has an Material (see Table 2 for locality abbreviations): AG: 1 RP3, elongated trigonid. The five main cusps have more or less equal 1 RP4, 1 RM1, 1 RM2; AG3: 1 LI1, 2 RI3, 1 LP2, 2 RP2, 1 Li3, 1 Rm1 heights, except for the hypoconid, which is somewhat lower. The (fragm.); CC3: 2 Li3, 1 Ri3, 1 Rp3; CG: 1 RI3, GLO5: 1Rm3; LM: 1 LC, protoconid-hypoconid and metaconid-entoconid connections are 1 LM2, 1 RM2; MB2A: 1 Ri3, 1 Rp3; MB2B: 1Rp3; MRU2: RI1; very thin and low. There is no sign of a posterior crest of the PED2C: 1RP3, 1RM2 (fragm.), 1 fragm.; PM: 1RM1; PM2: 1Rp4; paraconid, leaving the central trigonid valley open lingually. The PRD3: 1LP2, 1Lp4 (fragm.), 1 fragm.; PRD15C: 1LC, 1RM2 (fragm.), posterolophid ends freely against the entoconid, without contact- 2 Lc, 1 Rm3; ROM3: 2 LC (fragm.), 1 RM3, 1Rc; ROM5: 1 LI1; ROM7: ing the posterior cingulum. This latter configuration also 1 RI2, 1 RM3; ROM9: 1 Li3; ROM11: 1 LI3, 1 RP4 (fragm.); TOA: characterizes the talonid fragment MBB-2057. A strong posterior 1RM3; VIP: 1 RI1, 1 RI3, 1 LC, 1 RC, 1 LP3, 2 LDP3, 1 RDP3, 1 RM1, cingulum runs down from the entoconid, rounds the hypoconid 1 LM2, 2 LM3, 1 Li3, 1 Lp4, 1 Rp4, 1 Lm1, 2 Lm2, 2 Rm3. and (in the complete MBB-2059) finally reaches the level of the Measurements: see Table 2 and Figs. 3–6. paralophid. A fragment of a talonid from the same site (MBB-2056) Description and comparison: An external half of an I1 from Los with protoconid and metaconid present is assigned to Post- Aguanaces 3 (AG3-2101, MN11; Fig. 7(F)) has a curved shape. The palerinaceus cf. vireti as well. transition to the root is interrupted due to a cingulum-like extension at the base of the postero-lingual edge. The curvature is Genus Atelerix Pomel, 1848 stronger than in the I1 from La Gloria 10 assigned to Post- Type species: Erinaceus albiventris Wagner, 1841 palerinaceus cf. vireti (see above), and on the holotype from the latter species from Viladecaballs, which also is distinctly longer Atelerix aff. depereti Mein et Ginsburg, 2002 (2.7 mm). The length of the AG3 specimen (1.80 mm) is Figs. 2(D-H), 7(A-F), 8(I-J) intermediate between that of Atelerix depereti (1.6-1.7 mm) and 1986. Postpalerinaceus vireti - Adrover, pro parte (specimens P. intermedius (2.0-2.3 mm) from la Grive. Because the figured I1 from Los Aguanaces, Vivero de Pinos). crown from La Grive (Mein and Ginsburg, 2002: fig. 4) lacks a basal J.A. van Dam et al. / Geobios 61 (2020) 61–81 71 Fig. 6. Length-width scatters for p4, m1, m2, and m3. See Tables 1–4 for L and W values; see caption Fig. 3 for symbols. posterior cuspule, we tentatively assign AG3-2101 to A. aff. buccal side is convex. Its main cusp is pointed. The lingual face is depereti. An I1 from VIP (L =1.96 mm) also shows a postero-lingual bounded by a relatively weak cingulum, which smoothly trans- extension at its base. A buccal part of an incisor (MRU2-189, MN11) forms into a climbing ridge that runs anteriorly and climbs upward is similarly shaped as the one from AG3-2101. A similarly sized towards the main apex. Another ridge that separates buccal from (L =1.88 mm) buccal fragment from Ması́a de la Roma 7 (ROM5-21, lingual face runs down posteriorly from the apex, and meets with MN10) is tentatively attributed to A. aff. depereti as well. the lingual cingulum. In contrast to the configuration in Post- A small-sized I2 from Ması́a de la Roma 7 (ROM7-259, MN10; palerinaceus cf. vireti, the single root is antero-posteriorly flattened Fig. 7(E)) has a strongly flattened crown with a short anterior face with a curvature that mirrors that of the crown. The shape and elongated posterior face. The lingual side is concave and the conforms to the description of I2 from A. depereti from La Grive, on which a strong antero-posterior asymmetry has been observed. Three I3 were recovered from Los Aguanaces 3 (MN11; Fig. 2(D)). Their crowns contain a strong, antero-buccally placed pointed cusp, from which a pronounced ridge is descending along the buccal border, thereby creating a distinct cutting blade with a steep postero-lingual wall. This ridge ends in a postero-buccal cuspule. In one of the three specimens (AG3-2109) a postero- central cuspule is present as well. This cuspule is positioned on a faintly visible cingulum, which continues along the lingual side, reaching the level of the main cusp. A strong, antero-posteriorly flattened posterior root is preserved in AG3-2108. AG3-2114 is a postero-lingual part that is relatively unworn, containing a very pointed apex. A buccal fragment from Ması́a de la Roma 11 (ROM11-310, MN10) contains a main cusp from which a sharp ridge is descending buccally, ending in a small postero-buccal cuspule. All specimens have at least onsets of the posterior root Fig. 7. A-F. Atelerix aff. depereti from various localities. A: right P3, PED2C-1311; B: right P2, AG3-2112; C: right p3, MBA-2083; D: right M3, ROM7-260; E: right I2, preserved. This root is antero-posteriorly flattened and stands out ROM7-259; F: right I1, AG3-2101. G. Erinaceinae genus and species indet. from Los at an oblique angle (458) with regard to the crown base. AG3- Aguanaces 3, right M3, AG3-2107. Scale bar: 1 mm. 2108 and another specimen from Vivero de Pinos (VIP, MN11) also 72 J.A. van Dam et al. / Geobios 61 (2020) 61–81 have (part of) their anterior root preserved. This root is also Because such a lingual extension is less distinctly present in A. flattened and runs parallel to the posterior one. An I3 from Concud depereti from Sansan (formerly named ‘M. sansaniensis’, see same Cerro de la Garita (CG-754, MN12) is smaller than the other I3, figure), we assume that it developed during the evolution from this although this is partly due to dissolution and/or transport. It has a species towards A. aff. depereti. posterior cuspule, and another small cuspule along the lingual An unworn P3 from Pedregueras 2C (PED2C-1311, MN9; crown border. The presence of an anterior-lingual cingulum cannot Fig. 7(A)) has the shape of a right-angled triangle, the right angle be confirmed because of the rounding due to erosion. Whereas the being the antero-buccal corner. Its buccal part is dominated by a posterior root shows the characteristic oblique positioning of I3, large paracone, from which a sharp ridge is descending towards the the partly preserved anterior root is placed more or less posterior apex, from where it is curving back lingually. The lingual perpendicular to the crown base. In this respect (and in its part is separated from the buccal part by a groove. It contains a relatively small size) the specimen differs from other I3 described pointed protocone, which is smaller than the paracone. The here. We consider these differences to be too small for separating asymmetric shape of the P3 is more similar to that of A. depereti the tooth taxonomically from the rest of the material. Alternative- (Engesser, 1980: fig. 24) than to that of P. intermedius (Mein and ly, the specimen could represent a DI3. Ginsburg, 2002: Fig. 8; Engesser, 2009: fig. 49), of which the lingual An upper canine from Los Mansuetos (LM-4005, MN12; part is poorly developed, rounded and placed more centrally, and P. Fig. 2(E)) has its main cusp anteriorly placed in a slightly buccal vireti, of which the P3 from its type population Viladecaballs shows position. Its rounded posterior part is less wide than its anterior a very poorly developed lingual part containing only a low, part. From the main cusp a blunt, straight crest is running down anteriorly positioned protocone. A second A. aff. depereti P3 from centrally, where it ends in a small buccally positioned posterior Vivero de Pinos resembles PED2C-1311, although its posterior cuspule. A very weak anterior cingulum is present. The two roots border is less emarginated, and its preprotocrista is continuous. A are about equally strong, with the straight posterior one being the third, relatively small specimen from Los Aguanaces (AG, MN11) strongest. The anterior root is slightly curved posteriorly. Because has its posterior margin emarginated and its preprotocrista of its relatively small length (Fig. 3), we consider an assignment to uninterrupted. Three DP3 from Vivero de Pinos (VIP, MN11) have A. aff. depereti more probable than an assignment to the much their lingual parts strongly reduced compared to P3. A sign of a larger P. vireti or A. steensmai nov. sp. (see below). A posterior part protocone is nevertheless visible on two specimens. of a small-sized upper canine from Prado 15C (PRD15C-186, A fragment of a P4 from Ması́a de la Roma 11 (ROM11-309, MN10) is assigned to A. aff. depereti as well. Two anterior fragments MN10) contains the lingual parts of the protocone and hypocone. (ROM3-306 and 307, MN10) also suggest a similar, although Because cusps are sharp and curved rather than blunt, we somewhat smaller tooth size. Anterior crests are lacking on these tentatively assign the specimen to A. aff. depereti (and not to P. specimens. Although the posterior parts of the main cusp area not vireti). A buccal part of a P4 from Los Aguanaces shows a distinct preserved, onsets of both lingual and buccal cingula are visible. A small P2 from Prado 3 (PRD3-189, MN10; Fig. 2(G)) has a relatively narrow posterior part, although it is broader than in the specimen of A. depereti from Sansan (MN6) as figured by Engesser (1980: fig. 24b), which, unlike the specimen from Prado 3, also lacks a posterior cuspule. On the other hand, P2 from La Grive (MN7/8; Mein and Ginsburg, 2002) have their posterior part well developed, with lingual and buccal sides ‘‘également modérément bombées’’ (moderately inflated to a similar degree), a configura- tion fitting the one in Fig. 2(G). In addition, our specimen contains a well-developed anterior cingulum, a feature also described for the La Grive P2. Except for a slightly narrower posterior part, our specimen is almost identical to an unidentified erinaceine P2 from Cañada 6 (MN9) from the neighboring Calatayud-Montalbán Basin, 100 km North of Teruel (Lopez-Guerrera et al., 2011: fig. 2(18); note that ‘p2’ should be ‘P2’ in their text). As it cannot be decided with certainty whether this specimen belongs to A. depereti or A. aff. depereti, we think a designation A. cf. depereti for this early late Miocene specimen would be appropriate. Two P2 were recovered from Los Aguanaces 3 (AG2-2112 and 2113, MN11; Fig. 7(B)). Their posterior part is better developed compared to that of the Prado 3 specimen, with an occlusal outline that is more or less rectangular, except for the anterior edge, which is relatively rounded. The extended postero-lingual corner is lower than the rest of the crown, thereby creating a large, obliquely positioned cutting blade. From the main, pointed cusp a buccally placed, sharp ridge descends towards the postero-buccal cuspule. From this cuspule a cingulum runs downward and anteriorly towards a small lingual cuspule placed halfway the lingual border. Other cingula are present lingually and anteriorly. Complete Miocene erinaceine tooth rows found in France and Anatolia (Mioechinus oeningensis, M. tobieni) show that the sub- Fig. 8. A-F. Atelerix steensmai nov. sp. A: right M2, MDV2-4072, holotype; B: left M1 rectangular outline, as observed for the P2 from AG3 and CG, is (fragm.), MDV2-4064; C: right m2, MDV2-4074; D: left C, MDV2-4067; E: right i3, MDV2-2054; F: right P2, MDV2-4068 in occlusal (F1) and buccal (F2) views. G, caused by the lingual extension of the posterior part of the tooth H. Erinaceinae genus and species indet. from Los Aguanaces 3. G: left I2, AG3-2102; crown, resulting in an apparently oblique positioning with regard H: left I3, AG3-2103. I, J. Atelerix aff. depereti from Concud Cerro de la Garita and Los to the overall tooth row direction (e.g., Engesser, 1980: fig. 28). Mansuetos. I: left M2, LM-2002; J: left m2, CG-752. Scale bar: 1 mm. J.A. van Dam et al. / Geobios 61 (2020) 61–81 73 posterior emargination, which points to a low degree of fragments of M2 from Ması́a de la Roma 3 and Prado 15C were molarization. Also this feature would exclude an inclusion into tentatively assigned to this species as well. Postpalerinaceus vireti, the P4 of which hardly show a posterior Two unworn M3 from Ması́a de la Roma 3 (ROM3-301, MN9) emargination, and whose anterior and posterior borders tend to and Ması́a de la Roma 7 (ROM7-260, MN10; Fig. 7(D)) are very run parallel (Crusafont and De Villalta, 1947; see also López- similar in size and shape. Their outline is elongated and more or Guerrero et al., 2011). less ellipsoid. The ROM7 tooth is slightly more robust and its cusps Three M1 of MN11 age (AG, VIP, PM) have a sub-squared shape are somewhat better developed. Three cusps are present, the without a strong metastyle (as for example in P. intermedius). The height of which decreases in buccal direction. The lingual and W/L ratios of 1.13 and 1.04 for the specimens from AG and VIP (the intermediate cusp (protocone and paracone according to Butler, end part of the metastyle is broken off in the PM specimen, Alcalá 1948: fig. 13) are integrated into a main ridge. The more isolated, et al., 1991: pl. 1, fig. 38) are low compared to ratios of rectangular buccal cusp (parastyle) is connected to the intermediate cusp via a shapes, such as in ’Amphechinus’ robinsoni from Manchones (1.35), small ridge. No roots are preserved. A slightly larger, unworn M3 being somewhat more close to ratios of squared shapes, like in P. from Tortajada A (TOA-2001, MN11) is similarly shaped, although vireti from Viladecaballs (1.00; Crusafont and De Villalta, 1947) and its cusps are more individualized, as evidenced by the presence of a Can Llobateres (1.00-1.03; Crusafont and Gibert, 1974). Borders are valley behind the ridge between the almost equally important relatively straight, a condition that is also different from that in P. protocone and paracone. The main ridge, which integrates both vireti, in which emarginations characterize all sides (as in P4), main cusps, is slightly curved anteriorly, in contrast to the including the anterior one (see especially the drawings in straighter shape in the Ması́a de la Roma specimens. A tiny, short Crusafont and Gibert, 1974.) All four main cusps in M1 are well crest is running down posteriorly from the lingual cusp. The developed. The protocone is the strongest cusp. A distinct parastyle is low, as in the ROM specimens, but it is more isolated, metaconule is present, whereas a protoconule is only faintly being connected to the paracone by a very tiny ridge only. A small visible. An anterior cingulum is present, which broadens in buccal and short cingulum-like ridge is present postero-lingually. No direction. The strength of the posterior cingulum is variable. The root(s) are preserved. Two specimens from Vivero de Pinos (VIP, vestige of a lingual cingulum is present between the protocone and MN11) have a similar size, outline and simple cusp pattern as the hypocone in the VIP and AG specimens. specimen from Tortajada A. Like M2, also the M3 are similar in size An M2 from Los Mansuetos (LM-4002, MN12; Fig. 8(I)) has four and shape to M3 of the form classified as ‘Mioechinus’ sp. from well-developed main cusps. The position of the hypocone is Petersbuch (Ziegler, 2005: fig. 4L). slightly posterior to that of the metacone. Strong ridges connect A small i3 from Ması́a del Barbo 2A (MBA-2084, MN10) is very the protocone with the paracone on the one hand, and with the low-crowned, resembling the i3 of A. depereti from La Grive (Mein metacone and hypocone via the metaconule (present as a swelling) and Ginsburg, 2002). Its buccal border is relatively straight, on the other hand. The paracone and metacone are connected by a whereas the lingual border is strongly convex. Its antero-buccally low crest. The protoconule is lacking. A buccally widening placed main cusp is elongated longitudinally, constituting a cingulum is present along the anterior border, and a second narrow antero-posteriorly running ridge. Posteriorly, this ridge cingulum is present posteriorly. Whereas a buccal cingulum is ends in a small cuspule, from where a cingulum is running along lacking in LM-4002, another, fragmented specimen from the same the postero-lingual part of the tooth. Another i3 from the similarly site (LM-4001) shows a short cingulum-like structure in the aged site Ması́a de la Roma 9 (ROM9-158, MN10), which lacks its middle of the buccal border. LM-4002 resembles the much older lingual face, fits the MBA specimen well in size and morphology. M2 of A. depereti from Sansan (MN6, Engesser, 2009), except for a Four other i3 from Ması́a de la Roma 9, Los Aguanaces 3 and Concud much larger size and some details of the morphology, such as a 3 (AG3-2110, MN11; CC3 3016-3018, MN12) have a similar shape, complete instead of interrupted connection of the protoloph with except for a more developed postero-lingual corner and stronger the paracone. Similarities are even larger to the M2 from posterior cingulum, which looks like a true ridge. The relatively Petersbuch 6 (MN7-8, Germany) classified as ‘Mioechinus’ sp., large specimen from Los Aguanaces 3, which is postero-buccally both in terms of size and morphology (Ziegler, 2005: fig. 4K). damaged, has its posterior ridge positioned somewhat more Although M2 size is similar to that of P. vireti, the occlusal buccally than the MB2A specimen. A similarly aged specimen from morphology of the latter species is different, with the metaloph VIP has its posterior ridge placed relatively buccally as well. and metaconule placed in a more posterior position (both in Two lower canines from Prado 15C (PRD15C-185,187; MN10) Viladecaballs and Can Llobateres, MN9; Crusafont and De Villalta, are asymmetrical bucco-lingually, because of the buccally placed 1947; Crusafont and Gibert, 1974). In addition, the W/L ratio is central ridge. The main cusp is positioned anteriorly. The large lower in P. vireti (1.09 and 1.02) than in the Los Mansuetos lingual face is slightly concave. The central ridge, which descends specimen (1.18), which, in turn, are similar to average values for A. from the main cusp, ends in a small posterior cuspule that resides depereti from the various fissures of La Grive (MN7 fissures: on a cingulum. The single root (PRD15C-187) is flattened antero- 1.16 and 1.21, MN8 fissures: 1.13 and 1.19), but smaller than the posteriorly. Another specimen from Ması́a de la Roma 3 (ROM3- value for A. depereti from the MN6 locality Sansan (1.24). A still 302, MN9) has a damaged occlusal surface, preventing close larger value of 1.29 characterizes the M2 of ’Amphechinus’ robinsoni inspection of the central ridge and posterior cuspule. It is relatively from the MN6 site Manchones (Gibert, 1975; Fig. 4). short, showing a cingulum anteriorly and postero-lingually. Its M2 from Los Aguanaces and Vivero de Pinos (MN11) resemble root is flattened anteriorly. A lingual fragment with a partial root the Los Mansuetos specimen in size and shape. W/L ratios are from Cascante 7A (CASC7A-9, MN10) fits the ROM3 specimen in somewhat smaller (1.15 and 1.10, respectively). Whereas a terms of morphology and is attributed to A. aff. depereti as well. metaconule can be observed to be present on the AG specimen, An unworn p3 from Ması́a del Barbo 2A (MBA-2083, MN10; it is hardly visible on the specimen from VIP. Also, the morphology Fig. 7(C)) has a triangular shape with its buccal and lingual sides of a buccal fragment of an M2 from Pedregueras 2C (PED2C-1314, converging towards the anterior apex. Its moderately strong MN9) fits that of LM-4002 (Fig. 8(I)), showing a metaconule and a central cusp is pointed. From this cusp a sharp crest is running posterior emargination that is not acute. Unlike this specimen, towards the anterior apex, where a cuspule is formed. Strong PED2C-1314 has its protoloph interrupted close to the paracone, lingual and posterior cingula enclose a concave valley. The and a very thin buccal cingulum. Because no features were posterior cingulum is strongest lingually, where it tends to form observed that contradict an assignment to A. aff. depereti, two small a cuspule. Four other p3 (MBB-2052, CG-755-756, CC3-3020) have 74 J.A. van Dam et al. / Geobios 61 (2020) 61–81 a similar triangular shape, but their width-length ratio is lower. Paratypes: 1 LI2, 1 RC, 1 RP2, 2 RM1, 1 RM2, 2 Ri3, 1 Lc, 1 Lp4, Also these specimens are characterized by a pointed main cusp and 1 Lm1, 2 Lm2, 1 Rm2. a pointed anterior apex, which is bended up and carrying a small Type locality and biostratigraphy: Masada del Valle 2 (MDV2), cuspule. The posterior cingulum supports a centrally placed Teruel Basin, east Central Spain. Local Biozone: L; Mammal posterior cuspule. MBB-2052 (MN10) shows little cingulum Neogene Unit: MN12; Age: 7.6 Ma. (Van de Weerd, 1976; Van development, but possesses a distinct posterior cuspule and a Dam et al., 2001). smaller anterior cuspule. CG-755 (MN12; Fig. 2(F)) resembles Occurrence: besides type locality: Masada Ruea (MRU3; Fig. 1): MBA-2083, although it is smaller and more elongated with less 1 LP2, 1 Lp4, 1 Rm2. well developed cingula. The single and strong posterior root (as Measurements: see Table 3. preserved in both CG specimens) is directed posteriorly, but less Diagnosis: Large erinaceine, with size comparable to that of P. obliquely than in i3. It is antero-posteriorly flattened. The vireti. Strong development of cusps on upper and lower molars; specimen from Concud 3 (MN12) is somewhat rolled off, but still development of ridges subdued. Outline of M1-2 squared; buccal shows posterior and anterior cuspules. Its strong, obliquely placed and posterior emarginations very shallow. Upper incisors (I2) root is square-shaped. pointed and relatively symmetric; upper canine with anterior The combination of an anteriorly pointed shape and broad crest. P2 rounded and with two roots. M1-2 without protoconule, posterior part is characteristic for the p3 of A. depereti (Mein and but M1 with strong metaconule. M2 without metaconule and Ginsburg, 2002). By contrast, p3 of P. intermedius (Mein and without centrally placed crests (metaloph and hypocone crest), but Ginsburg, 2002: fig.17, re-identified herein as a left specimen) are with direct connections of the hypocone with both protocone and somewhat more inflated as especially expressed by a more metacone, resulting in a squared pattern of connections between rounded antero-buccal border. the four main cusps. i3 low-crowned, wide posteriorly and with The p4 is represented by two fragments only. A buccal fragment buccally placed central crest. Bulbous lower canine. m1 with from Puente Minero 2 (PM2-202, MN10) shows the remains of a paraconid and especially metaconid positioned relatively buccally. metalophid that is running backwards. A fragment from Prado 3 m2 with squared talonid with shallow buccal emargination, and (PRD3-188, MN10) could only tentatively be assigned to A. aff. with very poorly developed ridges between metaconid and depereti based on the attribution to this species of another element entoconid and between protoconid and hypoconid. Lingual ridge (P2). posteriorly of paraconid of m2 short, thereby creating an open The lower first molar is poorly represented as well. An isolated trigonid. talonid from an m1 from Vivero de Pinos suggests an original width Differential diagnosis: Differs from Atelerix (aff.) depereti, of 2.7-2.8 mm, much smaller than the width of m1 of Post- Mioechinus tobieni, Postpalerinaceus intermedius and P. cingulatus by palerinaceus vireti (Fig. 6). It is assigned to A. aff. depereti, the its larger size, and by its molar morphology, which is characterized smaller-sized form found at this site. An antero-lingual fragment by straighter borders and a stronger development of cusps at the (AG3-2111, MN11), which only contains paraconid and metaconid, cost of ridges. Differs from P. vireti in its flattened and more is also assigned to this species. symmetric I2, its development of straight borders and squared An m2 from Concud Cerro de la Garita (CG-752, MN12; Fig. 8(J)) pattern of the main ridges and the absence of a metaconule in M2, has its talonid and trigonid sub-equal in length. Its lingual side is the flat occlusal surface of i3, its less bulbous lower canine with slightly convex at the position of the metaconid. A thin and short central cusp positioned more anteriorly, and its m2 with ridge runs posteriorly of the cusp-shaped paraconid and partly straightened buccal border, well-developed cusps and poorly closes the trigonid. The posterolophid is not free ending, but meets developed longitudinal ridges. with the upward running posterior cingulum at the base of Description and comparison: entoconid. The posterior cingulum ends at the base of the Type material. Although much larger, the I2 from Masada del hypoconid, but resumes as a buccal cingulum running all the Valle 2 (MDV2-4051, MN12) has a crown morphology that is way to the anterior apex. A similarly-sized specimen from Vivero superficially similar to that of A. aff. depereti from ROM7 (MN10; de Pinos (MN11) has its paraconid more ridge-shaped than the described above), which contains a similarly shaped, relatively specimen from Concud Cerro de la Garita. narrow, lingually positioned climbing main ridge. The main cusp is Two minimally worn m3 from Prado 15C (PRD15C-3, MN10) slightly less pointed than in the ROM7 specimen, because of a less and an unworn m3 from Concud Cerro de la Garita (CG-753, MN12; concave postero-lingual face, yielding a more symmetric shape in Fig. 2(H)) are very similar in size and shape. They basically consist postero-lingual view. MDV2-4051 shows a short continuation of of a trigonid only. The protoconid is slightly stronger than the the cingulum beyond the meeting point with the posterior central metaconid. The paraconid is incorporated into a ridge that ridge, where a small cuspule is present. No root is preserved. connects it with the metaconid and which rounds the anterior The crown of the upper canine (MDV2-4067; Fig. 8(D)) is apex. The trigonid is open buccally. A younger specimen from La dominated by one strong main cusp and a longitudinal crest that Gloria 5 (MN13) is similar in size and shape, whereas two extends across the entire length of the tooth. The anterior part of specimens from Vivero de Pinos (MN11) are somewhat larger. One this crest is shifted lingually. The lingual face of the tooth is slightly of these has its anterior ridge extending relatively far backward concave. A narrow antero-lingual cingulum is present. The along the lingual side, without closing the trigonid. posterior cingulum stands out as a small plateau. Lingually, it reaches a point half-way the tooth, at the level of the main cusp tip. Atelerix steensmai nov. sp. Two strong roots are present, with the posterior one being the Figs. 8(A-F) strongest. It has its posterior side rounded. 2001. Erinaceinae indet. - Van Dam et al. The P2 (MDV2-4068; Fig. 8(F1,2)) has its central crest shifted Derivation of name: in honor of Dr. Karel Steensma, for his buccally. This crest ends in a cuspule posteriorly. No other significant contribution to mammal paleontology and stratigraphy, posterior cuspules are present. Posterior and anterior borders are including various fieldwork campaigns in the Teruel Basin and rounded, resulting in a symmetric shape in occlusal view. A poorly other Spanish basins. developed antero-lingual cingulum is present. The P2 has two Holotype: RM2, no. MDV2-4074, stored in the Department of antero-posteriorly flattened roots. Earth Sciences, Utrecht University, the Netherlands (Fig. 8(A)). L Two lingual parts of M1 (MDV2-4074, 4075; Fig. 8(B)) show a =4.59 mm, W =5.27 mm. very strong metaconule, situated buccally of the meeting point of J.A. van Dam et al. / Geobios 61 (2020) 61–81 75 hypocone crest (entoloph) and postprotocrista. Lingual borders are flattened I2 and i3, and the shape of the lower canine) echoes relatively straight, a condition opposite to that in Postpalerinaceus, features observed in A. aff. depereti, suggesting a phylogenetic in which the M1 shows a strong re-entrant between protocone and relationship. Because ridges are also relatively subdued in species hypocone. MDV2-4075 contains a little cingulum-like ridge of modern Atelerix (e.g., A. albiventris vs. Erinaceus intermedius; between proto- and hypocone. Gould, 2001), we prefer the assignment to this existing genus The complete M2 (MDV2-4072; Fig. 8(A), holotype) is relatively instead of erecting a new genus. squared due to its large length. All four main cusps are well developed and there is no sign of a meta- or protoconule. The Erinaceinae indet. connection between protocone and hypocone is straight and not Figs. 7(G), 8(G, H) curving inward to form a postprotocrista, the result being a Measurements: see Table 4. squared pattern of connections between the four main cusps. Description: Three teeth from Los Aguanaces 3 (an I2, I3 and Except for a strong protoloph, ridges are poorly developed. Cingula M3) could not be assigned to any of the other species. The I2 and I3 are absent both lingually and buccally of the metacone. The clearly belong to the same species, as both specimens show the anterior cingulum is strong and widening buccally. same particularly style of fossilization, with a beige crown and dark The shape of three i3 (MDV2-4053, 4054, 4055; Fig. 8(E)) is brown root. The M3 perhaps belongs to the same species as well, similar to that of A. aff. depereti, although the teeth are much larger but could also represent another form. and relatively wide posteriorly, with a central ridge that is placed The intermediately-sized I2 from (AG3-2102, MN11; Fig. 8(G)) more buccally. The remnants of a single, posterior root show that it shows a rounded circumference of both the crown and the root was pointing in posterior direction. (excluding an assignment to the Atelerix species described above), The bulbous lower canine is similarly-shaped as in A. aff. and has a very robust and protruding lingually and upward depereti, except that it is less asymmetrical bucco-lingually. It is running cingulum ridge that extends all the way up towards the similarly-sized as the lower canine on the holotype of Post- anterior part of the crown (excluding assignment to Postpalerina- palerinaceus vireti from Viladecaballs, but the latter specimen is ceus vireti). even more bulbous, with a central cusp that seems to be positioned The morphology of the I3 (AG3-2103; Fig. 8(H)) is very different less far anteriorly. from that of the other I3 described in this study. Its shape is The buccal half of a p4 shows a high protoconid, which descends compact instead of elongated (e.g., as in Engesser, 1980: figs. 22, towards a narrow posterior valley, behind which the buccal part of 23). The tooth is characterized by a flattened, triangular-shaped a tiny posterior ridge can be discerned. ‘‘carnassial’’ blade with both anterior and posterior ridge well The isolated trigonid from an m1 (MDV2-4071) has its developed. The posterior cingulum is elevated to form a plateau- paraconid and especially its metaconid positioned relatively like postero-lingual rim. This rim is not reaching as far anteriorly as buccally. This configuration is unlike that in P. vireti. its homologue in the I2. The I3 contains only one very strong root, A complete m2 (MDV2-4061; Fig. 8(C)) has cusps that are which is directed vertically. The presence of one root in I3 is not relatively isolated with large trigonid and talonid basins in very common in erinaceines; for instance, it has been observed in between. Metaconid and entoconid are stronger than protoconid populations of modern A. albiventris (Gould, 2001), thereby and hypoconid. The anterior apex is rounded with the paraconid creating a tension with the genus diagnosis (Frost et al., 1991). shaped as a ridge. Ridges between metaconid and entoconid and The M3 from the same site (AG3-2107, MN11; Fig. 7(G)) has a between protoconid and hypoconid are virtually absent. The sub-triangular to drop-like shape. Its intermediate cusp (paracone) lingual ridge posteriorly of the paraconid is short, which results in is only slightly smaller than its lingual cusp (protocone), to which it an open trigonid. The short posterior cingulum runs downward, is connected by a curved crest. The buccal cuspule (parastyle) is until it abruptly ends. Its upper end is connected to a thin relatively isolated from the paracone (more than in Atelerix aff. posterolophid. A weak buccal cingulum extends along the trigonid depereti; Fig. 7(D)) and is connected to it by a very thin ridge only. part of the tooth. Also, an isolated talonid (MDV2-4076) shows a The AG3 specimen has well-developed cingula along both long very weakly developed posterolophid that almost disappears in sides. It contains a large, broad root, because of the fusion of its the middle of the posterior border. lingual and buccal part. The type of outline, the presence of a Other material. Masada Ruea 3 (MN12), which is geographically parastyle, and the well-developed cingula, are features that are very close to Masada del Valle 2 (Fig. 1), has a very similar age (Van also observed in A. depereti (Mein and Ginsburg, 2002; Engesser, Dam et al., 2001). A P2 from this site (MRU3-364) is equal in size 2009). Differences include a smaller size and the presence of two and shape to MDV2-4068, with a sub-rectangular outline with (partly fused) roots in A. depereti. The shape of the tooth is very rounded corners, an anteriorly placed, sharp main cusp, and a similar to that of the recent Atelerix albiventris (Gould, 2001). buccally placed central ridge. Another observed feature described for Atelerix is the presence of a A fragment of a p4 (MRU3-363) shows a curved metalophid that small ridge behind the protocone, which may be homologous with descends posteriorly along the lingual border towards a small the hypocone (Butler, 1948). entoconid without forming a metaconid and leaving the trigonid Besides the AG3 material, three Erinaceinae fragments from wide open lingually. The posterolophid carries an elongated cusp Ması́a de La Roma 4B, Prado 3, and Patrimonio Forestal 5A in the middle and runs down towards the postero-buccal corner, (ROM4B-73, PRD3-190, PF5A-52) could not reliably be assigned to which is low compared to the postero-lingual corner. a specific element, let alone to one of the species (Fig. 9). An m2 (MRU3-231) resembles MDV2-4061 in its straight buccal border, in a small ridge posteriorly of the paraconid (leaving the trigonid partly open), and a posterior cingulum that extends 5. Discussion downward to grade into a very thin basal cingulum, which continues until the anterior apex (with an interruption at the level 5.1. Iberian distribution of Neogene erinaceines of the hypoconid). Remarks: The strong development of cusps at the expense of We have identified four different species of Erinaceinae in the ridges on the molars as observed in A. steensmai nov. sp. has not upper Miocene (MN9-MN13) of the Teruel and Calatayud- been observed so far in other Miocene lineages. On the other hand, Montalbán Basins. Large-sized erinaceines are represented by the morphology of various anterior elements (for instance, the Postpalerinaceus cf. vireti and Atelerix steensmai nov. sp. Middle- 76 J.A. van Dam et al. / Geobios 61 (2020) 61–81 Fig. 9. Temporal distribution of Erinaceinae between 17 and 2 Ma in the Teruel Basin (black; this paper), Sarrión depression (blue; Adrover, 1974; Crochet and Heintz, 1971), and Calatayud-Montalbán Basins (red; López-Guerrero et al., 2011; Van der Meulen et al., 2012; with approximate ages for Cañada 6 (11.0 Ma) and 12 (7.2 Ma) based on López-Guerrero et al. (2011) and van Dam et al. (2014). Ages local Zones after Van Dam et al. (2006, 2014, ongoing work). Ages MN and marine units after Hilgen et al. (2012) (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article). sized forms are represented by Atelerix aff. depereti, the most Gloria 5; Mein et al., 1990). Also, the Pliocene record of the Teruel common erinaceine, and by a very rare, indeterminate form from basin is very poor in Erinaceinae (Mein et al., 1990), with one Los Aguanaces 3, the genus and species identifications of which occurrence in MN14 (La Gloria 4) and one in MN16 (Escorihuela). have been left open (Erinaceinae indet.). Alpha diversity is Two other younger occurrences have been reported for the maximal in the middle part of the study interval (MN11), with neighboring Sarrión Depression, one from MN16 (Sarrión 1; three species recorded in the single locality Los Aguanaces 3. Adrover, 1974), and one from MN17 (La Puebla de Valverde, A small temporal gap in spiny hedgehog occurrence is recorded MN17; Crochet and Heintz, 1971). The material from the latter site within MN10 (corresponding to the upper part of Zone J2 and Zone (a partial mandible with m2 and partial m1, and a p4) was J3). More remarkably, however, is the virtual absence of spiny classified as Postpalerinaceus cf. vireti, although according to the hedgehogs after MN12, with only a single occurrence in MN13 (La authors it could also represent a separate species. J.A. van Dam et al. / Geobios 61 (2020) 61–81 77 The best represented species in our material, Atelerix aff. butleri in the Vallès-Penedès Basin (Crusafont et al., 1955). An depereti, covers the entire study interval. We consider this form to older entry of the group cannot be excluded given that, unlike be a direct descendent of A. depereti, a smaller-sized form known rodents (e.g., Daams et al., 1996), the late Oligocene to earliest from the late middle Miocene of France. Although Atelerix aff. Miocene (MN1-2) insectivorans from Spanish sites have hardly depereti is larger than A. depereti across most of its range, size been studied. Insectivorans from the younger part of the early reduction seems to have taken place towards the end of its record Miocene (MN3-4) are nevertheless well-documented (Van den (Concud Cerro de la Garita, MN12), with dimensions of at least the Hoek Ostende, 2003; Van den Hoek Ostende and Furió, 2005; anterior elements approaching that of A. depereti. Atelerix steensmai Van der Meulen et al., 2012), and no erinaceines were found. nov. sp. is restricted to the older part of MN12, and is considered to Later, during the later middle Miocene, Postpalerinaceus inter- be the result of a separate development towards larger size and medius, a form well-known from France, is known to have stronger bunodonty within the Atelerix genus. The presence of entered the Madrid Basin (Paracuellos 5, 13.8-13.7 Ma; Peláez- Postpalerinaceus (cf.) vireti in the Teruel Basin is not surprising Campomanes et al., 2003). ?Amphechinus survives on the Iberian given the temporal overlap with the record of this species in NE Peninsula at least until the middle-late Miocene transition Spain (Vallès-Penedès Basin, MN9-10). With our new finds, the late (Vallès-Penedès Basin; Gibert, 1975), with rarefication taking Miocene temporal occurrence of this species has been further place towards the end of the middle Miocene (Zone G; Van der extended to MN11. Afterwards (MN12), it seems to have been Meulen et al., 2012); replaced by the similarly-sized A. steensmai nov. sp. in the Teruel  11-10 Ma: replacement of ?Amphechinus by other species, such Basin. Interestingly, a large P4 identified as cf. Postpalerinaceus was as Atelerix aff. depereti in the eastern Spanish interior as found in the MN12 site Cañada 12 in the neighboring Calatayud- evidenced in the records from the Teruel Basin (this paper) Montalbán Basin (López-Guerrera et al., 2011). Its outline hardly and Calatayud-Montalbán Basin (Cañada 6; López-Guerrera shows a posterior emargination, and contains anterior and et al., 2011), and by Postpalerinaceus vireti in the Teruel basin posterior borders that almost run parallel, features that also (this paper), the Vallès-Penedès Basin (Crusafont and De Villalta, characterize the morphology of the P4 on the holotype of 1947; Crusafont and Gibert, 1974) and the Madrid Basin (Cerro Postpalerinaceus vireti from Viladecaballs (MN9, Vallès-Penedès de Batallones; Álvarez-Sierra et al., 2017). Except perhaps for the Basin). We therefore believe that its generic assignment could be latter occurrence, Erinaceinae have not been recorded so far for narrowed down to P. cf. vireti. the interval 9.5-9.0 Ma (upper J2 and J3 Zones); Because a revision of the Vallès-Penedès Basin erinaceines is  8.5 Ma: temporary increase in species richness (Teruel Basin), not intended here, we confine ourselves to a few observations on culminating in the presence of 4 species between 8.5 and 7 Ma the material from this basin. Firstly, from the viewpoint of both (two species of Atelerix, one Postpalerinaceus, and one unidenti- size and morphology, the exclusive presence of Postpalerinaceus fied form); vireti in the interval corresponding to later MN9 and MN10 may be  7 Ma: almost complete local (Teruel Basin) and probably considered doubtful. For instance, various specimens from Can regional (Iberian Peninsula) extinction of Erinaceinae around the Ponsic (10.1 Ma), a locality supposed to contain Postpalerinaceus Tortonian-Messinian transition. The few reported Pliocene vireti (Crusafont and Gibert, 1974), are consistently smaller than occurrences (La Gloria 4, Teruel Basin, MN14; Gorafe 1 and those from Can Llobateres (9.7 Ma) and Viladecaballs (9.4 Ma). Baza, Guadix-Baza Basin, MN14; Layna, Soria region, MN15; Such smaller-sized specimens also occur in the collection of the P. Escorihuela, Teruel Basin, MN16; Sarrión 1, Sarrión depression, vireti type locality Viladecaballs itself. Generally, these smaller- MN16) refer to rare and still undescribed material which is sized teeth tend to have a less derived habitus. Furthermore, either indicated as Erinaceidae indet. or Erinaceus sp. (Adrover, several differences between P. vireti material from Viladecaballs 1974; Mein et al., 1990; Van den Hoek Ostende and Furió, 2005; and Can Llobateres are difficult to reconcile with the presence of Furió et al., 2015, 2018; Piñero et al., 2017), although the latter only one species. These differences are most evident for P3 (Fig. 4): genus assignment should be regarded as uncertain. The P3 length is exceeding width in Viladecaballs (L = 2.15 mm, W = youngest known record of pre-modern (i.e., ‘non-Erinaceus’) 1.60 mm, W/L = 1.19; Crusafont and De Villalta, 1947), whereas erinaceines is an occurrence of Postpalerinaceus (P. cf. vireti), as the situation is reversed in the P3 from Can Llobateres, for which described for La Puebla de Valverde (MN17, Sarrión depression); the W/L ratio is only 0.85 (Crusafont and Gibert, 1974).  1.5 Ma: entry of undisputed Erinaceus in the Lower Pleistocene Furthermore, the figured M1 from Can Llobateres in the latter (with various records across Spain; Furió et al., 2015, 2018). study shows an extreme morphology with emarginations at all sides (as in P4), a configuration that is difficult to match with that Before we try to explain this spatiotemporal pattern in the of the type specimens from Viladecaballs. Unfortunately, the type Iberian distribution of spiny hedgehogs, we discuss the dental material is heavily worn, hampering any detailed comparisons functional morphology of the group, because it provides a useful with regard to occlusal feratures. In our opinion, the holotype source of information on preferred diets and environments of the cranial material of Postpalerinaceus material needs to be re-studied different lineages. (involving some additional preparation) in order to substantiate presumptions (e.g., Ziegler, 2005) that the lacrimal foramen is 5.2. Dental variation and diet situated outside the orbit (invalidating potential synonymy of Postpalerinaceus with Amphechinus and assignment of the species Dental morphology reflects dental function, which in the case of intermedius to Postpalerinaceus) and that the basisphenoid groove mammals includes a broad range of actions such as cutting, is absent (invalidating potential synonymy of Postpalerinaceus with crunching, piercing, etc. of both animal and plant food. Although Mioechinus). modern spiny hedgehogs basically consume small invertebrates, In a broader Neogene perspective, the following transitions in they are known to supplement or even replace these by other Erinaceinae history can be identified for the Iberian Peninsula: materials such as small vertebrates, eggs, and plant tissue (Corbet, 1988; Reeve, 1994). The dentitions of the more advanced fossil  16 Ma: entry of Erinaceinae around the early-middle Miocene forms described here are expected to reflect this multi-purpose transition with the presence of ?Amphechinus in the Calatayud- aspect. In an attempt to functionally and ecologically interpret the Montalbán Basin (Gibert, 1975; Van der Meulen et al., 2012) and erinaceine dentition, we relied on works such as that of Lucas Valencia region (Buñol; Robles et al., 1991), and Mioechinus (2004), who focuses on tooth shape in relation to the physical 78 J.A. van Dam et al. / Geobios 61 (2020) 61–81 properties of the different food types, and on the various studies of middle part of the M2 creates a basin that functions as a ‘mortar’, Freeman on modern bats (Freeman, 1979, 1984, 1992, 1998). Part the ‘pestles’ being the cusp-shaped entoconid and hypoconid of the of this latter group has a dental morphology that is comparable to m2 (Butler, 1948). The relatively straight lingual border of the M2 that of erinaceines, although dietary specializations in bats further contributes to a widening of its central basin. have proceeded further, covering almost all major dietary types Canines typically constitute the loci of the first action in (including carnivory, frugivory, ‘nectarivory’, etc.; Freeman, 1998), fracturing and immobilizing prey. Postpalerinaceus vireti has and the dentition might also be affected by cranial constraints rounded upper canines without an anterior longitudinal crest associated with their aerial locomotion. and only a weak posterior crest (observation on holotype). By Of the genera discussed in this work, Amphechinus has the contrast, the upper C of A. steensmai nov. sp. shows distinct anterior oldest and most primitive dental pattern. The dentition of this and posterior crests. Also, the posterior fragment of a C of Atelerix genus (especially in the oldest forms) still resembles the classical aff. depereti from Prado 15C shows at least the presence a clear tribosphenic design. The presence of many small sharp ridges all centrally positioned, posterior crest. In carnivores, strong and across the entire dentition points to the ability to surmount the rounded canines are thought to prevent breakage by prey bones ‘fracture toughness’ of insect cuticles, which are stiff composites (Van Valkenburgh and Ruff, 1987; Freeman, 1992). In bats, edges with fiber layers in different directions, thereby resisting crack on upper canines are thought to have a function in directing cracks propagation (Lucas, 2004). It can therefore be assumed that these both towards incisors (including clipping off inedible parts) and forms mostly consumed small invertebrates, which they could (pre)molars (further cutting by ectolophs), with movement manipulate using their relatively large incisors (a typical feature in confined by occlusion of upper and lower canines (Freeman, Erinaceinae; Butler, 2010), pierce and immobilize with their knife- 1992). Following this argument, species such as A. steensmai nov. like canines, and chop into small pieces using their molars and sp. would be better predisposed to durophagy (with the need for premolars. In this respect, the typical W-shaped ectolophs that making cracks) than to carnivory, whereas Postpalerinaceus would characterize insectivorous upper dentitions in Insectivora and be well equipped for both types. Chiroptera may be viewed as arrays of ‘mini-carnassials’ that break Land snails, a very common group of fossils in the Teruel basin up small invertebrates, while at the same time holding the (Albesa and Robles, 2006), may be expected to form part of the diet fragments in place (Freeman, 1979). The smaller size of especially of the spiny hedgehogs in this area as well. Living forms such as the oldest representatives would, however, sets a limit to prey size Erinaceus europaeus and Atelerix algirus (e.g., on Balearic Islands; (see also Pfretschner, 1997). Atelerix (aff.) depereti, the dentition of Reeve, 1994) are known to include small, thin-shelled snails into which contains sharp blades (e.g., I3, C, and P2; Figs. 2(D, E), 7(B); their diet. With its strong molar cusps and large basins, A. steensmai Engesser, 2009: fig. 48), and many sharp, small-sized ridges, would nov. sp. would have been able to also crack the shells of some of the still largely have conformed to such an ‘invertivorous’ diet. larger and more strongly built snails relatively easily. In addition to However, the beginning molarization (relatively squared molars) small invertebrates, small reptiles such as lizards and snakes could suggests a somewhat less ‘invertivorous’ diet than for Amphe- have been part of the diet, with the processing of scales requiring chinus, implying a more ‘mainstream’ hedgehog (varied) menu distinct crunching activity. An ecological parallel can be made with consistent with the diet of living Atelerix (Reeve, 1994; Santana relatively large-sized soricids such as Amblycoptus or et al., 2010). Beremendia. For the former, a malacophagous and/or carnivore More advanced Postpalerinaceus such as P. vireti show a component has been suggested as well (Prieto and Van Dam, 2012; reduction of the P3, molarization of the fourth premolars, and a Van Dam, 2004). It might even be hypothesized that Amblycoptus development towards square-shaped M1-2 containing additional was a direct competitor for the erinaceines, as the expansion of the cusps (meta- and protoconule). Postpalerinaceus shares these former at the MN12-13 transition as recorded in the Teruel Basin features with modern Erinaceus and Atelerix, which also have these coincides with the dwindling of the latter. A counter-argument two conules present (although in modern Atelerix the protoconule against such a competitive relationship is their size difference, may be missing; Gould, 2001). Anterior elements in Postpalerina- with length and width in Amblycoptus measuring about half those ceus are stout, rounded, and well secured. The presence of large and of the late Miocene Erinaceines (thereby approaching more that of robust teeth allows for durophagous action, i.e., the crunching and older, middle Miocene erinaceines). The consumption of both fracturing of hard items such as beetle carapaces, snails and scales gastropods and coleopterans has also been suggested for the Plio- of small reptiles. The remarkable horizontally flattened wear Pleistocene soricid genus Beremendia (Furió et al., 2010). surface on the lower canine (Fig. 2(C1,2)) and premolars as The functional interpretation of rare known cranial material encountered in La Gloria 10 could suggest a predisposition for supports dental interpretations. A strong sagittal crest on the skull crushing or resisting such hard objects (bone, reptile shale?). as observed in Postpalerinaceus vireti (Crusafont and De Villalta, Perhaps P. vireti also scavenged nests or similar places where 1947) fits both a diet including harder objects as well as softer young grow up (Álvarez-Sierra et al., 2017). In addition, analogous vertebrate tissue (to slice through), because of the need for strong to carnivores, the possession of a strong carnassial pair, as temporalis muscles to attach. Amphechinus edwardsi (Viret, 1938) observed in Postpalerinaceus (large P4 blade and elongated trigonid also possesses such a crest, but its skull is narrower than in P. vireti, of m1), would allow for efficient cutting across ‘soft solids’ such as suggesting a lower degree of durophagy and/or carnassial action skin or muscle of small vertebrates. Nonetheless, there will have (as in bats; Freeman, 1984). Although the presence of a sagittal been limits to such shearing action because of size. As observed in crest and various modifications in the dentition in more advanced the slightly smaller-sized modern Erinaceus, persistent biting erinaceines such as Postpalerinaceus, Erinaceus and Atelerix seem to rather than carnassial shearing may have been used for killing point to a widening of the dietary spectrum including small vertebrate prey (Reeve, 1994). vertebrates and/or hard objects, their mandibles do not show the Molarization (with an increase of cusp area at the expense of typical modifications that characterize jaws of Carnivora. For crest area) would also have facilitated the crushing and grinding of instance, the position of the condyle is not low-positioned plant tissue. The advantage of a full functioning hypocone that (approximately the level of the tooth row) as in carnivorans, but sinks into the trigonid, duplicating the action of the protocone in relatively high-positioned, as in insectivorans. This suggests that the talonid, is obvious (Pfretschner, 1997). In our material, the position and structure of the condyle in erinaceines (and other molarization and cusp development is particularly well developed insectivores) still allows for extra lateral movement in addition to in A. steensmai nov. sp., where the absence of structures in the the vertical movement required for shearing and/or crushing. On J.A. van Dam et al. / Geobios 61 (2020) 61–81 79 the other hand, the possession of a thicker, ventrally slightly seems to have affected the erinaceines there as well (no record rounded jaw in advanced forms such as Postpalerinaceus vireti and known after 9.4 Ma), their initial presence in this region requires an Erinaceus europaeus would be consistent with the inclusion of explanation. Perhaps ‘A.’ golpae, a species only known from this harder items in the diet (beetles, small bone, shell, lizard shale), area, differed from other erinaceines in preferring wetter or more whereas thin jaws (Amphechinus) would point to the consumption closed environments. On the other hand, the occurrence of of softer invertebrate food such as small insects, grubs, worms, etc. Postpalerinaceus vireti in both wetter sites such as Can Llobateres that could be probed and picked more easily using a long snout (Crusafont and Gibert, 1974), as well as in drier sites in the Teruel (Viret, 1938; Butler, 1948; see also Freeman, 1984, for a discussion region points to a broad climatic tolerance, which could be linked on bats, and Van Dam et al., 2011, for a discussion on shrews). to its diet, which supposedly includes a carnivorous component. Although biting force is expected to correspond to a larger The presence of yet another erinaceine species (‘P.’ cingulatus; coronoid process for the temporal muscles to attach (Turnbull, Ziegler, 2005b) in the wetter central European site Rudabanya 1970), the inclination of the coronoid process with regard to the (Bernor et al., 2004; Van Dam and Utescher, 2016) further supports tooth row is perhaps not necessarily related to diet, as it might also the notion of ecological differentiation within Miocene European depend on other, oppositely working evolutionary constraints such spiny hedgehogs. as tooth reduction. The inclination of the coronoid process also After a short absence between 9.5 and 8.8 Ma in the Teruel cannot be considered to be a very reliable taxonomic feature. For Basin, Erinaceinae presence resumes in Zone J4 (basal Turolian). example, whereas the (anterior edge of the) coronoid process is Maximum erinaceine diversity is attained in Zone K (MN11, ca. inclined on the holotype of P. vireti (Crusafont and De Villalta, 9-7.5 Ma), a period that is still relatively dry (Van Dam, 2006). The 1947: fig. 3), it is perpendicularly placed in the Cerro de Batallones coexistence of various species during this interval could therefore material that supposedly belongs to the same species (Álvarez- be explained by a widening of dietary spectra (Postpalerinaceus, A. Sierra et al., 2017: fig. 1). A similar ambiguity characterizes the steensmai), including harder and larger invertebrates such as mandible shape in Amphechinus, where the vertical ramus is terrestrial gastropods, or small vertebrates with hard parts such as almost vertical in A. edwardsi as described by Viret (1938: fig. 4), lizard scales. Alternatively, the occurrence of the latter species but positioned at an oblique angle (70-808) in both A. edwardsi and could be related to a temporary increase of humidity (early part of A. arvernensis as described by Butler (1948: fig. 22). Based on an MN12; Van Dam and Weltje, 1999; van Dam, 2006). inclined ascending ramus in the mandible, Ziegler (2005b) The pattern of erinaceine diversification in the rest of Europe preferred assignment of middle-sized erinaceine material from resembles that on the Iberian Peninsula, except for a number of Hungary (Rudabanya, MN10) to Postpalerinaceus (P. cingulatus) significantly older occurrences (predominantly Amphechinus) as rather than to the similarly-sized Atelerix depereti, the vertical evidenced by the late Oligocene-early Miocene records of France ramus of which is perpendicularly placed (Depéret, 1887). Given and Germany. During the middle Miocene Erinaceinae constituted the intra-generic variability of this character, the question of the a rare but consistent element of well-sampled sites in Central and generic assignment of ‘P.’ cingulatus could be considered to be still Eastern Europe (Hı́r et al., 2016; Prieto et al., 2015; Ziegler, 2005). open. Similarly, late middle Miocene erinaceines are well documented for France thanks to sites such as Sansan and La Grive. The late 5.3. European Erinaceinae and Neogene climate change middle Miocene (MN7-8), an interval during which spiny hedgehogs were relatively rare in Spain, seems to have been a According to terrestrial paleoclimate studies based on rodents, successful period for the group in France, Central Europe and the entry of Erinaceinae (Amphechinus) in Spain at 16 Ma would Anatolia, with presences of Postpalerinaceus, Atelerix and Mioe- coincide with either a transition to drier conditions (Van der chinus (Engesser, 1980, 2009; Mein and Ginsburg, 2002). By Meulen and Daams, 1992) or a generally dry period (Van Dam et al., contrast, the late Miocene record in Central Europe seems to be 2006: suppl. fig. 7). New and updated precipitation estimates also poorer than on the Iberian Peninsula, although a lower number of point to relatively dry and open conditions between 18-14 Ma (Van sites in the former region may cause some bias. The upper Miocene Dam et al., ongoing work). The subsequent rarefication of? record of France is characterized by frequent occurrences (Mein, Amphechinus during zone G (Van den Hoek Ostende and Furió, 1999). 2005; Furió et al., 2018) would be coeval with an increase in It cannot be established with certainty whether the latest precipitation. In line with the hypothesis of a preference of Miocene and Pliocene low in erinaceine diversity in the Iberian erinaceines for more open and drier environments, renewed drying Peninsula reflects a pan-European phenomenon or not. Occasional from 10 Ma corresponds to a rise of erinaceine occurrences occurrences are noted for the Pliocene of Central-Eastern Europe (Atelerix aff. depereti and Postpalerinaceus cf. vireti) in the Teruel (e.g., Sulimski, 1962; Rzebik-Kowalska, 2002), with first entry of Basin – note that at the same time relatively wet-adapted groups the modern genus Erinaceus (E. samsonowiczi) supposedly having such soricids and glirids disappear (Van Dam et al., 2001). A taken place around the Miocene-Pliocene transition (Doukas et al., prolonged environmental deterioration seems finally also to have 1995). However, it is only during the Pleistocene that this genus affected the Erinaceinae, however, which temporarily disappear starts to become relatively common. from the area just before the time of the J2-J3 transition: despite New compilations of sea surface temperatures for the the presence of well-documented upper J2 and J3 faunas such as Mediterranean and Northern Atlantic show an episode of very Peralejos C, La Roma 1-2, and Ması́a de La Roma 604 (Van de strong cooling around 7-6 Ma (Herbert et al., 2016). The sudden Weerd, 1976; Van Dam et al., 2001; Alcalá et al., 2005), not a single drops in temperature must have strongly affected environments in occurrence of Erinaceinae has been recorded so far for the interval Europe and the Mediterranean region, areas that had remained 9.5-9.0 Ma. In fact, small-mammal diversity as a whole reaches a very warm until then. It is tempting to link the strong decrease in low, a development which is thought to correspond to the Erinaceinae in the Mediterranean and Europe from the Messinian establishment of a truly dry summer (Van Dam, 2006). onwards to effects of this ‘Messinian glaciation’, perhaps in The presence of various erinaceines during the early late combination with renewed drying (van Dam, 2006). Although Miocene MN9-10 in northeastern Spain (Vallès-Penedès Basin) speculative, tectonic factors could have contributed to the demise seems to be at odds with the pattern described above, as conditions of spiny hedgehogs in this period as well, given the evidence for in this region were relatively wet at this time (Van Dam, 2006; accelerated uplift and basin formation across the Mediterranean Casanovas-Vilar and Agustı́, 2007). Although the ‘Vallesian crisis’ region, including Iberia (Jolivet et al., 2006). The resulting 80 J.A. van Dam et al. / Geobios 61 (2020) 61–81 topographic barriers could have contributed to preventing spiny P.S., Utescher, T., Van Dam, J.A., Werdelin, L., Ziegler, R., 2004. Recent advances on multidisciplinary research at Rudabánya, late Miocene (MN9), Hungary: a hedgehogs from returning to formerly inhabited areas. compendium. Palaeontographia Italica 89, 3–36. The fact that Atelerix species are currently restricted to Africa is Butler, P.M., 1948. On the evolution of the skull and teeth in the Erinaceidae, with consistent with an important role of temperature on its past special reference to fossil material in the British museum. In: Proceedings of the Zoological Society of London. 118, pp. 446–500. distribution. More generally, the combination of a preference for Butler, P.M., 1956. Erinaceidae from the Miocene of East Africa. Fossil Mammals of high temperatures and a tolerance towards aridity would fit the Africa 11, 1–75. (semi-) desertic landscapes in which the majority of modern Butler, P.M., 2010. Ch. 29, Neogene Insectivora. In: Werdelin, L., Sanders, W.J. (Eds.), Cenozoic Mammals of Africa. California University Press, Berkeley, pp. 573– Erinaceinae species (Hemiechinus, Paraechinus, Atelerix) are cur- 580. rently thriving. As also suggested by Corbet (1988), a parallel may Casanovas-Vilar, I., Agustı́, J., 2007. Ecogeographical stability and climate forcing in be drawn between the biogeographic histories of Atelerix and the the late Miocene (Vallesian) rodent record of Spain. Palaeogeography, Palaeo- climatolology, Palaeoecology 248, 169–189. ground squirrel Atlantoxerus, a common sciurid in the Iberian Corbet, G.B., 1988. The family Erinaceidae: a synthesis of its taxonomy, phylogeny, Peninsula during the Neogene until the Pliocene (García-Alix et al., ecology and zoogeography. Mammal Review 18, 117–172. 2007). Like the mainland distribution area of the northernmost Crochet, J.Y., Heintz, E., 1971. Insectivora (Mammalia) de la faune villafranchienne living species of Atelerix (A. algirus), whose African presence can de La Puebla de Valverde (prov. Teruel, Espagne). Bulletin du Museum National d’Histoire Naturelle 42, 776–779. probably be traced back to the Plio-Pleistocene boundary (Butler, Crusafont, M., Gibert, J., 1974. Nuevos datos sobre el genero Postpalerinaceus del 2010), the geographic range of the only living species of Vallesiense. Acta Geologica Hispanica 9, 1–3. Atlantoxerus (A. getulus) is restricted to northernmost Africa. Crusafont, M., De Villalta, J.F., 1947. Sur un nouveau Palerinaceus du Pontien d’Espagne. Eclogae Geologica Helvetica 40, 320–333. Crusafont, M., De Villalta, J.F., Truyols, Y., 1955. El Burdigaliense continental de la Acknowledgements cuenca del Vallès-Penedès. Memorias y Communicaciones del Instituto Geo- logico, Barcelona1–273, 12. Daams, R., Álvarez Sierra, M.A., Van der Meulen, A.J., Peláez-Campomanes, P., 1996. We are very grateful to Marta Palmero for making the specimen Paleoecology and paleoclimatology of micromammal faunas from upper Oligo- drawings. We thank Marc Furió for providing us with material of cene-lower Miocene sediments in the Loranca Basin, Province of Cuenca, Spain. Erinaceus europaeus and his comments on the manuscript. We In: Friend, P.F., Dabrio, C.J. (Eds.), Tertiary Basins of Spain: the stratigraphic record of crustal kinematics. Cambridge University Press, pp. 295–299. thank Marc Furió and Isaac Casanovas for their help with scanning De Jong, F., 1988. Insectivora from the upper Aragonian and the lower Vallesian of specimens at the Department of Geology of the Universidad the Daroca-Villafeliche area in the Calatayud-Teruel Basin (Spain). In: Freuden- Autònoma de Barcelona, and Laura Celià and Marta March for thal, M. (Ed.), Biostratigraphy and paleoecology of the Neogene micromam- malian faunas from the Calatayud-Teruel Basin (Spain). Scripta Geologica consulting the ICP collection. We thank Eduardo Espilez, Emman- Special Issue. pp. 253–286. uel Robert and Wilma Wessels for their help in studying material Depéret, C., 1887. Recherches sur la succession des faunes de vertébrés miocènes de from the Teruel, Lyon and Utrecht collections, respectively. Jérôme la Vallée du Rhône. Archives du Muséum d’Histoire naturelle de Lyon 4, 1–269. Doukas, C.S., Van den Hoek Ostende, L.W., Theocharopoulos, C.D., Reumer, J.W.F., Prieto and Florentin Cailleux are thanked for their constructive 1995. Insectivora. In: Schmidt-Kittler, N. (Ed.), The vertebrate locality Mara- reviews. We also greatly acknowledge the past efforts of Etienne mena (Macedonia, Greece) at the Turolian-Ruscinian boundary (Neogene). Moissenet and Rafael Adrover in discovering, excavating and Friedrich Pfeil, Munich, pp. 43–64. Engesser, B., 1980. Insectivora und Chiroptera (Mammalia) aus dem Neogen der studying many of the fossil sites in the Teruel Basin. Research Türkei. Schweizerische Paläontologische Abhandlungen 102, 45–149. (J.A.v.D.) was funded by Generalitat de Catalunya (CERCA Engesser, B., 2009. The insectivores (Mammalia) from Sansan (middle Miocene, Program), and facilitated by Department of Earth Sciences, Utrecht south-western France). Schweizerische Paläontologische Abhandlungen 128, 1–91. University. Research (L.A.) of Group E04_17R FOCONTUR was Freeman, P.W., 1979. Specialized insectivory: beetle-eating and moth-eating financed by Departamento de Innovación, Investigación y Uni- molossid bats. Journal of Mammalogy 60, 467–479. versidad (Gobierno de Aragón) and FEDER Aragón 2014-2020 Freeman, P.W., 1984. Functional cranial analysis of large animalivorous bats ("Construyendo Europa desde Aragón’’). (Microchiroptera). Biological Journal of the Linnean Society 21, 387–408. Freeman, P.W., 1992. Canine teeth of bats (Microchiroptera): size, shape and role in crack propagation. Biological Journal of the Linnaean Society 45, 97–115. References Freeman, P.W., 1998. Form, function and evolution in skulls and teeth of bats. In: Kunz, T.H., Racey, P.A. (Eds.), Bat Biology and Conservation. Smithonian Adrover, R., 1974. Un relleno kárstico plio-pleistocénico en el Cerro de los Espejos en Institution Press, Washington, DC, pp. 140–156. Sarrión (Prov. de Teruel, España) (Nota preliminar). Acta Geológica Hispánica 9, Frost, D.R., Wozencraft, W.C., Hoffmann, R.S., 1991. Phylogenetic relationships of 142–143. hedgehogs and gymnures (Mammalia: Insectivora: Erinaceidae). Smithsonian Adrover, R., 1986. Nuevas faunas de roedores en el Mio-Plioceno continental de la Contributions to Zoology 518, 1–69. región de Teruel (España). In: Interés bioestratigrafica y paleoecológico. Ins- Furió, M., Gibert, L., Ferràndez, C., Sevilla, P., 2015. The insectivores (Soricidae, tituto de Estudios Turolenses de la Excma. Diputacion Provincial de Teruel, Erinaceidae; Eulipotyphla; Mammalia) from Cueva Victoria (Early Pleistocene, Teruel. Murcia, Spain). Neues Jahrbuch für Geologie und Paläontologie 275, 151–161. Ai, H.S., He, K., Chen, Z.Z., Li, J.Q., Wan, T., Li, Q., Nie, W.H., Wang, J.H., Su, W.T., Jiang, Furió, M., Van den Hoek Ostende, L.W., Agustı́, J., Minwer-Barakat, R., 2018. X.L., 2018. Taxonomic revision of the genus Mesechinus (Mammalia: Erinacei- Evolution of the insectivore assemblages (Eulipotyphla, Mammalia) in Spain dae) with description of a new species. Zoological Research 39, 335–347. and their relation with Neogene and Quaternary climatic changes. Ecosistemas Albesa, J., Robles, F., 2006. Síntesis de los estudios sobre moluscos continentales 27, 38–51. neógenos del sector septentrional de la Depresión de Teruel: período 1775- Furió, M., Agustí, J., Mouskhelishvili, A., Sanisidro, O., Santos-Cubedo, A., 2010. 1998. Estudios Geológicos 62, 183–198. The paleobiology of the extinct venomous shrew Beremendia (Soricidae, Alcalá, L., Sesé, C., Herráez, E., Adrover, R., 1991. Mamı́feros del Turoliense inferior Insectivora, Mammalia) in relation to the geology and paleoenvironment of de Puente Minero (Teruel, España). Boletı́n de la Real Sociedad Española de Dmanisi (early Pleistocene, Georgia). Journal of Vertebrate Paleontology 30, Historia Natural (Sección Geológica) 86, 205–251. 928–942. Alcalá, L., Van Dam, J.A., Luque, L., Montoya, P., Abella, J., 2005. Nuevos mamı́feros García-Alix, A., Minwer-Barakat, R., Martín Suárez, E., Freudenthal, M., 2007. New vallesienses en Ması́a de La Roma (Cuenca de Teruel). Geogaceta 37, 199–202. data on mio-pliocene Sciuridae (Rodentia, Mammalia) from Southern Spain. Álvarez-Sierra, M.A., Garcı́a-Paredes, I., Gómez, A.R., Hernández-Balları́n, V., Van Comptes Rendus Palevol 6, 269–279. den Hoek Ostende, L.W., López-Antoñanzas, R., López-Guerrero, P., Oliver, A., Gibert, J., 1975. New insectivores from the miocene of Spain; I and II. In: Pro- Peláez-Campomanes, P., 2017. Los micromamı́feros del Cerro de los Batallones. ceedings of the Koninklijke Nederlandse Akademie Van Wetenschappen, Series In: Morales, J., et al. (Eds.), La colina de los Tigres Dientes de Sable: Los B. 78, pp. 108–133. yacimientos miocenos del Cerro de los Batallones. Torrejón de Velasco, Gould, G.C., 2001. The phylogenetic resolving power of discrete dental morphology Madrid, pp. 519–529. among extant hedgehogs and the implications for their fossil record. American Andrews, P., Cook, J., 1990. Owls, caves and fossils. In: Natural History Museum Museum Novitates 3340, 1–52. Publications. Natural History Museum, London. Herbert, T.D., Lawrence, K.T., Tzanova, A., Peterson, L.C., Caballero-Gill, R., Kelly, C.S., Bernor, R.L., Kordos, L., Rook, L., Agustı́, J., Andrews, P., Armour-Chelu, M., Begun, 2016. Late Miocene global cooling and the rise of modern ecosystems. Nature D.R., Cameron, D.W., Damuth, J., Daxner-Höck, G., De Bonis, L., Fejfar, O., Geoscience 9, 843–849. Fessaha, N., Fortelius, M., Franzen, J., Gasparik, M., Gentry, A., Heissig, K., Hilgen, F.J., Lourens, L.J., Van Dam, J.A., 2012. The Neogene period. In: Gradstein, Hernyak, N., Kaiser, T., Koufos, G.D., Krolopp, E., Jánossy, D., Llenas, M., Meszáros, F.M., Ogg, J.G., Ogg, G. (Eds.), A Geological Time Scale 2012. Elsevier, L., Müller, P., Renne, P., Ročk, Z., Sen, S., Scott, R., Szyndlar, Z., Topál, G., Ungar, Amsterdam, pp. 923–978. J.A. van Dam et al. / Geobios 61 (2020) 61–81 81 Hı́r, J., Venczel, M., Codrea, V., Angelone, C., Van den Hoek Ostende, L.W., Kirscher, U., Van Dam, J.A., 2004. Anourosoricini (Mammalia: Soricidae) from the Mediterranean Prieto, J., 2016. Badenian and Sarmatian s.str. from the Carpathian area: region: a pre-Quaternary example of recurrent climate-controlled North-South overview and ongoing research on Hungarian and Romanian small vertebrate range shifting. Journal of Paleontology 78, 741–764. evolution. Comptes Rendus Palevol 15, 863–875. Van Dam, J.A., 2006. Geographic and temporal patterns in the late Neogene (12-3 Jolivet, L., Augier, R., Robin, C., Suc, J.-P., Jean Marie Rouchy, J.M., 2006. Lithospheric- Ma) aridification of Europe: the use of small mammals paleoprecipitation scale geodynamic context of the Messinian salinity crisis. Sedimentary Geology proxies. Palaeogeography, Palaeoclimatology, Palaeoecology 238, 190–218. 188–189, 9–33. Van Dam, J.A., Utescher, T., 2016. Plant- and micromammal-based paleoprecipita- López-Guerrero, P., Garcı́a-Paredes, I., Van den Hoek Ostende, L.W., Van Dam, J.A., tion proxies: comparing results of the Coexistence and Climate-Diversity Álvarez-Sierra, M.Á., Hernández-Balları́n, V., Van der Meulen, A.J., Oliver, A., Approach. Palaeogeography, Palaeoclimatology, Palaeoecology 443, 18–33. Peláez-Campomanes, P., 2011. Cañada: a new micromammal succession from Van Dam, J.A., Weltje, G.J., 1999. Reconstruction of the Late Miocene climate of Spain the lower Vallesian of the Daroca area (Calatayud-Montalbán basin, Spain). using rodent palaeocommunity successions; an application of end-member Estudios Geológicos 67, 443–453. modelling. Palaeogeography, Palaeoclimatology, Palaeoecology 151, 267– Lucas, P.W., 2004. Dental Functional Morphology. How Teeth Work. Cambridge 305. University, Cambridge. Van Dam, J.A., Alcalá, L., Alonzo Zarza, A., Calvo, J.P., Garcés, M., Krijgsman, W., 2001. Mein, P., 1999. The small mammal vallesian and turolian succession of France. In: The Upper Miocene mammal record from the Teruel-Alfambra region (Spain). Agustı́, J., Rook, L., Andrews, P. (Eds.), Evolution of Neogene terrestrial eco- The MN system and continental Stage/Age concepts discussed. Journal of systems in Europe. Cambridge University Press, pp. 140–164. Vertebrate Paleontology 21, 367–385. Mein, P., Ginsburg, L., 2002. Sur l’âge relatif des différents dépôts karstiques Van Dam, J.A., Abdul Aziz, H., Álvarez Sierra, M.Á., Hilgen, F.J., Van den Hoek miocènes de La Grive-Saint-Alban (Isère). Cahiers scientifiques – Muséum Ostende, L.W., Lourens, L.J., Mein, P., Van der Meulen, A.J., Peláez-Campomanes, d’Histoire naturelle, Lyon 2, 7–47. P., 2006. Long-period astronomical forcing of mammal turnover. Nature 443, Mein, P., Martı́n-Suárez, E., 1994. Galerix iberica sp. nov. (Erinaceidae, Insectivora, 687–691. Mammalia) from the late Miocene and early Pliocene of the Iberian Peninsula. Van Dam, J.A., Krijgsman, W., Abels, H.A., Álvarez-Sierra, M.Á., Garcı́a-Paredes, I., Geobios 26, 723–730. López-Guerrero, P., Peláez-Campomanes, P., Ventra, D., 2014. Updated chronol- Mein, P., Moissenet, E., Adrover, R., 1990. Biostratigraphie du Neogene supérieur du ogy for middle to late Miocene mammal sites of the Daroca area (Calatayud- bassin de Teruel. Paleontologia i Evolució 23, 121–139. Montalbán Basin, Spain). Geobios 47, 325–334. Opdyke, N., Mein, P., Lindsay, E., Pérez-Gonzáles, A., Moissenet, E., Norton, V.L., Van de Weerd, A., 1976. Rodent faunas of the Mio-Pliocene continental sediments of 1997. Continental deposits, magnetostratigraphy and vertebrate paleontology, the Teruel-Alfambra region, Spain. Utrecht Micropaleontological Bulletins late Neogene of Eastern Spain. Palaeogeography, Palaeoclimatolology, Palaeoe- Special Publication, 2, pp. 1–217. cology 133, 129–148. Van den Hoek Ostende, L.W., 2003. Insectivores (Erinaceomorpha, Soricomorpha, Peláez-Campomanes, P., Morales, J., Álvarez Sierra, M.Á., Azanza, B., Fraile, S., Garcı́a Mammalia) from the Ramblian of the Daroca-Calamocha area. In: López-Mar- Paredes, I., Hernández Fernández, M., Herráez, E., Nieto, M., Pérez, B., Quiralte, tı́nez, N., Peláez-Campomanes, P., Hernán- dez Fernández, M. (Eds.), En torno a V., Salesa, M.J., Sánchez, I.M., Soria, D., 2003. Updated biochronology of the fósiles de mamı́feros: datación, evolución y paleoambiente. Coloquios de Miocene mammal faunas from the Madrid basin (Spain). Deinsea 10, 431–441. Paleontologı́a Volumen extraordinario. 1, pp. 281–310. Pfretschner, H.-U., 1997. Anpassung an der Zahnmorfologie an die Ernärung bei Van der Meulen, L.W., Furió, M., 2005. Spain. In: Vanden HoekOstende, L.W., rezenten und fossilen Säugetieren. In: Alt, K.W., Türp, J.C. (Eds.), Die Evolution Doukas, C.S., Reumer, J.W.F. (Eds.), The Fossil Record of the Eurasian Neogene Der Zähne - Phylogenie, Ontogenie, Variation. Quintessenz Verlags-GmbH, Insectivores (Erinaceomorpha, Soricomorpha, Mammalia), Part I. Scripta Geo- Berlin, pp. 423–447. logica, Special Issue 5. pp. 148–284. Piñero, P., Agustı́, J., Oms, O., Blain, H.-A., Laplana, C., Ros-Montoya, S., Martı́nez- Van der Meulen, A.J., Garcı́a-Paredes, I., Álvarez-Sierra, M.A., Van den Hoek Ostende, Navarro, B., 2017. Rodents from Baza-1 (Guadix-Baza Basin, Southeast Spain): L.W., Hordijk, K., Oliver, A., Peláez-Campomanes, P., 2012. Updated Aragonian filling the gap of the early Pliocene succession in the Betics. Journal of Verte- biostratigraphy: small mammal distribution and its implications for the Mio- brate Paleontology 37 e1338294. cene European chronology. Geologica Acta 10, 159–179. Prieto, J., Van Dam, J.A., 2012. Primitive Anourosoricini and Allosoricinae from the Van Der Meulen, A.J., Daams, R., 1992. Evolution of early-middle Miocene rodent Miocene of Germany. Geobios 45, 581–589. faunas in relation to long-term palaeoenvironmental changes. Palaeogeogra- Prieto, J., Van den Hoek Ostende, L.W., Hı́r, J., Kordos, L., 2015. The middle miocene phy, Palaeoclimatology, Palaeoecology 93, 227–253. insectivores from Hasznos (Hungary, Nógrád County). Palaeobiodiversity and Van Valkenburgh, B., Ruff, C.B., 1987. Canine tooth strength and killing behavior in Palaeoenvironments 95, 431–451. large carnivores. Journal of Zoology, London 212, 379–397. Qiu, Z., 1996. Middle Miocene Micromammalian Fauna From Tunggur, Nei Mongol. Viret, J., 1938. Étude sur quelques érinacéidés fossiles spécialement sur le genre Science Press Beijing, Beijing. Palaerinaceus. Travaux du laboratoire de Géologie de la Faculté des Sciences de Reeve, N., 1994. Hedgehogs. T. & A.D. Poyser, Ltd., London. . Lyon 34, 1–32. Robles, F., Belinchón, M., Garcı́a-Flor, J., Morales, J., 1991. El Neógeno continental de Ziegler, R., 1999. Order insectivora. In: Roessner, G.E., Heissig, K., Fahlbusch, V. Buñol y del Valle del Rı́o Cabriel. In: De Renzi, M., Márquez-Aliaga, A., Usera, J. (Eds.), The Miocene Land Mammals of Europe. Federal Republic of Germany, (Eds.), V Jornadas de Paleontologı́a. El estudio de la forma orgánica y sus Pfeil, Munich, pp. 53–74. consecuencias en paleontologı́a sistemática, paleoecologı́a y paleontologı́a Ziegler, R., 2005a. Erinaceidae and Dimylidae (Lipotyphla) from the late Middle evolutiva. pp. 205–215. Miocene of South Germany. Senckenbergiana Lethaea 85, 131–152. Rzebik-Kowalska, B., 2002. The Pliocene and early Pleistocene Lipotyphla (Insecti- Ziegler, R., 2005b. The insectivores (Erinaceomorpha and Soricomorpha, Mamma- vora, Mammalia) from Romania. Acta Zoologica Cracoviensia 45, 251–281. lia) from the Late Miocene hominoid locality Rudabánya. Palaeontographia Santana, E.M., Jantz, H.E., Best, T.L., 2010. Atelerix albiventris (Erinaceomorpha: Italica 90, 53–81. Erinaceidae). Mammalian Species 42 (857), 99–110. Ziegler, R., Dahlmann, T., Storch, G., 2007. Marsupialia, Erinaceomorpha and Sori- Sulimski, A., 1962. Supplementary studies on the insectivores from Weże 1 comorpha (Mammalia). In: Daxner-Höck, G. (Ed.), Oligocene-Miocene Verte- (Poland). Acta Palaeontologica Polonica 7, 441–498. brates from the Valley of Lakes (Central Mongolia): Morphology, phylogenetic Turnbull, W.D., 1970. Mammalian Masticatory Apparatus. 18, Fieldiana: Geology, and stratigraphic implications. Annalen des Naturhistorischen Museums in pp. 147–356. Wien 108A. pp. 53–164.