Stuck in the middle. A geographical appraisal of the oldest insectivores - and a marsupial - from the Vallès-Penedès Basin (early Miocene, Catalonia, Spain) Author Van Den Hoek Ostende, Lars W. Naturalis Biodiversity Center, Darwinweg 2, 2333 CR Leiden (The Netherlands) lars.vandenhoekostende@naturalis.nl Author Casanovas-Vilar, Isaac Institut Català de Paleontologia Miquel Crusafont. Edifici Z, c / de les Columnes, Campus de la UAB, 08193 Cerdanyola del Vallès, Barcelona (Spain) isaac.casanovas@icp.cat Author Furió, Marc Institut Català de Paleontologia Miquel Crusafont. Edifici Z, c / de les Columnes, Campus de la UAB, 08193 Cerdanyola del Vallès, Barcelona (Spain) and Departament de Geologia, Universitat Autònoma de Barcelona (08193), Cerdanyola del Vallès, Barcelona (Spain) marc.furio@uab.cat text Comptes Rendus Palevol 2020 2020-07-22 19 1 1 25 http://dx.doi.org/10.5852/cr-palevol2020v19a1 journal article 10.5852/cr-palevol2020v19a1 1777-571X 14203901 urn:lsid:zoobank.org:pub:9B7BA215-F7E3-4AF0-B764-43F0DD3EADB3 Amphiperatherium frequens (von Meyer, 1846) ( Fig. 3 A-G) MATERIAL AND MEASUREMENTS (in cm). — Spain . El Canyet , IPS19361 , mandible with m1-m4, partially encaged in sediment ; El Palau 3B, 1 m 2 sin., IPS86230 , 2.17 × 1.17 × 1.26; 1 m3 dext., IPS86231 , 2.11 × 1.18 × 1.12 ; Can Martí Vell II, M3 dext., IPS102010 , LvK 2.16 × WvK 2.36; LCr 2.12 × WCr 2.75; 1 m1 sin., IPS86262 , 2.00 × 1.00 × 1.12; 1 m3 dext., IPS86260 , 2.16 × 1.18 × 1.14 ; Can Martí Vell I, 1 M4 dext., IPS86243 , 1.10 × 2.30 ; Les Cases de la Valenciana 1, 1 M3 sin., IPS86622 , – × –, 2 fragmented upper molars, IPS86615 , IPS86623 ; Turó de les Forques , 1 M2 dext., IPS86822 , – × –; p2? dext., IPS86823 , – × 0.75, m3 sin., IPS85484 , 2.19 × 1.22 × 1.16 ; Sant Andreu de la Barca 1, dP3 dext., IPS86180 , LvK 1.63 × WvK 1.36; LCr 1.37 × WCr 1.68 . TABLE 1. — Localities studied with their corresponding stage and zone and their relative position (when possible). Abbreviations: srm , stratigraphic refinement method used to determine the relative position; L , lithostratigraphy; M , magnetostratigraphy; B , biostratigraphy.

Localities	Stage and zone	Relative position	Srm
La Vinya Vella	MN4	very high in succession	B+L
El Palau 3B	MN4?	–	B
El Canyet	MN4?	–	B
Les Escletxes del Papiol	Aragonian zone C	subzone Cb? younger than Vilobí	B+L
Sant Mamet	Aragonian zone C	subzone Cb? younger than Vilobí	B+L+M
Vilobí del Penedès	Aragonian zone C	subzone Cb? younger than Els Casots	B+L
Els Casots	Aragonian zone C	subzone Cb?	B+L
Can Julià	Aragonian zone C	–	B+L
Can Martí Vell 1-3	Aragonian zone C	subzone Ca?	B+L
Les Cases de la Valenciana	Aragonian zone C	subzone Ca?	B+L
El Molí de Can Calopa	Ramblian zone A	younger than Costablanca	B+L
La Costablanca 1-2	Ramblian zone A	–	B+L
Turó de les Forques	Ramblian zone A	below Costablanca	B+L
Sant Andreu de la Barca 1-3	Ramblian zone A	–	B

DESCRIPTION dP3 The occlusal outline is triangular.The parastylar region is broken in the only sufficiently complete specimen. The metaconus is the highest cusp. The paraconus is somewhat lower than the metaconus.The antimetaconus is almost as high as the paraconus, whereas the antiparaconus is reduced to a tiny elongated bulge between both.The protoconulus (= “paraconule” in Furió et al. 2012 ) is discernible as a small elevation of the preprotocrista. The preprotocrista connects with the base of the paraconus by means of a faint ridge. The protoconus is the most lingual cusp of the tooth, and the highest of the lingual lobe. The postprotocrista descends irregularly, so the most posterior inflexion can be interpreted as a weak metaconulus. The metaconulus is separated from the base of the metaconus by a notch. There is no cingulum surrounding the base of the tooth. M2 The only specimen found is incomplete, as the metaconus complex is missing.The parastyle is partially fragmented, but is clearly reduced. The paraconus is the highest cusp. The protoconus is much lower.The preprotocrista does not reach the base of the paraconus, as it is interrupted by a well-marked notch. The protoconulus and metaconulus are rather inconspicuous. The preserved part of the postpotocrista is a continuous ridge at the posterolingual margin of the tooth. M3 The metaconus is the highest cusp. The protoconus is clearly lower than the metaconus and only somewhat lower than the paraconus. The parastyle is formed by a hardly protruding, triangular area.The metaconulus is very small. The antiparaconus is pronounced; the antimetaconus is tiny, only represented by a small pimple placed close to an occlusal inflexion of the labial margin. The labial margin is concave. M4 The occlusal outline is triangular, and about twice as long in the labio-lingual than in the meso-distal direction. The paraconus is a sharp elevated cusp in a central position with two crests at an angle of 90º. One of the crests is completely parallel to the anterior margin, thus connecting with the parastyle. The other one is completely perpendicular to the anterior margin thus reaching the posterior corner and the purported metaconus.The protoconus is placed at the lingual corner. The basal cingulum covers the anterior margin, running from the protoconus to the base of the parastyle. The postprotocrista makes a soft ‘S’ curvature ending in a faint and slightly elevated metaconulus. Mandible The specimen from El Canyet only preserves the horizontal ramus. The lingual side is strongly dissolved and the labial one is covered by a hard sedimentary crust, so no significant mandibular character can be observed. p2 (?) The only specimen is damaged, lacking the posterior extension. This tooth is monocuspid and two-rooted. The two roots are very close to each other. The main cusp is trifaced. In lingual view, the crown is triangular. The lingual face is flat. The labial face is convex. In labial view, the anterior part of the base elevates in anterior direction reaching half of the total height of the main cuspid. m1 The trigonid has an elongated aspect, with a paralophid clearly longer than the protolophid. The protoconid is the highest cusp. The metaconid is somewhat lower, and the paraconid is lower than the metaconid. The talonid is clearly wider and shorter than the trigonid. The hypoconid is very large. The oblique cristid ends just lingually of the base of the protoconid; the hypolophid is curved and separated from the entoconid by a wide post-entoconid valley. The entoconid is small; the entocristid slopes down steeply. The anterior cingulid is strong; the posterior cingulum is well developed. A small portion of labial cingulid borders the re-entrant valley. m2 The relative height of the trigonid cuspids is the same as in m1. However, the proportion between length and width is different, with a paralophid only slightly longer than the protolophid. The talonid and trigonid have similar lengths and widths. The hypoconid is the highest cuspid of the talonid. The entoconid is well developed, but it is shorter than the hypoconid. The hypoconulid is tilted distally, thus providing a ‘twisted’ aspect to the postcristid. The hypoconulid and the entoconid are separated by a wide notch. There is a well-developed cingulid below the paralophid. The oblique cristid ends more labially than lingually, at about one third of the protolophid length. m3 The talonid is shorter and somewhat narrower than the trigonid. The hypoconulid stands out less than in m2. The rest of the characters are quite the same as in m2. REMARKS Occurrences of the herpetotheriid metatherian Amphiperatherium are quite frequent in the early Miocene of central Europe ( von Koenigswald 1970 ; Ziegler & Fahlbusch 1986 ; Ziegler 1990a , 1999 ; Klietmann et al. 2014a ). By contrast, this genus is rather rare in Spain , though its presence in the Iberian Peninsula was noted in Buñol ( Robles et al. 1991 ), Mas d’Antolino and Barranc de Campisano ( Furió et al. 2012 ) and Montalvos 2 ( Hordijk et al. 2015 ). Moreover, fossils were found in the Ramblian and early Aragonian sections of the Calatayud-Montalban basin (LHO, pers. obs.), but these collections are currently lost. The material from Las Cases de Valenciana was already described by Jovells-Vaqué et al. (2018). Species identification based on isolated molars is difficult, but the only European marsupial known from the Miocene is Amphiperatherium frequens von Meyer, 1846 . As the material generally fits this species both metrically and morphologically, its identification on the species level seems to be safe. Our material is too limited to venture assigning it to one of the subspecies as defined in central Europe.