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 Galerix symeonidisi Doukas, 1986 ( Fig. 3 K-M) MATERIAL AND MEASUREMENTS (in cm). — Spain . El Canyet , 1 mandible sin. with broken m2 and slightly damaged m3, IPS20978 , – × –; 1.67 × 1.19 ; Can Julià , 1 m3 dext., IPS86238 , 1.78 × 1.16; Can Martí Vell I: 1 P2 sin., IPS96863 , 1.67 × 0.86 ; 1 mandible dext. with m1 and m2, IPS96861 , 2.75 × 1.81; 2.32 × 1.72; 1 labial half of P4 dext., IPS96862 , 2.27 × – ; Can Martí Vell III, 1 M2 dext., IPS90033 , 1.75 × 2.29; 2 M2 sin., IPS90034 -35, 1.75 × 2.42; 1.79 × 2.48 ; Les Cases de la Valenciana 1, 1 M2 sin., IPS86824 , 1.86 × 2.47; 1 M3 sin., IPS86596 , 1.11 × –; 1 p1 sin., IPS86598 , 1.11 × 0.60; 1 p2 dext., IPS86597 , 1.14 × 0.56; 1 m3 sin., IPS86825 , 1.90 × 1.21 ; Vilobí de Penedès , 1 P3 lingual fragment, IPS86856 ; 2 damaged M1 dext., IPS86858 , 86860; 1 p4 sin., IPS86852 , 1.91 × 1.12 ; La Vinya Vella , 1 M1 dext., IPS106596 , 2.33 × 3.05 ; Sant Mamet 1, 1 M1 sin., IPS105162 , +1.94 × +2.60; 1 p4 dext., IPS105166 , 1.70 × 0.97; 1 m2 sin., IPS105164 , 2.36 × 1.48; 1 m2 dext., IPS105165 , 2.41 × – . DESCRIPTION P2 The outline is elliptical and mesiodistally elongated. The two roots of this tooth are divergent and the main cusp is a bit anterior to the centre of the premolar. P4 Only the labial side is preserved in the single available specimen. The parastyle is somewhat protruding. The preserved part of the posterior margin is covered by a thin but well-defined cingulum. M1 The two available specimens from VdP are damaged. Only the specimen from La Vinya Vella is completely preserved. The anterior width is quite similar to the lingual length. The posterolabial corner is elongated. The protocone-hypocone connection is stronger than the protocone-metaconule connection. The latter is even missing on one of the specimens. The posterior arm of the metaconule is short and it does not reach the posterior cingulum in any specimen. M2 The protocone-hypocone connection is strong. Two out of three M2 from Can Martí Vell III lack the protocone-metaconule connection. The posterior cingulum is continuous in the two specimens in which this character can be ascertained. M3 Even though the anterolabial part of the only available specimen is damaged, the last upper molar is triangular in occlusal view, in spite of the broken anterolabial part. The central basin is closed, except for a small notch near the paracone, thus partially interrupting the anterior crest. Mandible The only fragment from El Canyet is very damaged. It is a posterior fragment of the horizontal ramus with no significant characters other than a rather high mandibular corpus. The position of the mental foramen is unknown. p1 The outline of the occlusal surface is slightly ovoid. The tip of the unicuspid lies in the anterior part of the tooth; there is a small flattening behind it. The p1 is one-rooted. p2 In contrast to the p1, the p2 is double rooted. The outline is elliptical and the tip more centrally placed. There is no posterior flattening. p4 The premolar consists of a trigonid with a very short posterior flattening. The metaconid is well developed. In our hardly worn specimen, the paraconid is formed like a sharp ridge. The paralophid is interrupted by a clear notch between the paraconid and the protoconid. The posterior flattening is bordered by a low posterior ridge. m1 and m2 These elements show the invariable aspect typical of the genus, being the m1 larger than the m2 and having a more pointed paraconid, which is rounded in m2. Other than that, the only remarkable trait on these elements is the posterior cingulid, which rises obliquely the distal margin of the tooth, but it does not reach the postcristid. m3 Specimens have been found in Les Cases de la Valenciana 1, El Canyet and Can Julià. The element from El Canyet is not complete, thus hampering comparisons. The element from Can Julià only differs from that of Les Cases de la Valenciana described in Jovells-Vaqué et al. (2018) in having a slightly smaller talonid. REMARKS Galerix is a very common element in the Miocene faunas of the late and middle Miocene of Europe. The genus probably originated in Anatolia ( Van den Hoek Ostende 1992 ). The earliest representatives in Europe are known from MN 3 onwards. In Spain, the oldest Galerix known hitherto is G. remmerti Van den Hoek Ostende, 2003 a large-sized species which seems closely related to the central European G. aurelianensis (Van den Hoek Ostende 2003) . In addition, Van den Hoek Ostende (2003) noted the presence of a very large Galerix in some Ramblian assemblages, known from a couple of elements only. At the start of the Aragonian, G. remmerti is replaced by the much smaller-sized G. symeonidisi in the Daroca-Calamocha basin ( Van den Hoek Ostende & Doukas 2003 ). A similar situation occurs in central Europe, but here G. symeonidisi gradually replaces G. aurelianensis . Galerix symeonidisi is in its turn replaced by G. exilis . Initially, this replacement was considered a matter of gradual evolution ( Ziegler & Fahlbusch 1986 ) but according to the cladistic analysis of Borrani et al. (2018) , G. symeonidisi and G. exilis are not closely related species. Based on the large overall variation in some Galerix assemblage of the Daroca-Calamocha area, Van den Hoek Ostende and Doukas (2003) showed that in fact the two species must have occurred together. This discussion clearly demonstrated the difficulties of interpreting Galerix assemblages if more than one species could be present. In the case of the material from the Vallès-Penedès, identifications are hampered by the scarce material and the absence of characteristic elements (P3, p4) in most assemblages. In the P4 from SAB3, protocone and hypocone are separated, a condition also found in one specimen of G. remmerti from San Roque 4A ( Van den Hoek Ostende 2003). The M1 from Sant Mamet 1 and the M2 from Turó de les Forques are peculiar, as their metaconules do not have the crescent-shape that is considered typical for the Galericini ( Van den Hoek Ostende 2001 ). However, this morphotype is also known from Petersbuch 28 ( Klietmann et al. 2014a ) and from Zone B localities in the Daroca-Calamocha area (LHO, pers. obs.). The molar somewhat exceeds the largest specimen of Galerix remmerti from the inland localities, as do the m2 and M3 from the same locality. Also the p1 from Sant Andreu de la Barca 1 is remarkably large. The molars do fall in the size range of G. aurelianensis Ziegler, 1990 , a central European species that mainly differs from G. remmerti in size and presumably was part of the same migration event at the beginning of the Ramblian. Actually, the distinction between G. aurelianensis and G. remmerti is not always easy without a large sample. For instance, Klietmann et al. (2014a) suggested that the form present in the French localities of Beaulieu and Bouzigues ( Aguilar et al. 2003 ; Sigé et al. 1997 ) was likely G. remmerti . Would it be the case, the presence of G. aurelianensis in the Vallès-Penedès basin would create a biogeographic problem of difficult interpretation. Even the possibility that these two forms were conspecific and the differences observed are only a matter of latitudinal variation cannot be discarded. In the absence of further elements to judge, a conservative taxonomic solution is herein preferred, noting that these questions must deserve attention in future works. FIG . 3. — Marsupial and galericines from the early Miocene of the Vallès-Penedès: Amphiperatherium frequens von Meyer, 1846 ( A -G ). A , IPS86180. dP3 dext. (rev) SAB1. B , IPS102010. M3 dext. (rev) CMV-II. C , IPS86243. M4 dext. (rev). CMV-I. D , IPS86823. p2? dext. (rev). TFR. E , IPS86262. m1 sin. CMV-II. (E1-ocl; E2-lab). F , IPS86230. m2 sin. Palau 3B. (F1-ocl; F2-lab). G , IPS86260. m3 dext. CMV-II. Galerix remmerti Van den Hoek Ostende, 2003 ( H -J ). H , IPS85480. M2 dext. (rev). TFR. I , IPS85486. M3 sin. TFR. J , IPS85482. m2 dext. (rev). TFR. (J1-ocl; J2-lab). Galerix symeonidisi Doukas, 1986 ( K -M ). K , IPS106596. M1 dext. (rev). LVV. L , IPS96861. m1+m2 dext. (rev). CMV-I. (L1-ocl; L2-lab). M , IPS86238. m3 dext. (rev). CJ. (M1-ocl; M2-lab). Scale bar: 1 mm. The younger material of Galerix from the MN 4 localities in our study clearly belongs to a smaller species.Jovells-Vaqué et al. (2018) already identified the galericine from Cases de la Valenciana as Galerix cf. symeonidisi based on the size and the absence of a protocone-metaconule connections in two of the three upper molars. Their identification is now confirmed by a lingual fragment of a P3 from Vilobí del Penedès which clearly shows a hypocone, one of the main characteristics of G. symeonidisi . Of the two damaged M1 from this locality, one misses the protocone-metaconule connection. The complete M1 from La Vinya Vella (the only insectivore fossil found in this site) is tentatively ascribed to the same species due to the strong similarities with the material from Vilobí del Penedès. However, because of its somewhat larger size it could even belong to G. exilis , which is known to co-occur and ultimately replace G. symeonidisi in the Daroca-Calamocha area in Zone C ( Van den Hoek Ostende & Doukas 2003 ). This would be in line with La Vinya Vella being one of the youngest localities in this study. The material from LVV is, however, too limited and we thus refrain from ascribing it to a different species. In short, the localities of the Vallès-Penedès seem to fit the pattern observed in the inland basins, in which the large-sized G. remmerti is replaced by G. symeonidisi around the Ramblian-Aragonian boundary. Moreover, the large size of the larger species fits with that of the central European G. aurelianensis , suggesting a near continuous distribution between the northern species and G. remmerti . As yet, the material is limited and some caution must be observed in these conclusions, but they fit in nicely with the current ideas about early Miocene galericine distribution patterns.