Geobios 76 (2023) 53–73 Contents lists available at ScienceDirect Geobios journal homepage: www.elsevier.com/locate/geobio Original article A new rodent chronology for the late Neogene of Spain q Jan A. van Dam a,b,⇑, Pierre Mein c, Miguel Garcés d, Ronald T. van Balen e,f, Marc Furió b,g, Luis Alcalá h a Department of Earth Sciences, 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 Departament de Dinàmica de la Terra i de l’Oceà, Facultat de Ciències de la Terra, Geomodels Research Institute, Universitat de Barcelona, Barcelona, Spain e Department of Earth Sciences, Vrije Universiteit Amsterdam, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands f TNO-Geological Survey of the Netherlands, Princetonlaan 6, 3584 CB Utrecht, The Netherlands g Serra Húnter Fellow, Universitat Autònoma de Barcelona, Geology Department, Cerdanyola del Vallès, 08193 Barcelona, Spain h Parque de las Ciencias de Andalucía, Av. de la Ciencia, s/n, Granada 18006, Spain a r t i c l e i n f o a b s t r a c t Article history: The number of late Neogene Spanish micromammal-containing continental sections with a correlation to Received 10 March 2022 the Geomagnetic Time Scale is steadily growing. Nonetheless, well-calibrated sections with dense micro- Accepted 11 January 2023 mammal records are still rare, biostratigraphic correlations between basins are not straightforward, and Available online 23 January 2023 ages of uncalibrated sites are poorly constrained. Here, we aim at improving the chronology of Iberian micromammal sections and sites for the interval 8.5–2 Ma by: (i) analyzing qualitative and quantitative Keywords: similarities between rodent assemblages and turnover in the different basins, (ii) formulating a system of Miocene fifteen Iberian assemblage biozones, and (iii) constraining the ages of zone boundaries, assuming Pliocene Spain isochroneity across basins. Age uncertainty ranges for most known Iberian micromammal sites are Rodentia obtained by combining regional biozone boundary ages with local magnetostratigraphic records, sedi- Biostratigraphy mentation rates and/or evolutionary rates. In addition, our results include new, integrated stratigraphic Geochronology records from the Jumilla-La Celia and Teruel Basins, which are used to constrain the thus far poorly dated interval covering the latest Tortonian and earliest Messinian (8–7 Ma). Ó 2023 The Author(s). 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 et al., 1997; Oms et al., 1999; van Dam et al., 2001; Agustí et al., 2006a; Piñero et al., 2018; see Fig. 1 for the location of basins). The use of paleomagnetism for dating Spanish continental sed- Even though almost every subinterval is covered magnetostrati- iments dates back a long time (Dijksman, 1977). It was not until graphically by one or more sections, many late Neogene mammal the nineteen-nineties, however, that mammal-containing sections sites of the Iberian Peninsula still lack accurate age estimates. across the Iberian Peninsula started to be sampled in a systematic The main reason for this is the absence of a regional (Iberian) bios- way (Krijgsman et al., 1994, 1996; Garcés et al., 1996, 1998; Agustí tratigraphic framework with well-dated boundaries. Although a et al., 2001). Currently, the magnetostratigraphic framework for lack of resolution is plaguing both the macro- and micromammal the late Neogene and its correlation to the Geomagnetic Polarity record, the density of micromammal localities and specimens is Time Scale (GPTS) includes many sections, of which the most much higher, making this group the most suitable for biostratigra- important ones are situated in the Teruel Basin (La Gloria – Búnker phy. Within micromammals, rodents have traditionally been used de Valdecebro, Villalba Alta, Orrios, Escorihuela, Gea, Concud Esta- for the definition of biozones because of their abundance and rapid ción), Cabriel Basin (Cabriel, Venta del Moro, La Bullana, Fuente del evolution. The revision in this study thus embodies the integration Viso), Júcar Basin (Valdeganga), Fortuna Basin (Chorrico, Sifón de of local (basin-scale) rodent-based biozonations into a regional- Librilla), and Guadix-Baza Basins (Galera, Zújar, Gorafe) scale, Iberian stratigraphic framework. This framework, in turn, (Krijgsman et al., 1996; Garcés et al., 1997, 1998,1999; Opdyke will be used to make chronological inferences for most Iberian micromammal sites by combining ages of regional zone bound- q aries with local sedimentation and/or evolutionary rates. Corresponding editor: Gilles Escarguel. ⇑ Corresponding author at: Department of Earth Sciences, Utrecht University, The use of regional biozone boundaries as age calibration points Vening Meinesz Building A, Princetonlaan 8a, 3584 CB Utrecht, the Netherlands. for local stratigraphic interpolation requires these boundaries to be E-mail address: j.a.vandam@uu.nl (J.A. van Dam). isochronous. Although time equivalence of biozones is suggested https://doi.org/10.1016/j.geobios.2023.01.001 0016-6995/Ó 2023 The Author(s). Published by Elsevier Masson SAS. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 Fig. 1. Map of the Iberian Peninsula with Cenozoic basins in yellow. Names correspond to basins used in this study. by the strong parallels between the local biozonation systems cur- sequences have been mutually correlated to each other in a routine rently in use in terms of numbers and names of zones (Fig. 2), we fashion (Mein et al., 1990; van Dam et al., 2001; García-Alix et al., prefer to go one step further and perform actual computations of 2008a; Minwer-Barakat et al., 2012; Mansino et al., 2017; Piñero faunal similarity, both in terms of numbers of species and relative et al., 2018; Fig. 2). abundances. Specifically, we will draw upon ‘shared turnover’ (i.e., Between-region faunal similarities across Spain were already between-basin shared species last and first occurrences) in order to observed in the nineteen-eighties (Mein et al., 1983; Godoy et al., characterize the magnitude and geographical extent of potential 1983; Adrover, 1986), although at that time most biostratigraphic ‘events’ underlying community change at biozone boundaries. units were considered to have a large geographic domain, even up The latest Tortonian to earliest Messinian (8–7 Ma) is chrono- to the scale of the entire European continent. Accordingly, the just logically one of the most poorly constrained late Neogene intervals erected system of Neogene European Mammal units (MN1-17; on the Iberian Peninsula. In this paper, we will fill this gap by Mein, 1989; Agustí et al., 2001, Hilgen et al., 2012) was instantly including new faunal, magnetostratigraphic and cyclostratigraphic adopted as the starting point for further subdivision (e.g., data from the Jumilla – La Celia and Teruel Basins. ‘MN15a/b’, ‘MN16a/b’). However, with an increasing number of magnetostratigraphic studies appearing, first and last appearances of species and genera not seldomly turned out could be diachro- 2. Stratigraphic background nous at the continental scale, effectively reducing the domain of isochroneity to restricted bioprovinces only (van Dam, 2003; van 2.1. Biostratigraphy der Meulen et al., 2012). As strong diachrony is expected to corre- late to low faunal similarity (the common cause being a barrier of Existing Iberian latest Miocene and Pliocene local biozonation some kind), continents such as Europe are expected to be charac- schemes have a very similar structure, with approximately the terized by strong faunal heterogeneity. Indeed, the overlap same number of zones often carrying the same names and largely between late Neogene Iberian and Central European faunas containing the same dominant species. As a result, local zone (Daxner-Höck, 1980; Dahlmann, 2001; Franzen et al., 2013) typi- 54 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 Fig. 2. State-of-the-art correlation table between local Iberian biozones, newly proposed Iberian biozones, European-scale MN units, and Stages/Ages. Definitions of local zones are given implicitly (Range Zones), by first and last occurrences (FO/LO, for Concurrent-Range Zones) or by species lists (for Assemblage Zones). MN Zones are defined faunistically and not stratigraphically (Hilgen et al., 2012). cally consists of two or three species only. Even examples of zero polonicus Zone in the Guadix-Baza Basin. The first of these, which species overlap are known, e.g., between latest Miocene faunas of is represented by the sites of the Barranco del Blas (García-Alix the Western and Eastern Mediterranean; Schmidt-Kittler, 1995; et al., 2007), is a ‘floating’ zone separated from the next older zone van Dam, 2003). by the above-mentioned barren interval and (besides unidentifi- The similarities in number of zones and names between local able arvicoline material) containing species that can be correlated Spanish latest Miocene to earliest Pleistocene zonation schemes to either MN15 or MN16 (e.g., Ruscinomys cf. europaeus, Stepha- (Fig. 2) suggest that the Iberian Peninsula largely functioned as a nomys donnezani, Castillomys crusafonti). The Mimomys polonicus single bioprovince during this period. As can be seen in the table, Zone is a short zone between the Mimomys hassiacus and Kislangia the number of zones within approximately correlative larger strati- ischus Zones, of which the definition is based on a single and thus graphic units such as (sub-)Stages or MN units is mostly the same. far poorly documented occurrence of M. polonicus in Zújar 10 For example, MN11 and MN12 (lower and middle Turolian) are (Piñero et al., 2018). Also given the prospect of synonymizing M. covered by one and two local zones, respectively, in both the Bajo polonicus with M. hassiacus (Tesakov, 1998), the usefulness of this Segura and Teruel Basins. Similarly, MN13 (upper Turolian or ‘Ven- zone for overall Iberian biostratigraphy is doubtful. tian’) is covered by three named zones in the Teruel and Granada Time equivalence of zones is also suggested by the use of iden- Basin, and MN14 (lower Ruscinian) by three zones in the Teruel, tical zone names (Fig. 2), reflecting pan-Iberian occurrences of Alcoy, and Guadix-Baza Basins (with MN14 defined faunistically characteristic species. Examples are (fractions refer to the number and not stratigraphically; in the latter case it is shortened, because of basins): Huerzelerimys turoliensis (MN12: 2/2), Castromys inflatus of the early entry of the MN15-defining species Mimomys davakosi (MN13: 2/3), Paraethomys meini (MN13: 3/5, Apocricetus barrieri in Greece; (Hilgen et al., 2012; Hordijk and de Bruijn, 2009)). (MN13: 3/4), Paraethomys baeticus (MN14: 3/4), Trilophomys Finally, MN15 and MN16 are covered by two zones in both the Ter- (MN14: 4/4), Dolomys (adroveri) (MN15: 3/4), Kislangia (ischus) uel and Guadix-Baza Basins. ‘Mismatches’ can be explained by the (MN16: 2/3). The use of different names is sometimes semantic: occurrence of barren intervals: the lack of a ‘middle MN13’ zone in for instance, the Zone of ‘Mimomys archaique’ (MN15 Teruel Basin) the Bajo Segura basin (Martín-Suárez and Freudenthal, 1998: effectively applies to M. davakosi, as used in the names of corre- between CR31 and CR6) and the lack of an ‘upper MN140 interval sponding zones. In still other cases, the use of different names used in the Granada Basin (part of the ‘Barren Interval 2’ of García- is only a matter of coincidence, with the same combinations of spe- Alix et al., 2008a). Only two locally defined zones cannot be cies present in time-equivalent zones. straightforwardly correlated to other basins: the Stephanomys don- Although the concept of a single Iberian province with homoge- nezani Assemblage Zone in the Granada Basin and the Mimomys neous biostratigraphic units separated by isochronous zone 55 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 boundaries is suggested by Fig. 2, actual quantitative analyses of The two parallel sections of La Celia (JCLC) and Los Barracones faunal similarity and shared turnover are thus far lacking. It is (JCBA) in the Jumilla – La Celia Basin were sampled paleomagnet- one of the gaps we aim to fill in this work. ically. These two sections, which are separated geographically by 1–1.5 km, reveal a rapid lateral sedimentary facies change (Fig. 4 (F–H)). Whereas the JCLC section dominantly consists of conglom- 2.2. Tortonian-Messinian transitional interval erates, reddish mudstones, silts, and fine sands, pointing to a prox- imal sedimentation regime (debris flows; van Balen et al., 2015), Until now, the Tortonian-Messinian transition interval (which is the JCBA section mainly contains reddish to greenish mudstones approximately time-equivalent to the continental middle-upper and gypsum, indicating distal alluvial to shallow lacustrine Turolian, or Turolian-Ventian transition; Morales et al., 2013) is conditions. poorly constrained in continental Spanish records. Either a faunal gap is present despite a well-documented magnetostratigraphic record (e.g., Fortuna Basin (Agustí et al., 2006a; Garcés et al., 3.2. Magnetostratigraphy 1998; Piñero and Agustí, 2019) a magnetostratigraphic gap exists despite a rich and well-documented faunal record (e.g., Bajo Segura An average sampling resolution of about 1 site/meter was Basins and Teruel Basins; Martín-Suárez and Freudenthal, 1998; applied and samples were drilled in situ and oriented with a mag- van Dam et al., 2001). The correlation problems pertaining to this netic compass mounted on a clinometer device. A total of 130 sam- interval are reflected in a large age uncertainty for the approxi- ples were processed in the laboratory, sharing a similar lithology mately correlative MN12/13 transition (7.4–6.8 Ma; Hilgen et al., consisting of red clay, silt and silty sandstone. 2012). Paleomagnetic analysis consisted of stepwise thermal demag- The new magnetostratigraphic and cyclostratigraphic results netization of the natural remanent magnetization (NRM) of one for the late Tortonian and early Messinian that will be presented sample per site, in order to isolate the most stable characteristic in this work are based on sections in the Teruel and Jumilla-La remanent magnetization (ChRM). The remanence vector was mea- Celia Basins (Figs. 1–3). The Teruel Basin is a half-graben situated sured in the 2G-SRM750 superconducting rock-magnetometer of within the Iberian Chain and contains a fairly complete succession the laboratory of paleomagnetism of the CCiTUB (University of Late Miocene and Pliocene continental sediments (Anadón et al., Barcelona-CSIC). Measurement of the magnetic susceptibility fol- 1990; Alonso-Zarza and Calvo, 2000; Ezquerro et al., 2020). The lowed each demagnetization step in order to monitor chemical basin is well-known for its rich mammal record (van de Weerd, changes of the magnetic mineralogy upon heating. ChRM direc- 1976; Adrover, 1986; Mein et al., 1990; Alcalá, 1994; van Dam tions were picked by visual inspection of Zijderveld plots and cal- et al., 2001; Alcalá et al., 2018). A significant part of the record is culated by principal component analysis. In order to interpret the constrained magnetostratigraphically (Krijgsman et al., 1996; polarity of the samples, the latitude of the virtual geomagnetic pole Opdyke et al., 1997; Garcés et al., 1999). Locally, the age accuracy (VGP) was calculated from each ChRM direction, and positive (neg- of mammal sites has been further increased using orbitally forced ative) VGP latitudes were interpreted as normal (reversed) lithological alternations (early Late Miocene; Abdul Aziz et al., polarity. 2004; Abels et al., 2009b). The Jumilla-La Celia basin is a small basin situated along the 3.3. Micromammal paleontology Jumilla Fault Zone (JFZ) in the external part of the eastern Betic Cordillera (van Balen et al., 2015). The basin is known for its late One larger (CG5, n = 374) and four smaller micromammal local- Neogene volcanics aged 7.6–7.2 Ma (7.8–6.9 Ma with error bars ities (CG1,2,4,6,7, 40°240 200 N, 1°80 5200 W, total n = 43) were sampled included) based on K-Ar data (Nobel et al., 1981). A re- above and below the classic macromammal site of Concud Cerro de interpretation of micromammal data (sites La Celia and Los Gar- la Garita (CCCG; Fig. 4(E)) in the Teruel Basin. A previously gantones; van Dam et al., 2014) indicates a consistent MN12- described small micromammal sample from this classic level (code correlative age (between 7.6 and 7 Ma; Hilgen et al., 2012). CC in van de Weerd, 1976) was extended (total n = 467, new mate- rial coded CG). Furthermore, two new sites (La Roma 5A and 5B, 3. Material and methods R5A-B; 40°300 800 N, 1°00 4000 W, n = 225 and 425, respectively) were sampled north of the village of Peralejos about 500 m east of the 3.1. Sampled sections Tortonian-Messinian transition macro- and micromammal site La Roma 2 (Zone J3; Alcalá, 1994; Alcalá et al., 2005; Fig. 3). The stratigraphic context of both sites Three parallel sections were studied near the village of Concud is somewhat obscured by the local deposition of Quaternary grav- in the Teruel Basin (Fig. 3). These sections contain or are close to els and recent valley fill. Bed tracing suggests that R5A and B the well-known micromammal sites of Concud Cerro de la Garita belong to a highly condensed (10 m thick) interval bracketed (CC), Concud Barranco de las Calaveras (CCL) and Las Pedrizas by basal Zone L (Peralejos D) and Zone M1-2 faunas (El Capón 2) (Concud 2 and 3; CC2, CC3) as previously described (van de (faunal information in van Dam et al., 2001). Finally, a small collec- Weerd, 1976; Alcalá, 1994) (Fig. 4(A–E)). The Concud Cerro de la tion (Tortajada Pueblo, TOP, 40°240 3000 N, 1°40 0000 W; n = 6; Fig. 3) Garita (CCCG) section consists of a lower alluvial part consisting was sampled at the transition between carbonate/marls (Alfambra of red mudstones and an upper shallow lacustrine part dominated Fm.) and gypsum (Tortajada Fm.) near the village of Tortajada (for- by limestones and marls (Fig. 4(D)). The alluvial part was sampled mation names after van de Weerd (1976)). All newly sampled paleomagnetically. Although the magnetic signal in the Las Pedri- rodent teeth were measured using the methods as described by zas (CCLP) was too weak to perform a reliable magnetostrati- van de Weerd (1976) except for Gliridae, which were measured graphic study, the regular bedding pattern of this section and of according to Daams (1983). The newly sampled material is housed the Barranco de las Calaveras (CCBC) section could be interpreted in the Museo Aragonés de Paleontología (Teruel). cyclostratigraphically and was used as an additional tool in dating Besides a summary of the newly sampled late Miocene material, mammal sites. Seven hand samples from the CCLP section were we also include unpublished measurement statistics of previously subjected to X-ray diffraction (XRD) to check the mineral composi- described Late Miocene localities (Besems and van de Weerd, tion of the sediments. In addition, thin sections were prepared 1983; van Dam, 1997; van Dam et al., 2001) and faunal lists for from these samples to study microscopic features. Pliocene sites of the Teruel Basin for which previously only partial 56 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 Fig. 3. Geological map of the central part of the Teruel Basin with fossil sites. CP: El Capón; E: Escorihuela; O: Orrios; P: Peralejos; MD: Modorras; VA: Villalba Alta and Villalba Alta Río; LC: Lomas de Casares; BC: Barranco de Cuevas; R: La Roma; VB: Villalba Baja; CC: Concud Cerro de la Garita, Barranco de las Calaveras and Las Pedrizas; CG: Cerro Gordo; CL: Celadas; K: Las Casiones; TP: Tortajada Pueblo; RG: El Regajo; A: Alfambra; T: Tortajada; M: Masada del Valle and Masada Ruea, PM: Puente Minero; TC: Teruel Transformadores and Teruel Cementerio; V: Valdecebro; Q: El Arquillo; J: Teruel Judería; LM: Los Mansuetos; AL: Aljezar B; G: La Gloria and Los Aguanaces; B: El Búnker (de Valdecebro); TD: Teruel Desvío; PI: Poblado Ibérico; GE: Gea (=La Guea); CE: Concud Estación; CV: Concud Village; LA: La Calera; VI: Villastar; CU: Cubla. Abbreviations: P, Paleozoic; T, Triassic; J, Jurassic; C, Cretaceous; PG, Paleogene; N, Neogene; Q, Quaternary. information was published (Godoy et al., 1983; Adrover, 1986; itanomys adroveri was used to scale MN12-13 sites (thereby Mein et al., 1990). extending on van Dam’s (1997) application for the Teruel Basin). The size-ranked sequences were checked for logical inconsisten- 3.4. Biochronology cies in zone assignments (‘zone test’) or known superposition of beds (‘superposition test’). Ages of sites without a first-order magnetostratigraphic calibra- In case no constraints could be put on the age of a locality tion but positioned in a section with dated zone boundaries were except for its biozone, the age uncertainty range of the locality constrained assuming constant sedimentation rates between was taken maximally with regard to the biozone age uncertainty boundaries. In two cases quantified evolutionary stages were used range. Locality age precision was set at 0.01myr, which was also as an additional metric for age interpolation (cf. Legendre and taken as the minimum age difference between two localities of Bachelet, 1993; van Dam, 1997; Maul et al., 1998): (i) quantified which one was found to be ‘younger’ or ‘older’ than the other based anagenetic (evolutionary) change in character state distributions on faunal content (‘faunal stage duration convention’). Consis- of the Progonomys hispanicus – Occitanomys sondaari lineage was tently, a ‘faunal stage’ was defined as any biochronological unit used to scale Zone K localities in the Teruel Basin (principal compo- that could be distinguished from another unit by a different nent scores of van Dam, 1997), and (ii) molar size increase in Occ- species/morphotype inventory. A second convention (‘boundary 57 J.A. van Dam, P. Mein, M. Garcés et al. 58 Fig. 4. Chronological interpretation of middle Turolian (latest Tortonian – earliest Messinian) sediments in the Teruel and Jumilla-La Celia Basins. A. Local map with approximate section paths (in red) and sites (in blue) near Geobios 76 (2023) 53–73 Concud; K: kilometer mark along main road. B. Cyclically deposited beds in the Las Pedrizas (CCLP) section; cycle numbers in red, rock samples in blue. C. Internally laminated beds in the CCLP section, cycle numbers in red. D. Cyclically deposited beds in the Barranco de las Calaveras (CCBC) section; cycle numbers in red. E. Studied sections and correlations to the Geological Time Scale (Raffi et al., 2020). F. Local map with approximate section paths (in red) and sites (in blue) near La Celia; K: kilometer mark along main road; dark grey area: volcanic deposits; light grey area: diapir. G. Mudstones, gypsum and carbonate deposits in Los Barracones (JCBA) section. H. Conglomerates and mudstones in basal part La Celia (JCLC) section, with local fault indicated. J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 fauna convention’), was applied to rare boundary faunas, i.e., For the purpose of this study, two was considered the minimum assemblages that show an overlap of taxa that normally belong number of shared species events (FO and/or LO) required in order to the pre- or post-boundary zone. The maximum possible dura- to regard two local zone boundaries as time-equivalent, provided tion of this overlap was set to 0.1myr resulting in an uncertainty that the transition could not be marked as a ‘rare species’ event. of ±0.1myr for the age of a zone boundary as defined in a boundary Using Gaston’s (1994) criterion for rare species as those species fauna. confined to the upper quartile of an abundance-ranked species dis- Karstic localities (including well-known sites such as Layna, tribution, we checked STARA-B values against a value amounting to Moreda, Belmez and Sarrión; Adrover, 1986; Castillo, 1990; twice the average rare-common threshold percentage across all Freudenthal and Martín-Suárez, 1990) were excluded from this basin-zone combinations. Because most of our abundance distribu- study, as these localities tend to contain heterogeneous and poten- tions are strongly skewed towards rare species, we also checked tially biased populations as evidenced by abnormally large vari- STARA-B values against a threshold value based on the median of abilities in size and shape. the ranked species distribution. Boundaries characterized by zero or one shared event(s) (FO or LO) or with a below-threshold STAR- 3.5. Faunal similarity and isochrony A-B value (i.e., not in the ‘common regime’) were thus considered suspect with regard to isochroneity. Because the records of most Because of a lack of time-equivalent sets of magnetostrati- basins cover only a part of the studied interval, we chose to aggre- graphically calibrated sections with dense fossil records (across gate basins into three ‘areas’: South (S, Granada and Guadix-Baza two or more basins), the potential isochrony of local zones was Basin), East (E, Bajo Segura and Alcoy Basins) and Central (C, Teruel evaluated indirectly. This was done by inspecting basin-to-basin Basin). Because the records from the different basins within the similarities between (i) faunal assemblages, and (ii) species areas South and East tend to alternate stratigraphically (Fig. 2), appearances and disappearances (turnover). Whereas faunal simi- composite records for these areas could be constructed straightfor- larity helps in establishing the equivalence of local biozones across wardly (i.e., mostly without between-basin averaging). Because basins in order to define regional zones, similarities in turnover only two zone boundaries were documented simultaneously for (i.e., shared First and Last Occurrences, FO and LO) help to under- all three areas, we mainly focused on two-area comparisons. Of stand the geological and biological contexts that underlie regional the 12 boundaries that could thus be analyzed, 10 were shared zone boundaries and their potential isochronous nature. between South and Central, 5 between East and Central, and 2 The rationale of looking at shared turnover is rooted in the so- between East and South. called ‘magnitude-frequency relationship’, i.e., the observation that events with larger magnitudes occur less often than events with smaller magnitudes. The spatial corollary of this principle, which 4. Results and discussion not only applies to paleontology (e.g., the ‘Kill Curve’; Raup, 1991) but also to many other scientific disciplines including mete- 4.1. Jumilla-La Celia Basin orology, seismology, and astronomy, implies that stronger events are also felt over larger areas. Thus, when the community compo- Thermal demagnetization of the natural remanent magnetiza- sition in basin X changes significantly after having experienced a tion (NRM) was successful in cleaning a soft viscous component long period of relative stasis, the perturbing event may be assumed at temperatures of 200–250 °C in most of the La Celia and Los Bar- to be both strong and widespread. If a similar change occurs in racones samples (Fig. S1, Appendix A). Above this temperature, a nearby basin Y at a an approximately similar age with similar spe- stable dual-polarity characteristic remanent magnetization cies, a single underlying event may be assumed with the two local (ChRM) was revealed. The unblocking temperature of the ChRM zone boundaries necessarily being isochronous. typically ranged between 250 and 680 °C, with a marked drop of In our quantitative comparison of local biozones, we predomi- the remanence intensity at below 580 °C, thus suggesting the pres- nantly restricted ourselves to sites that were referred to in papers ence of both magnetite and hematite. in which zonations were defined. These papers include works on According to the quality of the demagnetization data, ChRM the Alcoy (Mansino et al., 2017), Bajo Segura (Martín-Suárez and directions were classified into three classes. Class 1 directions were Freudenthal, 1998), Guadix-Baza (Piñero et al., 2018), Granada calculated from linear demagnetization trajectories heading (García-Alix et al., 2008a) and Teruel Basins (Mein et al., 1990; towards the origin and reaching complete demagnetization of the van Dam et al., 2001; this paper). Besides the sites mentioned in NRM. Class 2 directions were determined from uncomplete and these papers, we additionally used several sites from the Bajo scattered demagnetization trajectories heading towards the origin. Segura and Granada Basin that could be easily fit in biostratigraph- Class 3 directions were defined by vector clusters denoting a stable ically (Boné et al., 1978; Gamonal et al., 2018). Sites with ambigu- component at 300–500 °C intermediate temperatures after ous zone assignments (including boundary faunas) were excluded. removal of the viscous component. The quality of paleomagnetic Faunal overlap was evaluated both qualitatively (which and directions was high on average, with an average intensity of how many species) and quantitatively (relative abundances). 1.2  103A/m and an angular error < 10°. Abundances were based on the total number of first, second and From a total of 98 samples in the JCLC section, 60 (61.2%) gave third molars. In case only measured specimens were available from directions of class 1 and 27 samples (35.2%) of class 2. Samples of the literature, these were used as an alternative for total counts. class 1 and 2 were considered to determine the magnetic polarity ‘Death assemblages’, as represented by molar counts, were left stratigraphy of the sections. 11 samples (11.2%) were classified as uncorrected regarding differences in life expectancy (van Dam, class 3 and not considered here for magnetostratigraphic purposes. 1997). ‘Shared turnover abundance rates’ (abbreviated as STARA With a significant set of samples of both polarities, the analysis and STARB for basins/areas A and B, in %) were computed as the yielded robust antiparallel normal and reversed mean directions sum of the relative abundances of disappearing and appearing spe- (Fig. 5; Table 1; Fig. S2, Appendix A), indicating a demagnetization cies (with respect to the adjacent lower and upper zone assem- procedure that was efficient in isolating the ChRM from the sec- blages, respectively). These were first separately calculated for ondary viscous components. From the JCBA section 7 samples each basin, and then averaged to yield one final statistic (STARA-B). (32%) gave directions of class 1 and 15 of class 2 (68%), which 59 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 Fig. 5. Magnetostratigraphy of the La Celia (JCLC), Los Barracones (JCBA) and Concud Cerro de la Garita (CCCG) sections. Positive and negative virtual geomagnetic pole (VGP) latitudes represent normal (black) and reversed (white) polarities, respectively. For lithological symbols, see Fig. 4. clearly defined a lower magnetozone of normal polarity and an removed at low temperatures of 150–200 °C. A very stable ChRM upper one of reversed polarity. component with high unblocking temperatures up to 620 °C was revealed by a linear decay trend towards the origin. According to 4.2. Teruel Basin its quality, ChRM directions of the CCCG section were grouped into class 1 (5 directions, 31%) and class 2 (11 directions, 69%), all defining 4.2.1. Paleomagnetic results a homogeneous normal polarity magnetozone (Fig. 4(F)). Only one The stepwise thermal demagnetization treatment of the CCCG sample in the middle of the studied interval showed an intermediate samples yielded clear-cut results (Fig. S1, Appendix A). A viscous temperature (<400 °C) component of reversed polarity that might north-directed component representing a recent overprint was represent a secondary component overprinting the ChRM signal. 60 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 Table 1 Mean values and Fisher statistics of paleomagnetic directions. Abbreviations: n, number of samples; Ds/Is, declination and inclination in stratigraphic coordinates; r, result of sum of unit vector; k, Fisher precision parameter; alpha95, Fisher semi angle of confidence; csd, circular standard deviation. Polarity n Ds Is r k alpha95 csd La Celia Normal 31 008.5 54.2 27.0 7.5 10.2 29.6 Reversed 61 180.6 49.6 55.2 10.4 5.9 25.2 All samples 92 003.0 51.2 82.1 9.2 5.2 26.8 Los Barracones Normal 15 341.3 56.8 13.6 10.3 12.5 25.3 Reversed 7 200.6 38.6 5.9 5.6 27.9 34.1 All samples 22 356.2 52.9 19.0 6.9 12.8 30.9 Concud Cerro de la Garita Normal 15 354.0 45.0 13.7 11.1 12.0 24.3 All samples 16 356.3 46.3 14.6 10.7 11.8 24.7 4.2.2. Cyclostratigraphic and sedimentological results thin sections except the one from LP6 show an abundant presence The CCLP section (Fig. 4(B, C, F)) consists of 24 ± 1–2 carbonate of black, 20-40 lm large peloids (‘clots’). Lower samples LP0, LP1 cycles with an average thickness of 1.0 m (st. dev. = 0.18 m). Bed and LP2 are clearly laminated, with irregular 100–200 lm thick thickness is relatively variable in the lower part of the section (cy- lamina consisting of an alternation of reddish micritic layers and cle 7–13) raising some doubt concerning the exact number of whitish cemented layers (‘laminated micrites’ sensu Freytet and cycles in this interval. An uncertainty of ± 1 cycle is thus assumed. Verrecchia, 2002). The whitish layers contain strongly fragmented A second uncertainty regarding the number of cycles is associated biogenic carbonate, although locally recognizable charophyte frag- with a poorly exposed interval in the middle part of the section, ments (stems, gyrogonids and encrustations) can be seen. Whereas which, given its thickness, could accommodate three cycles (18– clot accumulations follow the laminar structure in LP0, clots are 20). Because the possibility of two thick cycles cannot be excluded, more homogeneously distributed in the higher samples. A decrease however, another +1 uncertainty in the number of cycles is in lamination from LP0 and LP4 is accompanied by an increase in assumed. Furthermore, an additional, lower set of six cycles was the frequency of thin cracks filled with micrite/microsparite, and identified in the CCBC section (Fig. 4(E)) consisting of 1 m- with the preservation of complete shells. LP5 shows a further thick, marly limestone beds of which most have indurated tops. increase in the abundance and thickness of cracks and in the abun- This set, which could also be traced towards the CCCG section, dance of shell material (especially of well-preserved ostracode was lithostratigraphically correlated to the base of the CCLP sec- shells). The thin section of top sample LP6 (cycle 30), on the other tion, bringing the total number of cycles in the Concud carbonate hand, shows a massive, recristallyzed micrite without clearly rec- unit to 30 ± 2. ognizable fossils. Most carbonate beds in the CCLP section are boeuf-colored, The combined macroscopic, microscopic and mineralogical evi- powdery, and have a tabular/laminar structure (Fig. 4(C, D)). Based dence points to the evolution of a low-energy, shallow lake system on macroscopic features, the section can be subdivided into three with deepest lake levels corresponding to middle part of the LP parts corresponding to cycles 7–13, 14–20, and 21–30. In the lower section (cycles 14–20, ‘shallow permanent lake facies’, Abels part, lignitic bands and streaks of gastropod fragments are occa- et al., 2009a). Whereas indications for subaerial exposure, plant sionally present. In addition, the top parts of several beds contain reworking and nearness to the lake shore are generally lacking in 2–5 mm-thick vertical structures. The first four to five cycles of the middle part, evidence for such conditions are present in the the middle part (14–20) are relatively featureless macroscopically lower and especially upper part of the section (‘very shallow tran- except for the tabular/laminar bedding structure and the presence sient lake facies’, Abels et al., 2009a). However, except for the top of scattered gastropod fragments. Given the same boeuf color, the part, carbonate reworking in the LP section appears less intense poorly exposed beds on top (18–20, mentioned above) do not indi- than in other time-equivalent parts of the basin (e.g., palustrine cate a major change in lithology. On the other hand, cycle beds 21 limestones facies in Alonso-Zarza and Calvo, 2000). and 22, which represent the transition to the upper part, are lighter Because orbitally (precession)-controlled groundwater fluctua- (beige) colored, with bed 22 showing an indurated top part. In con- tions have been shown to underlie shallow lacustrine / mudflat trast to rock samples LP0, LP1 and LP2 from the middle part, LP3 oscillations in the Teruel Basin (MN9-10; Abels et al., 2009b), and from cycle 22 shows oxidized root remains as well as more abun- because a similar mechanism was hypothesized concerning the dant gastropods including completely preserved individuals origin of sedimentary cycles in our study interval (MN11/12; (Planorbarius) (see Fig. 4(B) for the position of the samples). The Alonso-Zarza et al., 2012), we feel confident in interpreting the upper part contains a series of beds with indurated tops (cycles sedimentary cyclicity in the CCLP section as the result of 24–27) and is closed off by a brownish carbonatic marl with ori- precession-controlled changes in humidity and lake level. (Obliq- ented gastropod fragments and dark pisoids (cycle 29) topped by uity can be confidently ruled out as the dominant orbital compo- a grey to beige indurated and recrystallized limestone (cycle 30). nent, as upward counting of 41-kyr cycles would lead to an age The XRD analysis shows comparable mineralogical composi- of 6.7 Ma for CC3, implying a MN13 instead of MN12 age). tions of the lower five samples (LP0 to LP4, cycles 15–16 and 20– 22). These samples consist of >90% calcite, with the remaining part 4.2.3. Paleontological results distributed in aragonite (3–5%), gypsum (0.5–1.5%), halite (0.1–1%) All rodent species found at Concud were previously described and quartz (0.1–0.5%). The top samples (LP5-6, cycles 29–30) have by van de Weerd (1976) with the following exceptions (see Suppl. a somewhat different mineralogical composition with less calcite Text, Appendix A): (i) four small-sized, primitive-looking molars (82%) and more aragonite (13%) and quartz (2.5%) in LP5, whereas from CCCG were designated as Murinae indet., and (ii) two speci- LP6 consists almost entirely of (99%) of calcite. mens from CG7 resembling P. barbarae, were named Parapodemus Microscopic inspection in plain polarized light shows that the aff. barbarae, because of their relatively large width compared to P. studied rocks can be classified as biomicrites (mudstone to wacke- barbarae. Possibly these specimens represent the evolutionary stone) with the larger components consisting of gastropod and transition towards P. meini (Martín-Suárez and Freudenthal, 1993). ostracode shell material and (to a lesser degree) charophyte frag- R5A and R5B yielded poorly diversified micromammal faunas ments (for representative examples, see Fig. S3, Appendix A). All that are somewhat atypical. Occitanomys adroveri is abundantly 61 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 present in both levels, whereas murines normally associated with All faunal lists and specimen measurements are included in this species in the basin in Zone L such as Parapodemus barbarae Tables S1 and S2 (Appendix A). and Huerzelerimys turoliensis are lacking (three very worn speci- mens in R5B may however represent the latter form; see Suppl. 4.3. Iberian rodent biozones Text, Appendix A). Instead of Parapodemus barbarae, R5A contains two small-sized 4.3.1. Justification molars identified as P. lugdunensis, a form otherwise known from As can be seen in Fig. 6(A), all abundance-based basin-to-basin lower Zone K. Similar small-sized specimens previously assigned similarities between assemblages for time-equivalent biozones to P. barbarae occur in Valdecebro 5 (VDC5; Adrover et al., 1986: except one have values in the range 50–95%. The overall high levels their Fig. 4(B, C, E)), a locality also lacking H. turoliensis. These spec- of similarity, which correspond to an average number of six overlap- imens are here re-assigned to P. lugdunensis. Finally, our re- ping species, underline the validity of the correlation scheme as inspection of the murine material from basal M1 site Masada del shown in Fig. 2. High numbers of seven to nine or ten shared species Valle 6 (MDV6, van de Weerd, 1976) revealed a small-sized and characterize the oldest late Miocene intervals (‘11’ and ‘12a’) and a slender M2, which was previously assigned to P. barbarae, but large part of the Pliocene (Ruscinian, ‘14a’ to ‘15b’). Numbers drop which we consider belonging to the same species (P. cf. lugdunen- to five in the early Villanyian (‘16a’ and ‘16b’) as recorded for the sis). Hence, the re-appearance of P. lugdunensis in the basin is thus Guadix-Baza – Teruel Basin pair. The low similarity between these confined to a short interval consisting of the upper part of Zone L two basins for level ‘13c’ (Fig. 6(A)) partly depends on the use of dif- and basal part of Zone M1. ferent chronospecies (in Apodemus and Stephanomys), which, in turn, In order to accommodate the deviating faunas of R5A and may be the result of the considerable time span covered by this zone VDC5 biostratigraphically, we define a new local zone, the Para- (see below) in combination with a low locality density. On the other podemus barbarae – Castromys inflatus Interval-Zone, which is hand, the high abundance-based similarity between the Bajo Segura coded by L2 (Fig. 2). The previously defined Parapodemus barbarae and Teruel Basin for level ‘12b’ is caused by the contribution of one Interval Zone (Zone L, van Dam et al., 2001), of which until now dominant species only (Occitanomys adroveri). the top was defined by the entry of Stephanomys ramblensis (van Shared turnover results are shown in Fig. 6(B) (for underlying de Weerd, 1976) is redefined as the P. barbarae Range Zone (L1). data, see Table S3, Appendix A). First, average threshold values of Whereas L1 thus contains P. barbarae and Huerzelerimys turolien- 1.1% and 2.7% were computed for separating rare from common sis, L2 lacks both these species but contains P. lugdunensis. In species using the quartile criterion (Gaston, 1994) and median cri- addition, L2 differs from L1 by the larger size of Occitanomys terion, respectively. In combination with our required minimum of adroveri. two appearing/disappearing species, we thus consider boundaries The present work seems the appropriate place to rename the with STARS-C, STARE-C, or STARE-S values exceeding 2.2-5.4% as basi- local Pliocene zones of Mein et al. (1990) in line with the interna- cally isochronous. tional stratigraphic guidelines (Interval Zones; Salvador, 1994; A very high STARE-C value of 150% involving no less than nine Fig. 2). We additionally apply the lettered codes N1-O4 to these species leaves little doubt for isochrony of the oldest local zone zones, thereby extending on the Miocene codes J1-M3 of van boundaries, i.e., the ones between the Huerzelerimys vireti and Dam et al. (2001) and the N1, N2 and O codes of Morales et al. Parapodemus barbarae Zones in the Teruel Basin and between the (2013). We (as in Mein et al., 1990) slightly depart from Morales Occitanomys sondaari and H. turoliensis Zones in the Bajo Segura et al.’s definitions in not using the entry of Promimomys as a local Basin (Fig. 6(B)). According to the magnitude-frequency relation- stratigraphic criterion. Although Promimomys has a wide distribu- ship, we are dealing with a strong event affecting a large area. tion and its FO is used as a marker for European unit MN14 (Hilgen The same reasoning applies to the uppermost boundary between et al., 2012), it is very rare on the Iberian Peninsula, rendering the zones with Mimomys hassiacus and zones with Kislangia ischus, event less useful for Iberian biostratigraphy. With sole occurrences which involves five species events and for which a STARS-C value in Teruel Basin localities La Gloria 4, Celadas 9 and 12, the entry of of 113% is computed. (Note that the M. hassiacus and M. polonicus Promimomys in Spain can be shown to post-date that in southeast- Zones in the Guadix-Baza Basin combined, see Section 2.1). ern Europe, where it is dated between 5.40 and 5.23 Ma (Hordijk Although all boundaries satisfy our combined criteria of a min- and de Bruijn, 2009; see Section 4.3.2 for discussion). For this rea- imum of two species and a STAR value >2.2 or even 5.4% (Fig. 6(B)), son, we basically retain Mein et al.’s (1990) ‘Zone à 2 Paraethomys’ several transitions deserve some additional discussion given their (renamed here as Paraethomys baeticus – Trilophomys interval relatively low values (<20%): Zone) and accompany it by the code N2, which now encompasses both Morales et al.’s Zone N2 (without Promimomys) and the lower  During the ‘12a’–‘12b’ transition, H. turoliensis and Eliomys sp. A part of their Zone ‘O’ (with Promimomys). N3 is used for Mein disappear with an associated STARE-C value of 12%. On the other et al.’s (1990) ‘Trilophomys Zone’. Finally, we introduce the new let- hand, both areas (East and Central) undergo a major change in ters O1-4 to cover Mein et al.’s (1990) four youngest zones (ap- their fauna: seven additional species (with abundance >0.1%) proximately corresponding to MN16 and MN17), thereby disappear in the central part (Teruel Basin) and eight (different) avoiding alphabetic overlap with the previously defined Oligocene species experience their FO or LO in the eastern Bajo Segura Zones P, Q, etc. in the Spanish interior region (Álvarez Sierra et al., Basin (Table S3, Appendix A). In this particular case, we consider 1987). it a probable scenario that one single event of considerable The revisions of Cricetinae by Freudenthal et al. (1991, 1998) magnitude resulted in extinction in one area and replacement urged a re-inspection of the Teruel Basin material of this group. in another. The fact that faunas in East and Central became very We concluded that the identifications of Neocricetodon occidentalis different during ‘12b’ (Central has only one species in common and Apocricetus alberti in van Dam et al. (2001) can be retained, but with East and South, Fig. 6(A)) points to a strong difference in that the Ib-N12 and N13a populations previously identified as N. environmental conditions and/or the temporary presence of a (‘Kowalskia’) occidentalis should be transferred to Apocricetus aff. barrier. Sampling (a poor representation of localities, e.g., two plinii. Furthermore, the cricetines from Valdecebro 6, Arquillo 1 sites in the Teruel Basin) may play an additional role; and Villastar, which, according to Adrover et al. (1993) belong to  The low STARE-S value compared to much higher STARS-C and the same species as the one from La Gloria 5, are considered to rep- STARE-C values at the ‘12b’–‘13a’ transition is due to the excep- resent A. alberti (for details, see Suppl. Text, Appendix A). tional dominance of Occitanomys adroveri (e.g., 72% in the Gran- 62 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 Fig. 6. Similarities between rodent assemblages and turnover for individual basins for the proposed system of Iberian zones (see section 4.3.2 for formal definitions, and Tables S1, S3, Appendix A, for selected localities and data, respectively). A. Lines and squares: relative abundance of shared species (averages for basin pairs); numbers in squares: numbers of shared species. Abbreviations: AL, Alcoy Basin; BS, Bajo Segura Basin; GB, Guadix-Baza Basin; GR, Granada Basin; TE, Teruel Basin. B. Bars: rates of shared turnover (FO + LO) between East, South, and Central areas (Fig. 1); numbers above bars: numbers of shared species events; dashed lines: average rare-common threshold level based on quartile (lower line) and median (upper line) of ranked abundance distribution (see Section 3.5). ada Basin) affecting the relative abundances of characteristic (as A. sp. indet. and A. gudrunae, respectively), this genus turns taxa entering together such as O. alcalai, Castromys inflatus into one of the dominant forms in both upper zones (as A. aff. and Stephanomys. Despite the low STAR value, we regard the gorafensis/A. gorafensis, and A. gudrunae in South and Central, common entry of these three species and the overall magnitude respectively) with percentages of 50% and 31%. If this expansion of faunal change during the transition (Fig. 6(B)) as a positive of the A. gudrunae – A. gorafensis lineage would be counted as a indication for a single underlying event affecting all three fau- boundary ‘event’, shared abundance would be considerably nas approximately coevally; higher. On the other hand, A. gudrunae is already common  Three shared species and a STARS-C value of 14.9% characterize (19%) in basal ‘13a’ site SIF1 in the Fortuna Basin (Piñero and the ‘13a’–‘13b’ transition. It should be noted that the strong Agustí, 2019), which is not included in the similarity calcula- expansion of Apodemus is not represented as an event in the cal- tions, because no local zonation has yet been defined for this culation. Present with very low numbers in both lower zones basin; 63 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 Table 2 Chronology of Iberian rodent zone boundaries. Ages are calibrated to the time scale of Raffi et al. (2012). Uncertainty ranges are based on ages of bracketing or transitional localities indicated. Rodent Lower Zone Upper Zone Lower bracketing Upper bracketing site* Transitional site Max. Min. Uncertainty Zone site* age age (M.y.) boundary (Ma) (Ma) top Ib- Mimomys realensis Valdeganga IV FA Allophaiomys pliocaenicus 1.90 1.80 0.10 N17b Ib-N17a-b Mimomys Mimomys realensis Villarroyo Fuentenueva, Alquería Valdeganga II 2.15 2.05 0.10 medasensis (Cortez de Baza 1,6) Ib-N16b- Kislangia ischus Mimomys Concud Estación 2 Valdeganga I 2.56 2.54 0.02 17a medasensis Ib-N16a-b Mimomys hassiacus Kislangia ischus Gea 0 Zújar 11 3.06 3.03 0.03 Ib-N15b- Dolomys adroveri Mimomys hassiacus Escorihuela B Zújar 10 Escorihuela C 3.19 3.18 0.01 16a Ib-N15a-b Mimomys davakosi Dolomys adroveri Villalba Alta 4 Villalba Alta 2 4.01 3.92 0.09 Ib-N14c- Trilophomys castroi Mimomys davakosi Villalba Alta Rio 1 Zújar 4 4.27 4.19 0.08 15a Ib-N14b-c Paraethomys Trilophomys castroi Sifón de Librilla 4.71 4.51 0.20 baeticus P Ib-N14a-b Apocricetus barrierei Paraethomys Fuente del Viso La Bullana 3 (Peralejos E) 4.91 4.87 0.04 baeticus Ib-N13c- Paraethomys meini Apocricetus barrierei Zorreras Sifón de Librilla 413 (La 5.47 5.33 0.14 14a Alberca) Ib-N13b-c Apocricetus alberti Paraethomys meini Sifón de Librilla 52 Venta del Moro 6.26 6.25 0.01 Ib-N13a-b Castromys inflatus Apocricetus alberti El Búnker 4/5 6.56 6.27 0.29 Ib-N12b- Parapodemus meini Castromys inflatus El Búnker, El Chorrico 17 (Masada del 7.02 6.90 0.12 13a Crevillente 17 Valle 6) Ib-N12a-b Huerzelerimys Parapodemus meini Los Aljezares 2B La Celia 7.18 7.09 0.09 turoliensis Ib-N11- Huerzelerimys vireti Huerzelerimys Balneario Fuente Podrida, Concud B/L (Crevillente 4) 7.68 7.65 0.03 12a turoliensis base Ib- Huerzelerimys vireti La Gloria 11 Los Aguanaces 7 8.75 8.63 0.12 N11 * Sites in regular print: magneto/cyclostratigraphic constraint; sites in italic print: sedimentary or evolutionary rate constraint; sites in brackets: derived constraint (see text).  Shared species events at the ‘13c’–‘14a’ transition include the Iberian zones are applied to all studied areas within the Penin- LO of Apocricetus alberti and the FO of A. barrierei. The Granada sula except for the La Cerdanya Basin (Pyrenees). The MN13- Basin additionally shares the entry of Celadensia nicolae with the correlative fauna Can Vilella from this basin has an atypical, more Teruel Basin. Whereas this form represents the most abundant central European-type of rodent fauna that lacks typical Iberian rodent in the latter basin, it is very rare in the former, however. murines (Agustí et al., 2006b). The only murine reported is Apode- STAR values for the Granada-Teruel and Guadix-Baza-Teruel mus gudrunae, known from MN13-correlative sites on the Penin- comparisons are 26.1% and 5.1%, respectively, with their aver- sula. The second form with a distinct Iberian record is the age value shown in Fig. 6(B). (Pooling the raw data for both insectivoran Amblycoptus jessiae (Furió and Agustí, 2017). This sori- southern basins would have been another option; in this case cid enters the Teruel Basin record around the M1-2 transition and the resulting STAR value would have been approximately equal becomes especially abundant in Zone M2 (also MN13, e.g., Las to the highest value.) We finally note that Apocricetus is rare in Casiones; van Dam, 2004). In our view, the Iberian Zone system the Teruel Basin and that not all cricetine molars found in local- cannot be applied to Can Vilella, which should be placed in another ities Celadas 3 and 4B could be assigned with certainty to A. bar- regional (French/Central European) system. rierei (as positively identified in Celadas 4B; see Table S3, Opting for assemblage zones and following the international Appendix A). stratigraphic rules (Salvador, 1994), we name the new zones after the (or one of the) most commonly occurring species (Table 2). In Overlooking all the evidence on faunal similarity and shared order to preserve the aspect of chronological ranking, a parallel turnover, we consider the transfer of Iberian zone boundary ages system of codes was set up. These codes start with ‘Ib-N’ (Iberian across basins a viable step in the process of locality age estimation Neogene) and are followed by a number (11, 12, etc.) that roughly for our entire study period. corresponds to the time-equivalent MN unit, and a letter to indi- cate the actual Zone (Ib-N12a, Ib-N12b, etc.). For consistency rea- 4.3.2. Iberian rodent zones sons, we retain the number 17 as in MN17, although this unit The new regional zonation presented here (Table 2; Fig. 7) will now formally belongs to the Quaternary (Gradstein et al., 2012). also include MN17-correlative biozones, even though these were The use of MN unit numbers is only meant for convenience (i.e., defined in one basin only (Guadix-Baza Basin) and could therefore rapid chronological appreciation) and has no significance in terms not be subjected to faunal similarity analysis. The possibility of of strict correlation. We will now name and describe all biozones, applying these zones to a wider area is nevertheless suggested mention the most important qualitative and quantitative differ- by observations in the Júcar Basin (Valdeganga section, eastern ences between basins, and provide age uncertainty intervals for Spain), where the two name-giving species for the Guadix-Baza zone boundaries (Table 2). For the sake of brevity, we use short- zones (Mimomys medasensis and M. cf. reidi) appear in the same ened forms of the zone codes in the main text body, i.e., omitting order (Mein et al., 1978; Opdyke et al., 1997). the prefix ‘Ib-’. 64 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 Fig. 7. Chronological placement of Iberian micromammal localities with age uncertainty 200kyr. Shown are magnetostratigraphically calibrated sections (thick red lines), cyclostratigraphically calibrated sections (thick green lines), sections for which constant sedimentation rates are assumed (thin black lines), intervals with locality ages estimated otherwise (dashed lines), and Ib-N Zone age uncertainty ranges (grey bars). Ages are calibrated to the time scale of Raffi et al., 2020. For full locality names, see Table S1 (Appendix A). The Huerzelerimys vireti Zone (Ib-N11, early Turolian, late Tor- N12a site Aljezar 2B (ALJB) in the Teruel Basin, as inferred from tonian) is quantitatively dominated by small-sized murines. extrapolating evolutionary rates (M1 size) of Occitanomys adroveri Whereas Occitanomys sondaari is dominant in the Teruel Basin, (see Section 4.4.4) calibrated to cyclostratigraphically inferred Parapodemus lugdunensis and the cricetine Neocricetodon are more locality ages in the Concud sections (see Section 4.4.1). The upper important in the Bajo Segura Basin. The jumping mouse Eozapus limit (7.09 Ma) corresponds to the age of N12b site La Celia shows a modest but persistent presence in the Teruel Basin but (Jumilla-La Celia Basin; Section 4.4.1, options II/III). The marine is absent in the Bajo Segura Basin. Tortonian-Messinian boundary (7.25 Ma) correlates to the upper The base of N11 is dated in the Teruel Basin to lie between 8.75 part of N12a. (La Gloria 11) and 8.63 Ma (Los Aguanaces 7) based on magne- Assemblages belonging to the Castromys inflatus Zone (Ib- tostratigraphic results for the La Gloria – El Búnker sections N13a, late Turolian, early Messinian) strongly differ from those in (Krijgsman et al., 1996; van Dam et al., 2001). the previous Zone. Stephanomys is now well represented in central In the Huerzelerimys turoliensis Zone (Ib-N12a, middle Turo- Spain (Teruel and Madrid Basins) where it has replaced Occi- lian, late Tortonian), Parapodemus lugdunensis is replaced by the tanomys adroveri, while this latter form remains dominant in the larger-sized P. barbarae. Occitanomys (O. adroveri) loses its domi- South (Granada Basin). Presence of Stephanomys and absence of nance within murines. Within the cricetines, Neocricetodon is Occitanomys adroveri also characterizes the East (Fortuna Basin). replaced by Apocricetus aff. plinii in the Teruel Basin, and by Crice- At the same time, the locus of maximum abundance of Castromys tulodon meini-lucentensis lineage in the Bajo Segura and Fortuna moves northward (Teruel Basin). Basins. A lower limit of 6.88 Ma for the N12b-13a boundary follows The N11-N12a boundary is dated in the Cabriel Basin at 7.68– from the age of 6.87 Ma for N13a site El Chorrico 17 in the Fortuna 7.65 Ma based on the magnetostratigraphy and sedimentation rate Basin (Garcés et al., 1998) and the reservation of an additional of the Cabriel North section (Opdyke et al., 1989, 1997), with local- (minimum) duration of 0.01myr for basal N13a site Masada del ity Fuente Podrida placed at the base of C4n.1n (Table S4, Appendix Valle 6 (Teruel Basin). Although an explicit faunal list for El Chor- A). rico 17 has not yet been published, the locality is part of a set of The Parapodemus meini Zone (Ib-N12b, middle Turolian, early sites (Agustí et al., 2006a: Table 1) that is characterized by the Messinian) is characterized by the absence of Huerzelerimys. Of all presence of an advanced Castromys species, and of Apodemus zones, inter-basinal differences are the largest. Whereas Occi- gudrunae and Apocricetus alberti, forms that imply a faunal stage tanomys (O. adroveri) becomes even more dominant in the Teruel younger than Masada del Valle 6. This latter locality is considered Basin and the Parapodemus barbarae-Castromys lineage disappears, to represent a separate ‘faunal stage’ (see Section 3.4) between the the latter lineage becomes well represented in the Bajo Segura and ‘regular’ N12b and N13a faunal stages based on the absence of Occ- especially in the Granada Basin. itanomys alcalai, the presence of Parapodemus cf. lugdunensis (see The N12a-12b boundary is estimated to lie between 7.18 and Suppl. Text, Appendix A) and the insectivoran Archaeodesmana 7.09 Ma. The lower limit corresponds to the maximum age of luteyni (Rümke, 1985). Based on the assumption of a constant sed- 65 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 imentation rate in the Crevillente section (Bajo Segura Basin), the limit can be inferred based on magnetostratigraphy at Sifón de Lib- maximum age of the boundary can be constrained to 7.02 Ma rilla 413 and 430 (Fortuna Basin, Agustí et al., 2006a; Piñero and (maximum age for Crevillente 17, see Section 4.4.3). Agustí, 2019). Although the presence of Debruijnimys julii (known With Castromys disappearing and Occitanomys adroveri rarify- from N14b site La Bullana 2B and N14c site Baza) could suggest a ing, other murine taxa such as Occitanomys alcalai and Apodemus higher zone, the inclusion of SIF413 into Zone N14a is not only sup- gudrunae increase their relative abundance in the Apocricetus ported by its numerical age (5.27 Ma; Garcés et al., 1998; Agustí alberti Zone (Ib-N13b, late Turolian, middle Messinian). The crice- et al., 2006a; Piñero and Agustí, 2019), which would exclude tine Apocricetus (re-)appears in the various basins as the name- N14b and c, but also by the absence of a large-sized Paraethomys giving A. alberti. Although in small numbers, the entry of the lineage, of which the entry corresponds to the base of Zone cricetodontine Blancomys is noteworthy. N14b. With sites Sifón de Librilla 413 and 430 (5.13 Ma and The N13a-b boundary cannot be estimated very accurately. The 5.27 Ma, respectively) belonging to N14a, the N13c-14a boundary magnetostratigraphically constrained site El Búnker 4/5 (Teruel can be assumed to be older than 5.27 Ma. On the other hand, a Basin) is a ‘boundary locality’ (see Section 3.4), given the combined boundary older than the Mio-Pliocene boundary (5.33 Ma) can be presence of Castromys inflatus and the insectivoran immigrant inferred from the classical fauna of La Alberca, a site interbedding Amblycoptus jessiae. The fauna, with an age estimate between with marine Messinian sediments that can be assigned to N14a 6.46 and 6.37 Ma (Table S4, Appendix A), can be placed between (presence of Apocricetus barrierei). Here we conservatively set the typical N13a faunas with Castromys inflatus (El Búnker, La Gloria minimum age of the N13c-14a boundary to 5.33 Ma. 6) and N13b faunas (Las Casiones, Las Casiones Superior, Las The proximity of the younger limit of the N13c-N14a boundary Casiones Superior A, Villastar, Valdecebro 6, La Gloria 5) without to the Miocene-Pliocene boundary and marine Messinian-Zanclean this species but with Apocricetus alberti, Blancomys sanzi and boundary (5.33 Ma) leaves the possibility open that both bound- Amblycoptus jessiae (van Dam, 2004). Applying the ‘boundary fauna aries are (almost) coeval. The approximate correlation of N13c- convention’ (see Section 3.4), the age of N13a-b boundary is con- N14a boundary to the transition from European Mammal unit strained to 6.56–6.27 Ma. The younger limit is a very conservative MN13 to MN14 as implied by our zone numbering system thus estimate, because placing boundary at 6.27 Ma would leave almost also entails an approximate correlation of this MN transition to no space to accommodate N14b localities, given the age of the the Miocene-Pliocene boundary (5.33 Ma). However, it should be N13b-c boundary (see below). noted that such a correlation is based on the ‘faunistic interpreta- The composition of the Paraethomys meini Zone (Ib-N13c, late tion’ of the MN system. If the ‘stratigraphic interpretation’ of the Turolian, late Messinian) is like that of the previous zone, except MN system would be adopted (see Hilgen et al. (2012) for these for the addition of the African immigrant Paraethomys meini. two interpretations), the key defining event would be the First His- Whereas a low percentage of 5% of this species characterizes torical Appearance of Promimomys in Europe, which is dated to the Teruel Basin, larger shares of 20–30% characterize the southern 5.40–5.23 Ma in Greece (see Section 4.2.3). Although this age range and eastern basins. is comfortably close to that of the Ib-N13c-14a boundary, it has The entry of Paraethomys defining the N13b-c boundary is mag- also become clear that direct use of the Promimomys entry in Spain netostratigraphically constrained in the El Chorrico and Sifón de may lead to erroneous age correlations, because of its potential Librilla sections in the Fortuna basin (Garcés et al., 1998; Agustí diachroneity as evidenced by some rare occurrences in Zone et al., 2006a; Piñero and Agustí, 2019). Assuming constant sedi- N14b only. In this respect, the First Historical Appearance of Occi- mentation rates within polarity zones and ignoring sample size tanomys brailloni could be a better candidate to mark the transition effects, the base of N13c can be dated to lie between  6.26 Ma from MN13 to MN14 (Piñero and Verzi, 2020), especially because (SIF52) and 6.20 Ma (SIF61 = Librilla,; Gibert et al., 2013) in the this species is also recorded for the first time in MN14 faunas in Sifón de Librilla section, or between 6.15 Ma (AU2-r) and the eastern Mediterranean region (CORRECT) with a dated first 6.03 Ma (AU3-a) in the El Chorrico section. However, because Para- occurrence between 5.20 Ma and 5.04 Ma (Hordijk and de Bruijn, ethomys is already present in the 6.25 Ma old locality Venta del 2009). On the other hand, O. brailloni is absent from many N14a- Moro in the Cabriel Basin (Gibert et al., 2013), we use 6.26– b-c sites on the Iberian Peninsula where it starts to become a com- 6.25 Ma as the boundary age between N13b and c, while regarding mon resident only from N15a onwards (Table S1, Appendix A). the absence of Paraethomys in the AU2 levels as a sampling effect. After the rapid disappearance of Celadensia, compositions The composition of the Apocricetus barrierei Zone (Ib-N14a, become more homogeneous across the Iberian Basins in the Para- Turolian-Ruscinian transition, Messinian-Zanclean transition) is ethomys baeticus Zone (Ib-N14b, early Ruscinian, early Zanclian). not very different from that of the Paraethomys meini Zone, except Paraethomys baeticus, P. meini and Occitanomys alcalai are the most in the North (Teruel Basin), where the ‘microtoid cricetid’ immi- important species. Apocricetus is the most abundant lineage in the grant Celadensia nicolae constitutes 45% of the rodent fauna. The Cabriel Basin. murine Occitanomys brailloni is for the first time present in the East The N14a-b boundary is constrained to 4.91–4.87 Ma. The max- and South (Fortuna Basin, Guadix—Baza Basin, Caravaca), whereas imum age of 4.91 Ma is based on the age of 4.93 Ma of N14a local- hamsters (Ruscinomys, Blancomys, Apocricetus) increase their share ity Fuente del Viso (Cabriel Basin; Opdyke et al., 1997) lowered by in the South (Purcal 4, Granada Basin). The large percentage of an additional 0.01myr in order to accommodate the N14a-b Paraethomys meini in the Alcoy Basin (55%) is remarkable. boundary fauna Peralejos E (Teruel Basin). Similarly, the minimum The lower age limit of the N13c-14a boundary is represented by age of 4.87 Ma is constrained by the assumed minimum age of the localities Zorreras 2B and 3A in the Sorbas Basin, which are 4.85 Ma for the N14b site La Bullana 3 (Cabriel Basin), which is cor- magnetostratigraphically constrained to C3r (Martín-Suárez et al., related to C3n.3n (see Section 4.4.3 and Suppl. Text, Appendix A). 2000). With the immigrant Debruijnimys almenarensis and rela- The transition to the Trilophomys castroi Zone (Ib-N14c, early tively small-sized murines present, the Zorreras localities can be Ruscinian, early Zanclean) is characterized by a second immigra- confidently assigned to Zone N13c. Their age can be narrowed tion of ‘microtoid cricetids’ as represented by Trilophomys, a form down to 5.47 Ma by acknowledging their stratigraphic position of unknown cricetine origin not directly related to Celadensia within the ‘Zorreras Member’, a sediment package of 8 precession (Fejfar et al., 2011). In its way of expansion, Trilophomys mimics cycles that is time-equivalent to the latest Messinian Lago Mare Celadensia in becoming immediately, although temporary, domi- phase (Krijgsman et al., 2001) and which ends with the Pliocene nant in central Spain (50% in the Teruel Basin), while forming flooding event at 5.33 Ma. Ages well over 5 Ma for the upper age only a modest addition to the existing fauna in the South (5% 66 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 presence in the Alcoy and Guadix-Baza Basins; the latter basin wit- uel Basin, where it is associated with Kislangia ischus in local Zone nesses a distinct expansion of Apocricetus instead). O4, which correlates to N16b.). With Stephanomys cordii present, but Trilophomys still absent, The addition of the large-sized arvicoline Kislangia ischus marks Sifón de Librilla P (Fortuna Basin) can be regarded as a ‘boundary the transition to the Kislangia ischus Zone (Ib-N16b, late Vil- fauna’ between Zones N14b and 14c. We therefore take its magne- lanyian, late Piacenzian). In the Teruel Basin, the medium-sized tostratigraphically dated age of 4.61 Ma as an estimate for the cen- M. hassiacus-polonicus lineage remains dominant, while in the tral age of the N14b-c boundary, while adding ±0.1myr (4.71– Guadix-Baza Basin this position is filled by the smaller-sized Mim- 4.51 Ma, see convention in Section 3.4). omys gracilis-stehlini lineage, with the Stephanomys donnezani – In the Mimomys davakosi Zone (Ib-N15a, middle Ruscinian, thaleri still being well-represented (35%). middle Zanclean) Trilophomys is largely replaced by the arvicoline The N16a-16b boundary (3.05 Ma) is constrained between Mimomys davakosi. Instead of M. davakosi and Trilophomys, the 3.06 Ma, the age of Gea 0 (Teruel Basin, N16a, see Section 4.4.2) eastern fauna of La Juliana is still dominated by Paraethomys and 3.03 Ma, the age of Zújar 11 (Guadix-Baza Basin, N16b; Oms (two lineages). et al., 1999). The age of the N14c-15a boundary is constrained to lie between The highest two Zones are only represented quantitatively in 4.27 Ma and 4.19 Ma, based on magnetostratigraphic calibrations the Guadix-Baza Basin. In the Mimomys medasensis Zone (Ib- in the Villalba Alta Río section (Teruel Basin; Opdyke et al., 1997) N17a, late Villanyian, Gelasian), arvicolines now also reach domi- and Zújar section (Guadix-Baza Basin; Oms et al., 1999) with ages nance in the South, where they are especially well represented of 4.27 Ma for Villalba Alta Río 1 (N14c) and 4.19 Ma for Zújar 4 by the name-giving species. Kislangia is represented by K. aff. cap- (N15a, with Mimomys cf. davakosi). pettai and K. gusii. The large-sized arvicoline Dolomys adroveri replaces Mimomys The N16b-17a boundary is constrained to lie between 2.56 Ma davakosi in the Dolomys adroveri Zone (Ib-N15b; late Ruscinian, (Concud Estación 2, Teruel Basin) and 2.54 Ma. The latter age stems late Zanclean). It occupies an important share of the rodent fauna from our re-interpretation of the Valdeganga section with an in both the central (Teruel Basin) and southern (Guadix-Baza inferred age of 2.54 Ma for N17a level Valdeganga I (Section 4.4.2). Basin) parts of the Peninsula. Relatively heterogeneous Ib-N14c- The age of 2.44 Ma for N17a site Galera G (Guadix-Baza Basin; 15a faunas with Paraethomys and Apocricetus are replaced by a Garcés et al., 1997) is consistent with this interpretation. The poorly diversified fauna in the latter basin that, besides by Dol- N16b-17a boundary slightly post-dates the Pliocene-Pleistocene omys, is dominated by the murines Apodemus atavus and Castil- boundary and marine Piacenzian-Gelasian boundary (2.58 Ma). lomys. In contrast to the Teruel Basin, only one Stephanomys The transition to the Mimomys realensis Zone (Ib-N17b, latest lineage (S. donnezani – thaleri) is present and the proportion of Tri- Villanyian, Gelasian-Calabrian transition) is associated with lophomys is low. Likewise, the single representative fauna from the another expansion of arvicolines at the cost of Stephanomys. Other most western basin (Asta Regia 30 , Jerez Basin, Castillo and Agustí, arvicolines besides M. realensis (a form that probably includes M. 1996) shows the dominance of Apodemus atavus and Castillomys ‘cf. reidi’) reported from the Guadix-Baza Basin are M. aff. pliocaeni- and S. donnezani – thaleri at the cost of Dolomys. cus (descendant of M. polonicus) and Kislangia aff. cappettai. The age of the N15a-b boundary was based on the sequence of According to the magnetostratigraphic results of Garcés et al. localities in the Villalba Alta Río - Villalba Alta composite section (1997) for the Galera section (Baza Basin) the N17a-b boundary (Mein et al., 1990; Opdyke et al., 1997). A potential biostratigraphic is situated between Galera 2 (2.25 Ma) and Galera H (1.98 Ma). discrepancy characterizes the transitional sequence, however, with Furthermore, the magnetostratigraphic correlation of N17a site VA2 (3.92 Ma) containing N15a species Mimomys davakosi and VA4 Villarroyo (La Rioja) to a position in the Reunion chron (C2r.1r- (4.01 Ma) containing N15b species Dolomys adroveri (see Sec- 1n; 2.155–2.120 Ma; Pueyo et al., 2016) implies a maximum age tion 4.4.2; Table S4, Appendix A), implying an alternation of the of 2.15 Ma for Zone N17b. Using the lithostratigraphic correlation two arvicoline species during this short interval. As a conservative of Alquería (N17b) to the Galera section (Oms et al., 2000), the solution, we bracket the N15a-b boundary between 4.01 Ma and upper boundary can be lowered to 2.04 Ma. Although an alterna- 3.92 Ma. The marine Zanclean-Piacenzian boundary (3.60 Ma) cor- tive interpretation of the upper part of the Galera section was relates to the middle part of Zone N15b. put forward by others (Scott et al., 2007), our re-interpretations The Mimomys hassiacus-polonicus lineage replaces Dolomys in of the Valdeganga section (Júcar Basin; Mein et al., 1978; Opdyke the Mimomys hassiacus Zone (Ib-N16a, early Villanyian, early Pia- et al., 1997; Section 4.4.2) and Fuentenueva section (Guadix-Baza cenzian) in the Teruel Basin, where arvicolines replace murines as Basin, Scott et al., 2007) confirm the original Galera ages, with a the dominant group. Stephanomys (S. donnezani – thaleri) remains minimum age of 2.11 Ma for N17a/b transitional level VDG-II, important in the Guadix-Baza Basin. and an age of 2.05 Ma for N17b site Fuentenueva 1. Combining Based on Oms et al.’s (1999) magnetostratigraphic reinterpreta- all information, the uncertainty interval for the N17a-b boundary tion of Opdyke et al.’s (1997) results for the Escorihuela Section in can be set to 2.15–2.05 Ma. the Teruel Basin, the age of the N15b-16a boundary is constrained The top of the N17b Zone is chosen to coincide with the entry of to lie between Escorihuela B (3.19 Ma, N15b, with Dolomys adro- rootless voles, a widespread event commonly used to define the veri) and Escorihuela/Escorihuela A (3.11 Ma, N16a, with Mimomys Villanyian-Biharian boundary (e.g., Tesakov, 1998). Whereas Allo- hassiacus; see also Section 4.3.3). A third level, Escorihuela C phaiomys is generally considered to be the first rootless form to (3.16 Ma) can be regarded as a ‘boundary fauna’, because it lacks enter Europe, also other rootless forms (in Mimomys) have been both Dolomys and M. hassiacus (but contains M. gracilis). (Mein found in similarly aged sediments in the Guadix-Baza Basin et al. (1990), on the other hand, consider Escorihuela C to belong (Agustí, 1998; Oms et al., 2000). Given the diverging opinions on to the ‘Zone à M. gracilis + M. hajnackensis’, which is equivalent to local Guadix-Baza Basin stratigraphy (Oms et al., 2000, Scott N16a.) The magnetostratigraphically inferred age of 3.18 Ma for et al., 2007), we refrain from making a final calibration of these N16a site Zújar 10 in the Guadix-Baza Basin can be used to signif- sedimentary sequences to the GPTS (Orce section; see Suppl. Text, icantly narrow down the boundary to 3.19–3.18 Ma. Although Appendix A). Instead, we use the western European correlation of ZU10 does not contain Mimomys hassiacus, it contains M. cf. poloni- the Allophaiomys event to the Olduvai chron (C2n, 1.93–1.78 Ma) cus, which is viewed as its temporal successor (Oms et al., 1999; or just before (Repenning et al., 1990; Maul and Markova, 2007). Piñero et al., 2018). (We also note that the presence of Mimomys For the Iberian domain, the implied maximum age of 2.0– cf. polonicus in the Guadix-Basin pre-dates its presence in the Ter- 1.8 Ma for the top of Ib-N17b can be further reduced to 1.90– 67 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 1.8 Ma, given our estimated age of 1.90 Ma for N17b site Valde- age of 7.09 Ma for Ib-N12b site LCE. As will be shown in section ganga IV (see Section 4.4.2). Section 4.4.4, these options also allow several other N12a localities that are younger than CC3 to be properly placed. 4.4. Chronology of sections and sites Although a more detailed sedimentological study of the observed small-scale cycles would be necessary to establish a final 4.4.1. Tortonian-Messinian transition interpretation (‘tuning’) of the Las Pedrizas section, the following The chronology of the Tortonian-Messinian transition interval correlations can be hypothesized from a chronostratigraphic point in Spain was refined using (i) the new stratigraphical results in of view (Fig. 4(F)): the Jumilla-La Celia and Teruel Basins, and (ii) the ages of Iberian Zone boundaries Ib-N11-12a and N12a-12b.  The association of the featureless middle part (which we take to The magnetic polarity pattern of the combined La Celia and Los include cycles 13–21) to the interval encompassing nine preces- Barracones sections shows three reversed and two normal polarity sion minima with reduced amplitude corresponding to the 400- intervals (Figure 4 and 5). In addition, two very short normal polar- kyr/2.4-myr eccentricity minimum between 7.4–7.2 Ma; ity magnetozones were identified in the La Celia section at 42 m  The correlation between a possible deepening starting in cycle and 56 m. Whether these represent true geomagnetic features or 13 (end of interval with lignite streaks and root structures, lam- artifacts of the sediment remanence acquisition remains unknown inar microstructure in thin section LP0-1) with the transition to us. But given their short thickness, their relevance for correlation towards the first low-amplitude precessional minimum at with the GPTS is negligible since the time scale may be incomplete 7.386 Ma; for very short geomagnetic chrons.  The correlation of a possible shallowing starting in cycle 21–22 Especially the La Celia site is biostratigraphically relevant (van (reduced lamination, increase in the amount of well-preserved Dam et al., 2014). Although the exact spot of the La Celia site could gastropod fragments, and the appearance of palustrine features not anymore be recovered, its position directly N of the main road in the form of microscopic cracks and indurated bed tops) with near the road junction in the La Celia village (Agustí et al., 1985: an interval of more extreme precessional minima starting at their Fig. 7) leaves only one serious option for its stratigraphic posi- 7.190 Ma. tion as the one corresponding to a brown-colored (relatively organic-rich) mudstone horizon 5 m above the present main road The best match of the cyclostratigraphic pattern to the preces- (at 85 m in Fig. 4(E)). (Lower and higher levels are reddish-colored sional curve is the one in which the middle, deeper part is taken and presumably barren.) With the Parapodemus meini – Castromys maximally, i.e., including cycle 13 (instead of 14) and including littoralis lineage present, La Celia belongs to the Parapodemus meini transitional cycle 21 (as counted in Fig. 4(E), i.e., with four cycles Zone (Ib-N12b). By contrast, the lower positioned sites of Los Gar- in the ‘non-exposed’ interval). Downward counting of cycles yields gantones, which contain Huerzelerimys turoliensis and Parapodemus an age of 7.657 Ma for cycle 1 (precession minimum) and Ib-N12a barbarae, belong to the lower Huerzelerimys turoliensis Zone (Ib- localities CCL/CCB, i.e., just within the Ib-N11-12a boundary uncer- N12a). tainty range as inferred in the Cabriel Basin (7.68–7.65 Ma; Based on the polarity pattern alone, three options for the corre- Table S4, Appendix A). Downward extrapolation (using the same lation to the GPTS should be considered (Fig. 4(F)): sedimentation rate) yields an age well into long normal chron C4n.2n (7.73 Ma) for the uppermost normal sample in the CCCG I. Correlation of N1 to C3Br.2n and of N2 to C3Br.1n. The pat- section. We can therefore safely assume that our polarity interval tern of two short normal intervals separated by a longer N corresponds to chron C4n.2n (Fig. 4(E)). reversed fits the equivalent pattern in the GPTS. On the other Because of the various cyclostratigraphic uncertainties, we will hand, a strong increase in sedimentation rate (or hiatus) allow the whole sequence to be placed 1 or 2 precession cycles must be assumed to explain the absence of a third normal higher (Fig. 4(E)). As will be discussed in Section 4.4.4, the resulting interval in the upper part that would be equivalent to C3Bn; minimum age for the youngest Teruel Basin N12a locality still pre- II. Correlation of N1 to C3Br.2n (as in option I) and of the top of dates the next higher (Ib-N12a-b) Zone boundary. On the other N2 to the top of C3Bn. The long, reversed interval conve- hand, a lower placement by one cycle would imply an age of niently correlates to C3Ar, but either C3Br.1n or C3Br.1r is 7.675 Ma for the CCL/CCB localities (cycle 1), which is older than missing; the allowed minimum age of 7.66 Ma for a N12a site (given the III. Correlation of the base of N1 to the base of C4n.1n and the age of 7.68 Ma for N11 site Balneario and maximum age of top of N2 to the top of C3Bn (as in option II). This option is 7.67 Ma for N11-12a boundary site Crevillente 4). in accordance with the substantial length of N1 in the JCBA section, which seems too long to correspond to C3Br.2n. This 4.4.2. Magnetostratigraphic updates Teruel and Júcar Basins latter chron or CrBr.3r is missing. Besides the major update for the Tortonian-Messinian transi- tion, several smaller updates to the Iberian magnetostratigraphic In order to choose between the three magnetostratigraphic framework were performed (Table S4, Appendix A). The first two options, we additionally make use of (i) the time-equivalent record represent combinations of previously published results, the third in the Teruel Basin, and (ii) biochronology based on molar size. involves new/revised correlations of individual sites to existing Magnetostratigraphic option I for the JCLC and JCBA sections polarity patterns, whereas the last update combines previous pale- implies an age of 7.26 Ma for N12b site LCE, which at the same omagnetic and faunal results with a newly inferred regional zone time functions as a minimum age for N12a site CC3, positioned boundary age: at cycle 18 in the CCLP section (Fig. 4(A, E)). The presence of 17 cycles (350kyr) between N12a sites CCB/CCL and CC3 implies a 1. The correlation to the GPTS of the polarity pattern of the upper minimum age of 7.60 Ma for the former, in agreement with the part of the late Miocene El Búnker section (Teruel Basin; age of 7.68–7.65 Ma for the Ib-N11-12a boundary as established Krijgsman et al., 1996; van Dam et al., 2001) is confirmed by in the Cabriel Basin (Opdyke et al., 1989, 1997; Table S4, Appendix more recent work in the Cabriel Basin (Gibert et al., 2013). A). While option I leaves 0.04myr (three precession cycles, Fig. 4 The latter work includes an age estimate for key site Venta (E)) for the exact placing of the Las Pedrizas section, options II and del Moro that is incompatible with younger options for the Ter- III leave more space to accommodate the section given a younger uel Basin polarity pattern; 68 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 2. Opdyke et al. (1997) discussed two options (C2An.1n or C2An as because of its transitional character. While Occitanomys size in this a whole) for the correlation of the upper normal interval in the site is like that of O. sondaari from its N11 type population Torta- late Pliocene Gea section in the Teruel basin. The second option jada A, the murine composition is atypical, with Parapodemus can be eliminated, however, given the magnetochronologic re- entirely represented by P. gaudryi, a form only marginally present interpretation of the nearby Escorihuela section (Oms et al., in N11 in the Teruel Basin (e.g., Los Aguanaces 3; van Dam, 1997) 1999); and even more rarely present in Zone N12a (van Dam et al., 2001, 3. Teruel Basin sites Villalba Alta Río 1–4, Villalba Alta 5, Orrios 1, 2014). Our re-inspection of the cricetine species named ‘Kowalskia 3 and 4 were previously correlated to the GPTS based on polar- aff. fahlbuschi’ by de Bruijn et al. (1975) in Crevillente 4, revealed a ity patterns as recorded in the Villalba Alta Río - Villalba Alta morphological stage intermediate between Neocricetodon occiden- composite and Orrios (ORR) sections (Opdyke et al., 1997). talis (typical for N11) and Apocricetus (aff.) plinii, (typical for Here, we use the lithostratigraphic schemes from Mein et al. N12a, but with some occurrences in N11 as well). Although the (1990: their Figs. 4 and 5) to correlate sites Villalba Alta 1–4, correlation of Crevillente 4 to the composite section of Martín- Orrios 7, and Orrios 9C to the main sections (and hence to the Suárez and Freudenthal (1998) has turned out to be complicated GPTS) as well. Secondly, ages of Escorihuela/Escorihuela A and lithostratigraphically (Freudenthal et al., 1991), the site can be Escorihuela B in the late Pliocene Escorihuela section of van biostratigraphically placed between Cabriel Basin sites Balneario Dam et al. (2006) are slightly modified based on the combined (N11, 7.68 Ma) and Fuente Podrida (N12a, 7.65 Ma) (see also Suppl. stratigraphic information in Mein et al. (1990) and Opdyke et al. Text, Appendix A). (1997). Thirdly, locality ages for the nearby Concud-Estación The Arenas del Rey section in the SW part of the Granada Basin section (Opdyke et al., 1997) are estimated on the basis of the contains 15 localities (series of La Mina, La Dehesa and Calerico; sedimentation rate for C2An.1n as recorded in the nearby Gea Boné et al., 1978; Padial-Ojeda, 1986; García-Alix et al., 2008a) of section which the large majority belongs to Zones N13b-c. The highest site 4. The magnetostratigraphic correlation of the Valdeganga sec- (La Dehesa D5; Boné et al., 1978) contains Celadensia as a rare ele- tion in the Júcar Basin to the GPTS (Opdyke et al., 1997) is re- ment (Castillo et al., 1990) and is accordingly assigned to N14a. By interpreted in the light of more recent paleomagnetic results varying the position of the N13b-c and N13c-14a boundaries in the Guadix-Baza Basin (Oms et al., 1999), which, in combina- between locality levels, age ranges for all sites can be estimated tion with update 2 above, results in an 3.06–3.03 Ma age range under the assumption of a constant sedimentation rate. This proce- for the N16a-b boundary. Given this range, the lowest levels of dure resulted in an average age uncertainty of 0.14myr. Further- the Valdeganga section (Valdeganga 14–15, N16b) should cor- more, magnetobiostratigraphic work in the Purcal section in the relate to C2Ar.1n (3.03–3.58 Ma; Fig. 7). The remaining part of northeastern part of the basin (Martín Suárez et al., 1998) has the Valdeganga section, which mostly consists of samples of shown that the three superposed sites Purcal 3 (N13c), Purcal 4 reversed polarity, thus correlates to C2r (Matuyama chron; in and Purcal 7 (both N14a) are all positioned in a reversed chron, accordance with the interpretation of Opdyke et al. (1997). for which the only viable option is C3r (6.02–5.24 Ma). Age ranges Based on an upward extrapolation of the sedimentation rate, for these sites were constructed using the age range of 5.47– an age of 1.92 Ma for the uppermost normal sample (i.e., within 5.33 Ma for the N13c-14a boundary and assuming a constant sed- the Olduvai chron, C2n, 1.95–1.78 Ma) and 1.90 Ma for the imentation rate between 10–40 cm/kyr (this latter range was con- uppermost N17b mammal site Valdeganga IV is estimated. servatively chosen given observed rates of  20–25 cm/kyr in nearby lacustrine successions; Fernández and Soria, 1996). 4.4.3. Sections without first-order correlation to the GPTS Mansino et al. (2017) presented the stratigraphic positions of 16 Several sections in the Bajo Segura, Granada, Alcoy, and Teruel sites near Alcoy. The sequence starts with the sites of Alcoi Forn, Basin lack first-order magnetostratigraphic control, but can be which can be assigned to N13c-14a, and which are overlain by dated by combining regional biostratigraphy with local sedimenta- the main section of Gormaget Ravine, containing sites belonging tion rates. A summary of the results is presented below. (Details to N14b. The uppermost sites Alcoi 2C/2D can be assigned to can be found in the Suppl. Text and Table S4, Appendix A, which N14c based on the presence of Stephanomys cordii and Apocricetus includes age-depth calculations.). cf. angustidens. By varying the ages of the N14a-b, N14b-c and Although the Crevillente section (which consists of the Castro N14c-15a boundaries between their minimum and maximum val- and Crevillente Dam sections; de Bruijn et al., 1975; Freudenthal ues and assuming a constant rate of sedimentation, age ranges for et al., 1991) in the Bajo Segura Basin has no magnetostratigraphic all sites could be estimated. The average uncertainty amounts to record, locality ages can be constrained by combining stratigraphic 0.17myr. heights (Martín-Suárez and Freudenthal, 1998) with ages of zone The Lomas de Casares section in the Teruel Basin (Mein et al., boundaries (between N11, 12a, 12b, 13a and 13c). With the 1990) contains ‘middle Turolian’ (i.e., potentially N12a-b) levels N11b-12a boundary constrained to 7.68–7.65 Ma, the minimum (Barranco de Cuevas 1–4) in its basal part, and N14a (Lomas de age for N12b to 6.89 Ma, and the maximum age for N13c to Casares 2/5, 3) and N15a (Lomas de Casares 1) sites in its upper 6.25 Ma, ages ranges for sites can be estimated assuming a con- part. The age of latter site can be constrained biochronologically stant rate of sedimentation. This procedure results in an average to lie between the magnetostratigraphically constrained ages of locality age uncertainty of 0.19myr (Table S4, Appendix A). The Villalba Alta 3 (4.15 Ma) and 1 (4.04 Ma). On the basis of a maxi- age estimate of 7.02–6.89 Ma thus inferred for N12b site Crevil- mum age of 7.67 Ma for the Barranco de Cuevas sites, a maximum lente 17 adds in constraining the maximum age for N13a locality age of 4.14 Ma for Lomas de Casares 1, a minimum age of 4.87 Ma BUN in the El Búnker section (Teruel Basin) to 7.00 Ma. (Because for Lomas de Casares 2/5 (because older than N14b site La Bullana of a low paleomagnetic sampling density and sedimentation rate, 3), and the assumption of a constant sedimentation rate in the considerable uncertainty in maximum and minimum age for El Lomas de Casares section, an age range of 5.20–5.07 Ma for Lomas Búnker and El Búnker 4/5 was inferred by allowing chron boundary de Casares 3 is calculated, which is consistent with the inferred levels to vary between bracketing paleomagnetic samples maximum age range for N14a (5.46 Ma). The above calculations (Table S4, Appendix A)). partly depend on the correlation of the Ib-N14b sites of La Bullana The locality Crevillente 4 (de Bruijn et al., 1975), which was not (3 and 2B) to C3n.3n and C3n.2r, respectively. As further explained included in later systematic papers by Freudenthal, Martín-Suárez in the supplementary text (Appendix A), we regard a still younger and colleagues, is relevant from a stratigraphic point of view correlation of the La Bullana sites to C3n.2n and C3n.1r (Piñero and 69 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 Agustí, 2020) unlikely, as it is incompatible with the age of the ermost CCB/CCL level only contains seven measurable O. adroveri N14b-c boundary as defined by the age of boundary fauna Sifón M1, we used CCCG (7.57–7.52 Ma, n = 20) and CC3 (7.28– de Librilla P. 7.24 Ma, n = 57) as our primary calibration points for further age inter- and extrapolation (Table 3). As a result, three of the four above mentioned sites have predicted ages of 7.27-7.26 Ma, 4.4.4. Localities with ages inferred from evolutionary rates whereas the estimated age for Aljezar B is clearly younger (7.18– No tooth element category successfully passed both the ‘zone 7.14 Ma). A similar result is implied by the size of M2, which also test’ and ‘superposition test’ using the combined data for the three points to Aljezar B as clearly being the youngest N12a site zones Ib-N12a, 12b and 13c (for data and analysis, see Table S5, (Table S5, Appendix A). We therefore conclude that magnetostrati- Appendix A). However, when the data were confined to N12a graphic option I for the Jumilla-La Celia Basin (implying a mini- (which contains most sites), the sequence of M1 average length mum age of 7.27 Ma for N12a) is not the most probable option. measurements (n  10) passed both tests (Table S5, Appendix A). The rejection of option I implies that the short reversed chron (Lowering the sampling threshold to n  5 allowed more localities C3Br.3r is not recorded, and that N1 represents an amalgamation to be included but led to violations in both tests.) Whereas the plot of C4n.1n and C3Br.2n. In choosing between options II and II, we of M1 length against calibrated age showed a monotonously prefer the former, given the substantial length of interval N1 in increasing trend, trends in other elements (M2, m2, and to a lesser the Los Barracones section. degree m1) as well as in the first principal component PC1 of all The application of the ‘best-fit’ cyclostratigraphic scenario for four size variables, showed a more erratic pattern including a the Las Pedrizas section (see Section 4.4.1), which at the same time strong size difference around 7.5 Ma. (Fig. S4, Appendix A). Such represents the oldest possible solution, leads to a secondary incon- an increase is difficult to explain given the overall taxonomic sta- sistency with regard to N12a level Tortajada C/D, for which a size- bility across the whole zone. based age range of 7.70–7.65 Ma is estimated (Table 3), which The more irregular trend in second molars could be related to partly exceeds the maximum age of 7.68 Ma for the N11-N12a their larger natural variability. In his study of the lower dentition boundary. Moreover, an even older age for Ib-N12a locality Torta- in 17 different mammal groups, Gingerich (1974) showed that jada B follows from its position  7 m below Tortajada C/D (van within-population size is the least variable in first molars, the pos- Dam et al., 2001). Statistical error may play a role, because the lat- sible reason being their early development preventing the expres- ter level just contains the minimum required amount of 10 speci- sion of sexual dimorphism. In their review of upper and lower mens. Other elements (m1, m2) appear to point an age younger tooth length and width variation in Cenozoic European hamsters than that of Crevillente 15, which is constrained to maximally and true mice, Freudenthal and Cuenca Bescos (1984) and 7.61 Ma on the basis of sedimentation rates (see Section 4.2.3). Freudenthal and Martín-Suárez (1990) find the lowest variation Because we give priority to the sedimentation rate method over in M1, m1 and m2 length. Additional variability in M2 in murines the size-based method, while accepting 7.61 Ma as a maximum, could be due to the large difference in measured occlusal length the average age of Tortajada C/D is (slightly arbitrarily) set to between worn and unworn specimens and/or to the use of slightly 7.61 Ma as well (7.63–7.58 Ma), thereby leaving space to accom- different conventions regarding orientation (e.g., compare van de modate lower site Tortajada B within N12a (set at 7.66–7.61 Ma). Weerd (1976) and Martín-Suárez and Freudenthal (1993)). Here, An age close to the N11-12a boundary of the Tortajada sites is con- we opted for M1 length as our final metric for age inter- and sistent with the virtual absence of the hamster Hispanomys (only extrapolation. This choice differs from the one in van Dam one specimen in Tortajada C), a feature shared with the other basal (1997), who applied a measure based on all four elements (princi- N12a sites CCB and CCL (7.66–7.62 Ma). The extension of the range pal component 1, PC1) to a more restricted set of sites. of the typical N11 hamster Neocricetodon occidentalis into the basal Four Ib-N12a sites (Los Mansuetos, Concud Cerro de la Garita 5, part of N12a (CCL; see Suppl. Text, Appendix A) also points to a Tortajada and Aljezar B) have a larger Occitanomys adroveri M1 size short transitional stage in the hamster community around the than Concud 3 (CC3), the highest locality in the cyclostratigraphi- N11-12a boundary. This transition is time-equivalent to the cally calibrated Concud series (see Section 4.4.1). Because the low- Table 3 Size-based age estimations for localities with ten or more M1 of Occitanomys adroveri, using cyclostratigraphically inferred ages for CG5 and CC3 as tie points. Locality Code Basin Ib-Zone Tie points Mean L(M1) N Size-based Reference measurements (Ma) age (Ma) max. min. max. min. Valdecebro 5 VDC5 Teruel N12b 2.100 32 Adrover et al. (1986) Crevillente 14 CR14 Bajo Segura N13a 2.094 13 Martín-Suárez and Freudenthal (1993) Crevillente 17 CR17 Bajo Segura N12b 2.080 48 Martín-Suárez and Freudenthal (1993) Los Aljezares B ALJB Teruel N12a 2.060 80 7.18 7.14 Adrover (1986) La Roma 5A R5A Teruel N12b 2.044 14 this paper Tortajada TO Teruel N12a 2.039 17 7.26 7.22 van de Weerd (1976) Concud Cerro de la Garita 5 CG5 Teruel N12a 2.037 10 7.27 7.22 this paper Los Mansuetos LM Teruel N12a 2.036 68 7.27 7.23 van de Weerd (1976) Concud 3* CC3 Teruel N12a 7.283 7.242 2.032 57 7.28 7.24 van de Weerd (1976) La Roma 5B R5B Teruel N12a 1.994 23 7.42 7.38 this paper Masada del Valle 5 MDV5 Teruel N12a 1.988 24 7.44 7.40 van de Weerd (1976) Cubla 2 CU2 Teruel N12a 1.962 30 7.53 7.49 Besems and van de Weerd (1983), this paper Villalba Baja 2 VB2 Teruel N12a 1.954 18 7.56 7.52 van de Weerd (1976); this paper Concud Cerro de la Garita* CCCG Teruel N12a 7.570 7.524 1.952 20 7.57 7.52 van de Weerd (1976); this paper Casa del Acero ACERO Fortuna N12a 1.950 20 7.58 7.53 Agustí et al. (1981) Masada del Valle 2 MDV2 Teruel N12a 1.938 69 7.62 7.57 van de Weerd (1976) Crevillente 15 CR15 Bajo Segura N12a 1.933 24 7.64 7.59 Martín-Suárez and Freudenthal (1993) Tortajada C/D TOC/D Teruel N12a 1.917 10 7.70 7.65 van de Weerd (1976); this paper * Used for calibration. 70 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 remarkable change in the true mice community as recorded across Declaration of Competing Interest the N11-12a boundary and in transitional fauna Crevillente 4 (see Section 4.4.3). The authors declare that they have no known competing finan- The age of 7.27–7.22 Ma as predicted for CG5 based on molar cial interests or personal relationships that could have appeared size implies the presence of a hiatus between CCCG (7.57– to influence the work reported in this paper. 7.52 Ma) and CG5 (Fig. 4(E)). A period of non-deposition is feasible given the position of the section close to the Mesozoic basement Acknowledgments (Fig. 3) in combination with a rather heterogeneous lithology in the upper part of the section (including white and grey nodular We dedicate this paper to Hans de Bruijn, who sadly passed limestones, lignitic clays and grey marls). away in 2021. Hans was one of the founding fathers of micromam- We additionally examined the possibilities of using size mal paleontology and a principal initiator of systems of rodent- increase in the long-lasting lineage Stephanomys ramblensis – S. based biostratigraphy in Spain and many other areas. We thank dubari – S. cordii – S. margaritae – S. vandeweerdi – S. minor – S. pri- Raef Minwer Barakat for allowing us to use his quantitative data etaensis – S. balcellsi – S. ‘progressus’ (Adrover, 1986) for dating sites for Guadix-Baza Basin rodents, Everett Lindsay for his information belonging to the latest Miocene (Zones N13a,b,c) and Pliocene on the lithostratigraphy of sections in the Teruel Basin, and Plinio (mostly data from Zones 14a-b). Upper first molars were selected Montoya for his help in clarifying the stratigraphy of La Bullana. in order to include the four calibrated sites Sifón de Librilla 3, 61, We thank Alexey Tesakov for sharing his insights on Dolomys and P, and Venta del Moro (which all contain eight M1, which is more other arvicoline lineages. We acknowledge Karel Steensma for his than the frequency of other elements). After compiling data for 21 invaluable help during the various field campaigns, and Lorenzo sites, the size ranking of localities suffered from violations of the Vilas and Cayetano Herrero for discussions during field work in zone test, with size values in southern basins often being lower the La Celia area. Eduardo Espílez and Luis Mampel are thanked than in the Teruel Basin (not shown). Furthermore, as shown in for taking rock samples for sedimentology, and João Trabucho, Jose the case of Occitanomys, the method can be assumed to work best Pedro Calvo and Ana Alonso for their help in characterizing the within zones (i.e., within the context of a relatively stable commu- sedimentary paleoenvironment of the Las Pedrizas section, Anita nity). Unfortunately, in the case of Stephanomys, no zone satisfied van Leeuwen for carrying out XRD analyses at Utrecht University, the combined requirement of a sufficient number of specimens Dolors Barsó for preparing the thin sections at the Faculty of Earth and a minimum of two calibrated tie points for the size-based Sciences at University of Barcelona, and Jasper Huijsmans for his approach to be successfully applied. assistance with microscopy. We thank the three reviewers for their Finally, locality rankings based on molar size trends in another comments and Frits Hilgen for his valuable feedback regarding the potentially suitable lineage, Apodemus gudrunae – A. gorafensis structure of the manuscript and the Teruel Basin cyclostratigraphy. (N13a-15a) yielded too many zone inconsistencies in order to be This work was funded by MCIN/AEI/10.13039/501100011033 (R + reliably used for chronological purposes. Local differences could D + I project PID2020-117289GB-I00), AEI-FEDER EU (to M.F.), have played a role here given the relatively early (Zone ‘13a’) AGAUR (Consolidated Research Group, 2017 SGR 960 to M.F. and appearance of A. gorafensis in the Fortuna Basin (Piñero and 2017 SGR 596 to M.G.), the Generalitat de Catalunya/CERCA Pro- Agustí, 2019), casting some doubts on the validity of a simple gram (to J.A.v.D. and M.F.), and the Department of Science, Univer- ancestor–descendant relationship (Martín-Suárez and Mein, sity and Society of Knowledge of the Government of Aragon 1998; García-Alix et al., 2008b). (Reference Research Group E04_20R FOCONTUR funding to Fun- dación Dinópolis-L.A.). We also thank the support of Department of Education, Culture and Sport of the Government of Aragon, 4.4.5. Remaining localities and its General Directorate of Cultural Heritage for the permissions Age ranges for localities that could not be constrained by sedi- to excavate in several paleontological sites of Teruel. mentary or evolutionary rates were taken maximally with regard to the (maximum) age range for their zone or faunal stage (for Appendix A. Supplementary material an overview of sites per basin, see Suppl. Text, Appendix A). Supplementary material to this article can be found online at https://doi.org/10.1016/j.geobios.2023.01.001. 5. Conclusions References A review of 8.5–2 Ma-old rodent faunas and calibrated sections across the Iberian Peninsula has resulted in a new regional bio- Abdul Aziz, H., van Dam, J.A., Hilgen, F., Krijgsman, W., 2004. Astronomical forcing in zonation with estimated age uncertainty ranges for boundaries Upper Miocene continental sequences: implications for the Geomagnetic Polarity Time Scale. Earth and Planetary Science Letters 222, 243–258. and micromammal sites. Although the magnitude of between- Abels, H.A., Abdul Aziz, H., Calvo, J.P., Tuenter, E., 2009a. Shallow lacustrine zone turnover differs from boundary to boundary, approximate microfacies document orbitally paced lake-level history in the Miocene Teruel isochrony for most Iberian Zone boundaries is suggested by shared Basin (NE Spain). Sedimentology 56, 399–419. Abels, H.A., Aziz, H.A., Ventra, D., Hilgen, F.J., 2009b. Orbital climate forcing in relative abundance patterns for both the faunal assemblages and mudflat to marginal lacustrine deposits in the Miocene Teruel Basin (Northeast their turnover. New stratigraphic and paleontological data for the Spain). Journal of Sedimentary Research 79, 831–847. Tortonian-Messinian transitional interval (8–7 Ma) in the Teruel Adrover, R., Alcalá, L., Mein, P., Moissenet, E., Orrios, J., 1986. Mamíferos del Turoliense medio en la Rambla de Valdecebro, Teruel. Estudios Geológicos 42, and Jumilla – La Celia Basin have significantly improved the age 495–509. control for this thus far poorly calibrated interval on the Iberian Adrover, R., Mein, P., Moissenet, E., 1993. Roedores de la transición Mio-Plioceno de Peninsula. la región de Teruel. Paleontologia i Evolució 26–27, 47–84. Adrover, R., 1986. Nuevas faunas de roedores en el Mio-Plioceno continental de la región de Teruel (España). Interés bioestratigrafica y paleoecologico. Instituto de Estudios Turolenses de la Excma. Diputacion Provincial de Teruel, Teruel. Data availability Agustí, J., 1998. A review of Late Pliocene to Early Pleistocene arvicolid evolution in Spain. Paludicola 2, 8–15. Agustí, J., Gibert, J., Moyà-Solà, S., 1981. Casa del Acero; Nueva fauna turoliense de Research data are included in the Supplementary Tables for this vertebrados (Mioceno superior de Fortuna, Murcia). Butlleti Informatiu - article Institut de Paleontologia de Sabadell 13, 69–87. 71 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 Agustí, J., Moyà-Solà, S., Gibert, J., Guillén, J., Labrador, M., 1985. Nuevos datos sobre Freytet, P., Verrecchia, E.P., 2002. Lacustrine and palustrine carbonate petrography: la bioestratigrafía del Neógeno continental de Murcia. Paleontologia i Evolució an overview. Journal of Paleolimnology 27, 221–237. 18, 83–93. Furió, M., Agustí, J., 2017. Latest Miocene insectivores from Eastern Spain: Evidence Agustí, J., Cabrera, L., Garcés, M., Krijgsman, W., Oms, O., Parés, J.M., 2001. A for enhanced latitudinal differences during the Messinian. Geobios 50, 123–140. calibrated mammal scale for the Neogene of Western Europe; State of the art. Gamonal, A., Mansino, S., Ruiz-Sánchez, F.J., Crespo, V.D., Corbí, H., Montoya, P., Earth-Science Reviews 52, 247–260. 2018. Sierra del Colmenar 1A, a new late Messinian (Late Miocene) locality in Agustí, J., Garcés, M., Krijgsman, W., 2006a. Evidence for African-Iberian exchanges the Bajo Segura Basin (SE Spain): biostratigraphic and palaeoenvironmental during the Messinian in the Spanish mammalian record. Palaeogeography, implications. Historical Biology 30, 380–381. Palaeoclimatology, Palaeoecology 238, 5–14. Garcés, M., Agustí, J., Cabrera, L., Parés, J.M., 1996. Magnetostratigraphy of the Agustí, J., Oms, O., Furió, M., Perez-Vila, M.J., Roca, E., 2006b. The Messinian Vallesian (Late Miocene) in the Vallès-Penedès Basin (northeast Spain). Earth terrestrial record in the Pyrenees: the case of Can Vilella (Cerdanya Basin). and Planetary Science Letters 142, 381–396. Palaeogeography, Palaeoclimatology, Palaeoecology 238, 179–189. Garcés, M., Agustí, J., Parés, J.M., 1997. Late Pliocene continental Alcalá, L., 1994. Macromamíferos Neógenos de la Fosa de Alfambra-Teruel. Instituto magnetochronology in the Guadix-Baza Basin (Betic Ranges, Spain). Earth and de Estudios Turolenses - Museo Nacional de Ciencias Naturales, Teruel. Planetary Science Letters 146, 677–687. Alcalá, L., van Dam, J.A., Luque, L., Montoya, P., Abella, J., 2005. Nuevos mamíferos Garcés, M., Krijgsman, W., Agustí, J., 1998. Chronology of the late Turolian deposits vallesienses en Masía de La Roma (Cuenca de Teruel). Geogaceta 37, 199–202. of the Fortuna basin (SE Spain): implications for the Messinian evolution of the Alcalá, L., Luque, L., van Dam, J.A., Rubio, J.C., 2018. Cuenca de Teruel y Fosa del eastern Betics. Earth and Planetary Science Letters 163, 69–81. Jiloca. In: Alcalá, L., Calvo, J.P., Simón, J.L. (Eds.), Geología de Teruel. Instituto de Garcés, M., Krijgsman, W., van Dam, J.A., Calvo, J.P., Alcalá, L., Alonzo Zarza, A., 1999. Estudios Turolenses, Teruel, pp. 99–129. Late Miocene alluvial sediments from the Teruel area: Magnetostratigraphy, Alonso-Zarza, A.M., Calvo, J.P., 2000. Palustrine sedimentation in an episodically magnetic susceptibility, and facies organization. Acta Geológica Hispanica 32, subsiding basin: the Miocene of the northern Teruel Graben (Spain). 171–184. Palaeogeography, Palaeoclimatology, Palaeoecology 160, 1–21. García-Alix, A., Minwer-Barakat, R., Martín-Suárez, E., Freudenthal, M., 2007. A late Alonso-Zarza, A.M., Meléndez, A., Martín-García, R., Herrero, M.J., Martín-Pérez, A., Ruscinian (Pliocene) rodent fauna from the Granada Basin (SE Spain). Journal of 2012. Discriminating between tectonism and climate signatures in palustrine Vertebrate Paleontology 27, 1066–1070. deposits: Lessons from the Miocene of the Teruel Graben, NE Spain. Earth- García-Alix, A., Minwer-Barakat, R., Martín-Suárez, E., Freudenthal, M., 2008a. Science Reviews 113, 141–160. Biostratigraphy and sedimentary evolution of Late Miocene and Pliocene Álvarez Sierra, M.A., Daams, R., Lacomba, J.I., López-Martínez, N., Sacristán-Martín, continental deposits of the Granada Basin (southern Spain). Lethaia 41, 431– M.A., 1987. Succession of micromammal faunas in the Oligocene of Spain. 446. Münchener Geowissenschaftliche Abhandlungen A 10, 43–48. García-Alix, A., Minwer-Barakat, R., Martín-Suárez, E., Freudenthal, M., 2008b. Anadón, P., Moissenet, E., Simón, J.L., 1990. The Neogene grabens of the Eastern Muridae from the Mio-Pliocene boundary in the Granada basin (Southern Iberian Caín (eastern Spain). Paleontologia i Evolució, Memória Especial 2, 99– Spain). Biostratigaphic and phylogenetic implications. Neues Jahrbuch für 130. Geologie und Paläontologie Abhandlungen 248, 183–215. Besems, R.E., van de Weerd, A., 1983. The Neogene rodent biostratigraphy of the Gaston, 1994. Rarity. Chapmann & Hall, London. Teruel-Ademuz Basin (Spain). Proceedings of the Koninklijke Nederlandse Gibert, L., Scott, G., Montoya, P., Ruiz-Sánchez, F., Morales, J., Luque, L., Abella, J., Akademie van Wetenschappen, Series B 86, 17–23. Lería, M., 2013. Evidence for an African-Iberian mammal dispersal during the Boné, E., Dabrio, C.J., Michaux, J., Pena, J.A., Ruiz, B.A., 1978. Stratigraphie et pre-evaporitic Messinian. Geology 41, 691–694. paleontologie du Miocène supérieur d’Arenas del Rey, Bassin de Grenade Gingerich, P.D., 1974. Size variability of the teeth in living mammals and the (Andalousie, Espagne). Bulletin de la Societe Belge de Geologie 87, 87–99. diagnosis of closely related sympatric fossil species. Journal of Paleontology 48, Castillo, C., 1990. Paleocomunidades micromamímiferos de los yacimientos 895–903. karsticos del Neógeno Superior de Andalucia oriental. Tesis Doctoral, Godoy, A., Olivé, A., Moissenet, E., 1983. Mapa y memoria explicativa de la Hoja 567 Universidad de Granada, 255p. (Teruel) del Mapa Geológico de España, a escala 1:50.000 (2a Serie). IGME, Castillo, C., Agustí, J., 1996. Early Pliocene rodents (Mammalia) from Asta Regia Madrid. (Jerez basin, Southwestern Spain). Proceedings Koninklijke Nederlandse Gradstein, F.M., Ogg, J.G., Schmitz, M.D., Ogg, G., 2012. A geological time scale 2012. Akademie van Wetenschappen, Serie B 99, 25–43. Elsevier, Amsterdam. Castillo, C., Freudenthal, M., Martín Suárez, E., Martínez, M.Y., Rivas, P., 1990. New Hilgen, F.J., Lourens, L.J., van Dam, J.A., 2012. The Neogene Period. In: Gradstein, F. localities with fossil micromammals in the Pliocene of Granada basin (Spain). M., Ogg, J.G., Schmitz, M.D., Ogg, G. (Eds.), A geological time scale 2012. Elsevier, Scripta Geologica 93, 41–46. Amsterdam, pp. 923–978. Daams, R., 1983. The dental pattern of the dormice Dryomys, Myomimus, Hordijk, K., de Bruijn, H., 2009. The succession of rodent faunas from the Mio/ Microdyromys and Peridyromys. Utrecht Micropaleontological Bulletins. Special Pliocene lacustrine deposits of the Florina-Ptolemais-Servia Basin (Greece). Publication 3, 1–115. Hellenic Journal of Geosciences 44, 21–103. Dahlmann, T., 2001. Die Kleinsäuger der unter-pliozänen Fundstelle Wölfersheim in Koufos, G.D., Vasileiadou, K., 2015. Miocene/Pliocene mammal faunas of southern der Wetterau (Mammalia: Lipotyphla, Chiroptera, Rodentia). Courier Balkans: implications for biostratigraphy and palaeoecology. Palaeobiodiversity Forschungsinstitut Senckenberg 227, 1–129. and Palaeoenvironments 95, 285–303. Daxner-Höck, G., 1980. Rodentia (Mammalia) des Eichkogels bei Mödling Krijgsman, W., Langereis, C.G., Daams, R., van der Meulen, A.J., 1994. (Niederösterreich). 1. Spalacinae und Castoridae. 2. Übersicht über die Magnetostratigraphic dating of the middle Miocene climate change in the gesamte Nagetierfauna. Annalen des Naturhistorischen Museums in Wien 83, continental deposits of the Aragonian type area in the Calatayud-Teruel basin 135–152. (Central Spain). Earth and Planetary Science Letters 128, 513–526. de Bruijn, H., Mein, P., Montenat, C., van de Weerd, A., 1975. Corrélations entre les Krijgsman, W., Garcés, M., Langereis, C.G., Daams, R., van Dam, J.A., van der Meulen, gisements de rongeurs et les formations marines du Miocène terminal A.J., Agustí, J., Cabrera, L., 1996. A new chronology for the middle to late d’Espagne méridionale; provinces d’Alicante et de Murcia. Proceedings of the Miocene continental record in Spain. Earth and Planetary Science Letters 142, Koninklijke Nederlandse Akademie van Wetenschappen, Series B 78, 282–313. 367–380. Dijksman, A.A., 1977. Georeversals as recorded in the Miocene redbeds of the Krijgsman, W., Fortuin, A.R., Hilgen, F.J., Sierro, F.J., 2001. Astrochronology for the Calatayud-Teruel Basin (Spain). Ph.D. Thesis, Utrecht University, 156p. Messinian Sorbas basin (SE Spain) and orbital (precessional) forcing for Ezquerro, L., Luzón, A., Simón, J.L., Liesa, C.L., 2020. Segmentation and increasing evaporite cyclicity. Sedimentary Geology 140, 43–60. activity in the Neogene-Quaternary Teruel Basin rift (Spain) revealed by Legendre, S., Bachelet, B., 1993. The numerical ages; a new method of datation morphotectonic approach. Journal of Structural Geology 135, 104043. applied to Paleogene mammalian localities from Southern France. Newsletters Fejfar, O., Heinrich, W.D., Kordos, L., Maul, L.C., 2011. Microtoid cricetids and the of Stratigraphy 29, 137–158. early history of arvicolids (Mammalia, Rodentia). Palaeontologia Electronica 14, Mansino, S., Fierro, I., Tosal, A., Montoya, P., Ruiz-Sánchez, F.J., 2017. Micromammal 27A. biostratigraphy of the Alcoi Basin (eastern Spain): remarks on the Pliocene Fernández, J., Soria, J., 1996. Evolución sedimentaria en el borde norte de la record of the Iberian Peninsula. Geologica Acta 15, 121–134. Depresión de Granada a partir del Turoliense terminal. Acta Geológica Martín Suárez, E., Oms, O., Freudenthal, M., Agustí, J., Parés, J.M., 1998. Continental Hispanica 1996, 73–81. Mio-Pliocene transition in the Granada Basin. Lethaia 31, 161–166. Franzen, J.L., Pickford, M., Costeur, L., 2013. Palaeobiodiversity, palaeoecology, Martín-Suárez, E., Freudenthal, M., 1993. Muridae (Rodentia) from the Lower palaeobiogeography and biochronology of Dorn-Dürkheim 1 – a summary. Turolian of Crevillente (Alicante, Spain). Scripta Geologica 103, 65–118. Palaeobiodiversity and Palaeoenvironments 93, 277–284. Martín-Suárez, E., Freudenthal, M., 1998. Biostratigraphy of the continental upper Freudenthal, M., Cuenca Bescos, G., 1984. Size variation of fossil rodent populations. Miocene of Crevillente (Alicante, SE Spain). Geobios 31, 839–847. Scripta Geologica 76, 1–28. Martín-Suárez, E., Mein, P., 1998. Revision of the genera Parapodemus, Apodemus, Freudenthal, M., Lacomba, J.I., Martín-Suárez, E., Peña, J.A., 1991. The marine and Rhagamys and Rhagapodemus (Rodentia, Mammalia). Geobios 31, 87–97. continental upper Miocene of Crevillente (Alicante, Spain). Scripta Geologica 96, Martín-Suárez, E., Freudenthal, M., Krijgsman, W., Rutger, F.A., 2000. On the age of 1–8. the continental deposits of the Zorreras Member (Sorbas Basin, SE Spain). Freudenthal, M., Martín-Suárez, E., 1990. Size variation in samples of fossil and Geobios 33, 505–512. Recent murid teeth. Scripta Geologica 93, 1–34. Maul, L.C., Masini, F., Abbazzi, L., Turner, A., 1998. Geochronometric application of Freudenthal, M., Mein, P., Martín-Suárez, E., 1998. Revision of Late Miocene and evolutionary trends in the dentition of fossil Arvicolidae. In: van Kolfschoten, T., Pliocene Cricetinae (Rodentia, Mammalia) from Spain and France. Treballs del Gibbard, P.L. (Eds.), The dawn of the Quaternary; proceedings of the SEQS- Museu de Geologia de Barcelona 7, 11–93. EuroMam symposium 1996. Nederlands Instituut voor Toegepaste Geowetenschappen TNO, Haarlem, Netherlands, pp. 565–572. 72 J.A. van Dam, P. Mein, M. Garcés et al. Geobios 76 (2023) 53–73 Maul, L.C., Markova, A.K., 2007. Similarity and regional differences in Quaternary Raup, D.M., 1991. A Kill Curve for Phanerozoic marine species. Paleobiology 17, 37– arvicolid evolution in Central and Eastern Europe. Quaternary International 160, 48. 81–99. Repenning, C.A., Fejfar, O., Heinrich, W.D., 1990. Arvicolid rodent biochronology of Mein, P., Moissenet, E., Truc, G., 1978. Les formations continentales du Néogène the Northern Hemisphere. In: Fejfar, O., Heinrich Wolf, D. (Eds.), International Supérieur des Vallées du Júcar et du Cabriel au NE d’Albacete (Espagne), symposium on Evolution, phylogeny and biostratigraphy of arvicolids biostratigraphie et environnement. Documents des Laboratoires de Géologie de (Rodentia, Mammalia). Geological Survey, Prague, pp. 385–418. la Faculté des Sciences de Lyon 72, 99–147. Rümke, C.G., 1985. A review of fossil and Recent Desmaninae (Talpidae, Mein, P., Moissenet, E., Adrover, R., 1983. L’extension et l’âge des formations Insectivora). Utrecht Micropaleontological Bulletins. Special Publication 4, 1– continentales pliocènes du fossé de Teruel (Espagne). Comptes-Rendus des 241. Séances de l’Académie des Sciences, Serie 2 (296), 1603–1610. Salvador, A., 1994. International Stratigraphic Guide. A guide to stratigraphic Mein, P., Moissenet, E., Adrover, R., 1990. Biostratigraphie du Néogène supérieur du classification, terminology, and procedure. The Geological Society of America, bassin de Teruel. Paleontologia i Evolució 23, 121–139. Boulder. Mein, P., 1989. Updating of MN zones. In: Lindsay, E., Fahlbusch, V., Mein, P. (Eds.), Schmidt-Kittler, N., 1995. The vertebrate locality Maramena (Macedonia, Greece) at European Neogene mammal chronology. NATO ASI Series A, pp. 73–90. the Turolian-Ruscinian boundary (Neogene), Münchner Geowissenschaftliche Minwer-Barakat, R., García-Alix, A., Suarez, E.M., Freudenthal, M., Viseras, C., 2012. Abhandlungen 28. Pfeil, Munich. Micromammal biostratigraphy of the Upper Miocene to lowest Pleistocene Scott, G.R., Gibert, L., Gibert, J., 2007. Magnetostratigraphy of the Orce Region (Baza continental deposits of the Guadix basin, southern Spain. Lethaia 45, 594–614. Basin), SE Spain: New chronologies for Early Pleistocene faunas and hominid Morales, J., Peláez-Campomanes, P., Abella, J., Montoya, P., Gibert, L., Scott, G., occupation sites. Quaternary Science Reviews 26, 415–435. Cantalapiedra, J.L., Sanisidro, O., 2013. The Ventian mammal age (Latest Tesakov, A.S., 1998. Voles of the Tegelen fauna. In: van Kolfschoten, T., Gibbard, P.L. Miocene): present state. Spanish Journal of Palaeontology 28, 149–160. (Eds.), The dawn of the Quaternary; proceedings of the SEQS-EuroMam Nobel, F.A., Andriessen, P.A.M., Hebeda, E.H., Priem, H.N.A., Rondeel, H.E., 1981. symposium 1996. Nederlands Instituut voor Toegepaste Geowetenschappen Isotopic dating of the post-alpine Neogene volcanism in the Betic Cordilleras, TNO, Haarlem, Netherlands, pp. 71–134. Southern Spain. Geologie en Mijnbouw 60, 209–214. van Balen, R.T., Forzoni, A., van Dam, J.A., 2015. Active faulting and folding along Oms, O., Dinares-Turell, J., Agustí, J., Parés, J.M., 1999. Refinements of the European Jumilla Fault Zone, northeastern Betics, Spain. Geomorphology 237, 88–97. mammal biochronology from the magnetic polarity record of the Plio- van Dam, J.A., 1997. The small mammals from the upper Miocene of the Teruel- Pleistocene Zújar section, Guadix-Baza basin, SE Spain. Quaternary Research Alfambra region (Spain): paleobiology and paleoclimatic reconstructions. 51, 94–103. Geologica Ultraiectina 156, 1–204. Oms, O., Agustí, J., Gabas, M., Anadón, P., 2000. Lithostratigraphical and correlation van Dam, J.A., 2003. European mammal-based chronological systems: past, present of micromammal sites and biostratigraphy of the upper Pliocene to lower and future. Deinsia 10, 85–96. Pleistocene in the Northeast Guadix-Baza Basin (southern Spain). Journal of van Dam, J.A., 2004. Anourosoricini (Mammalia: Soricidae) from the Mediterranean Quaternary Science 15, 43–50. region: a pre-Quaternary example of recurrent climate-controlled north-south Opdyke, N., Mein, P., Lindsay, E., Pérez-Gonzáles, A., Moissenet, E., Norton, V.L., range shifting. Journal of Paleontology 78, 741–764. 1997. Continental deposits, magnetostratigraphy and vertebrate paleontology, van Dam, J.A., Alcalá, L., Alonzo Zarza, A., Calvo, J.P., Garcés, M., Krijgsman, W., 2001. late Neogene of Eastern Spain. Palaeogeography, Palaeoclimatology, The upper Miocene mammal record from the Teruel-Alfambra region (Spain). Palaeoecology 133, 129–148. The MN system and continental stage/age concepts discussed. Journal of Padial-Ojeda, J., 1986. Estudio de los roedores y lagomorfos del Mioceno continental Vertebrate Paleontology 21, 367–385. de la depresión de Granada. Tesis Doctoral, Universidad de Granada 303. van Dam, J.A., Abdul Aziz, H., Álvarez Sierra, M.Á., Hilgen, F.J., van den Hoek Ostende, Opdyke, N., Mein, P., Moissenet, E., Pérez-González, A., Lindsay, E., Petko, M., 1989. L.W., Lourens, L.J., Mein, P., van der Meulen, A.J., Peláez-Campomanes, P., 2006. The magnetic stratigraphy of the Late Miocene sediments of the Cabriel Basin, Long-period astronomical forcing of mammal turnover. Nature 443, 687–691. Spain. In: Lindsay, E., Fahlbusch, V., Mein, P. (Eds.), European Neogene mammal van Dam, J.A., Furió, M., van Balen, R.T., 2014. Re-interpreting the biochronology of chronology. ASI Series A, NATO, pp. 507–514. the La Celia and Los Gargantones mammal sites (Late Miocene, Murcia, Spain). Piñero, P., Agustí, J., 2019. The rodent succession in the Sifón de Librilla section Geobios 47, 155–164. (Fortuna Basin, SE Spain): implications for the Mio-Pliocene boundary in the van de Weerd, A., 1976. Rodent fauna of the Mio-Pliocene continental sediments of Mediterranean terrestrial record. Historical Biology 31, 279–321. the Teruel-Alfambra region, Spain. Utrecht Micropaleontological Bulletins. Piñero, P., Agustí, J., Oms, O., 2018. The late Neogene rodent succession of the Special publication 2, 1–185. Guadix-Baza basin (south-eastern Spain) and its magnetostratigraphic van der Meulen, A.J., García-Paredes, I., Álvarez-Sierra, M.A., van Den Hoek Ostende, correlation. Palaeontology 61, 253–272. L.W., Hordijk, K., Oliver, A., Peláez-Campomanes, P., 2012. Updated Aragonian Piñero, P., Verzi, D.H., 2020. A new early Pliocene murine rodent from the Iberian biostratigraphy: Small mammal distribution and its implications for the Peninsula and its biostratigraphic implications. Acta Palaeontologica Polonica Miocene European chronology. Geologica Acta 10, 159–179. 65, 719–731. Pueyo, E.L., Muñoz, A., Laplana, C., Parés, J.M., 2016. The Last Appearance Datum of Hipparion in Western Europe: magnetostratigraphy along the Pliocene– Further reading Pleistocene boundary in the Villarroya Basin (Northern Spain). International Journal of Earth Sciences (Geologische Rundschau) 105, 2203–2220. Martín-Suárez, E., 1988. Sucesiones de micromamíferos en la Depresión de Guadix- Raffi, I., Wade, B.S., Pälike, H., 2020. The Neogene Period. In: Gradstein, F.M., Ogg, J. Baza (Granada, España). Tesis Doctoral, Univ. Granada, 241 p. G., Schmitz, M.D., Ogg, G. (Eds.), Geologic time scale 2020. Elsevier, Amsterdam, pp. 1141–1215. 73