DeMiguel et al. BMC Biology (2021) 19:6 https://doi.org/10.1186/s12915-020-00939-5 RESEARCH ARTICLE Open Access Palaeoecological differences underlie rare co-occurrence of Miocene European primates Daniel DeMiguel1,2* , Laura Domingo3,4, Israel M. Sánchez2, Isaac Casanovas-Vilar2, Josep M. Robles2 and David M. Alba2 Abstract Background: The two main primate groups recorded throughout the European Miocene, hominoids and pliopithecoids, seldom co-occur. Due to both their rarity and insufficiently understood palaeoecology, it is currently unclear whether the infrequent co-occurrence of these groups is due to sampling bias or reflects different ecological preferences. Here we rely on the densely sampled primate-bearing sequence of Abocador de Can Mata (ACM) in Spain to test whether turnovers in primate assemblages are correlated with palaeoenvironmental changes. We reconstruct dietary evolution through time (ca. 12.6–11.4 Ma), and hence climate and habitat, using tooth-wear patterns and carbon and oxygen isotope compositions of enamel of the ubiquitous musk-deer Micromeryx. Results: Our results reveal that primate species composition is strongly correlated with distinct environmental phases. Large-bodied hominoids (dryopithecines) are recorded in humid, densely-forested environments on the lowermost portion of the ACM sequence. In contrast, pliopithecoids inhabited less humid, patchy ecosystems, being replaced by dryopithecines and the small-bodied Pliobates toward the top of the series in gallery forests embedded in mosaic environments. Conclusions: These results support the view that pliopithecoid primates preferred less humid habitats than hominoids, and reveal that differences in behavioural ecology were the main factor underpinning their rare co- occurrence during the European Miocene. Our findings further support that ACM hominoids, like Miocene apes as a whole, inhabited more seasonal environments than extant apes. Finally, this study highlights the importance of high-resolution, local investigations to complement larger-scale analyses and illustrates that continuous and densely sampled fossiliferous sequences are essential for deciphering the complex interplay between biotic and abiotic factors that shaped past diversity. Keywords: Hominoids, Pliopithecoids, Primate evolution/adaptation, Palaeodiet, Stable isotopes, Tooth wear, Feeding behaviour, Palaeobiology * Correspondence: demiguel@unizar.es 1 ARAID foundation / Universidad de Zaragoza, Departamento de Ciencias de la Tierra, and Instituto Universitario de Investigación en Ciencias Ambientales de Aragón (IUCA), Pedro Cerbuna 12, 50009 Zaragoza, Spain 2 Institut Català de Paleontologia Miquel Crusafont, Universitat Autònoma de Barcelona, Edifici ICTA-ICP, C/ Columnes s/n, Campus de la UAB, 08193 Cerdanyola del Vallès, Barcelona, Spain Full list of author information is available at the end of the article © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. DeMiguel et al. BMC Biology (2021) 19:6 Page 2 of 15 Background unlikely to hold for different clades such as pliopithe- Fossil primates from the Miocene of Europe are gener- coids and hominoids, characterised among others by dif- ally rare and absent from most sites, and when recorded, ferent locomotor adaptations—leading to the proposal different primate species only seldom co-occur within a that these two groups probably had different habitat single locality (stratigraphic horizon). As a result, there preferences, which only enabled their coexistence under is an ongoing debate about the factors underpinning the particular ecological conditions [1]. Early ecomorpholo- geographic and chronostratigraphic distribution of Mio- gical analyses based on ungulate hypsodonty (a proxy for cene primates in this continent [1–6]. Before the disper- vegetation structure also used to infer palaeoprecipita- sal of cercopithecoids (Old World monkeys) into Europe tion) have concluded that both hominoid and by the early Turolian (ca. 8.5 Ma, late Miocene), two pliopithecoid-bearing localities from the European Mio- main groups are recorded there: pliopithecoids, generally cene were more humid than those lacking primates [2]. considered a Eurasian clade of stem catarrhines (i.e. pre- More recent work based on hypsodonty further showed ceding the cercopithecoid-hominoid split [7]), and homi- that pliopithecoids generally inhabited more humid envi- noids (crown catarrhines more closely related to extant ronments (i.e. with higher moisture and/or rainfall) than apes and humans than to cercopithecoids [8–10]). Both hominoids, although probably less humid than those in groups presumably dispersed from Africa to Eurasia fol- which both groups co-occur [4]. However, given the lowing the closure of the Tethys Seaway during the late small number of fossil localities recording both taxa, middle Miocene and subsequently diversified across the such comparisons lack statistical power and may fail to continent giving rise to multiple genera and species. consider palaeoenvironmental differences across geog- There are approximately one hundred known localities raphy and time throughout the Miocene, especially at recording either or both of these groups—almost 20% the regional and local scales. corresponding to Abocador de Can Mata (ACM) in Focusing on faunal elements that accompany primates Spain—although they only co-occur in less than 10% of within a single area and over a restricted time span them, with pliopithecoid-bearing localities being slightly would allow us to test whether turnovers in the primate more abundant (ca. 55 vs. 45%) than hominoid-bearing assemblage are correlated to local changes in palaeo- ones [4, 5]. European hominoids are generally larger environmental conditions. The composite stratigraphic than pliopithecoids and considered great apes (homi- sequence of ACM, located in the area of els Hostalets de nids), except for the small-bodied Pliobates, interpreted Pierola within the Vallès-Penedès Basin (NE Iberian as a stem hominoid [11]. Peninsula [14]) (Fig. 1a–c), and spanning more than 1 There are very few sites in the European Miocene Myr (12.6–11.4 Ma [5, 15]), offers an unparalleled op- where hominoids and pliopithecoids co-occur [1, 4, 5], portunity to test this hypothesis for several reasons. and in most cases, fossils of each group come from dif- First, the ACM sequence has delivered one of the most ferent localities within the same site (e.g. different karstic diverse primate assemblages from the European Mio- fissure fillings from La Grive) or it is uncertain whether cene, including both hominoids and pliopithecoids [5]. their remains came from the same stratigraphic horizon Second, the co-occurrence of hominoids and pliopithe- (e.g. Castell de Barberà [6]). Strong taphonomic evidence coids has only been recorded in one out of the 19 ACM supporting sympatry is only available from Rudabánya in primate-bearing localities, and the distribution of each Hungary [12] and ACM (locality ACM/C5-C3 [5]). group throughout the series does not appear random These localities therefore offer a unique opportunity to [5]. Finally, thanks to continuous palaeontological sur- evaluate the palaeoenvironmental conditions that en- veillance during the construction of a landfill, most of abled the coexistence of pliopithecoids and hominoids. the fossil finds are accurately dated based on detailed Given the rarity of primate remains among mammalian litho-, bio- and magnetostratigraphic correlations [5, 16, assemblages from the European Miocene, the infrequent 17]. This offers the opportunity to test alternative expla- co-occurrence of two different primate species at a sin- nations for the variable temporal distribution of both gle locality might be, at least in part, a sampling artefact primate groups during a restricted time span and within [1]. However, the lack of co-occurrence in many well- a uniform depositional setting. sampled primate-bearing localities would rather support With this aim in mind, here we present a reconstruc- the view that their infrequent coexistence is a real tion of the local climate and palaeoenvironments phenomenon that requires an explanation. through the ACM sequence based on tooth wear and The competitive exclusion principle [13] predicts that dental enamel stable carbon and oxygen isotope values species occupying the same ecological niche cannot co- (represented by the notation δ13C and δ18O) of the ru- exist on the long-term, ultimately leading to the preva- minant Micromeryx—(Fig. 1d and Figure S1)—a repre- lence of one over the other, or to the progressive sentative of the family Moschidae (musk-deer) [18, 19]. divergence of their respective niches. This explanation is Because the diet of any plant-eating mammal is a direct DeMiguel et al. BMC Biology (2021) 19:6 Page 3 of 15 Fig. 1 (See legend on next page.) DeMiguel et al. BMC Biology (2021) 19:6 Page 4 of 15 (See figure on previous page.) Fig. 1 Abocador de Can Mata (ACM) and the moschid Micromeryx. a Geographical situation and general geological context of the Vallès-Penedès Basin. b Detailed geological map of the basin and the sequence of Abocador de Can Mata (ACM) (black dot). c Correlation of the composite local magnetostratigraphy of ACM series with the Geomagnetic Polarity Time Scale (modified from Alba et al. [5]). European Land Mammal Ages, Mammal Neogene (MN) units and local biozones of the Vallès-Penedès Basin are shown on the left. The shadowed region indicates an unsampled interval of the Vallès-Penedès record. The stratigraphic positions of the ACM localities studied in this work are shown to the right on the composite lithostratigraphic column. Note that the bottom boundary of the lowermost local biozone is unknown. d Life reconstruction of a Micromeryx azanzae male. Art by I.M.S link with the habitats in which it lives, we used the diet with the Fisher exact test. For the chi-square test, the re- (i.e. ecology) of this ruminant to inform about ACM pri- sults show that Micromeryx morphotype 1 is different mate ecological preferences and habitats. The selection from morphotype 2 (p = 0.0283), whereas non-significant of Micromeryx is based on the following reasons: (1) the differences are between morphotypes 2 and 3 (p = record of this taxon throughout the ACM stratigraphic 0.2255) and between morphotypes 1 and 3 (p = 0.2019). sequence, characterised by abundant isolated teeth and On average, mesowear results indicate a browsing on dentognathic fragments, allows us to construct a con- soft vegetation and low levels of abrasives (endogenous tinuous tooth-wear and isotopic record; (2) by focusing phytolith-rich grasses and dicotyledonous, and exogen- on a single genus, we can characterise more consistently ous dust and grit), although Micromeryx morphotype changes in the vegetation cover, food abrasiveness, etc., 1—with more rounded cusps and higher MS (Table 1)— across the selected time interval, avoiding thus biases shows a shift toward the exploitation of tougher and due to different physiologies; and (3) Micromeryx was more abrasive foods than the others. ubiquitous in the Miocene of Iberia, inhabiting a varied range of biomes from more or less open savannas to Stable isotope data (sub) tropical forests [18, 20], and exhibiting an extraor- The difference between carbonate (δ18OCO3) and phos- dinary versatility in terms of exploitation of nutrients phate (δ18OPO4) oxygen isotopic composition can be and resources. The combination of all these factors justi- used to monitor possible bioapatite diagenetic alteration. fies the suitability of employing Micromeryx as a case Micromeryx tooth enamel did not undergo extensive study to investigate the environmental and climatic post-burial alteration since the difference calculated be- shifts that took place during the latest middle Miocene tween δ18OCO3 and δ18OPO4 (Δ18OCO3-PO4 = δ18OCO3 − in the area of els Hostalets de Pierola. δ18OPO4) values for the whole dataset (8.6 ± 0.8‰) is within the range obtained when considering modern Results mammals (~ 8.6–9.1‰ [22, 23]) (Additional file 1: Sup- Tooth wear plementary information, Note 3). The fossil material studied consists of dentognathic re- Micromeryx morphotypes (Table 1 and Additional file 6: mains and isolated teeth of Micromeryx. Although ini- Table S1) yielded tooth enamel δ13C values indicative of tially a single species of Micromeryx was reported from woodland to woodland-mesic C3 grassland conditions ACM [15], the currently available dental material indi- (see Additional file 1: Supplementary information, Note cates the presence of three different morphotypes that 4 for a detailed explanation of the calculated δ13C cut- likely represent different species (Additional file 1: Sup- off values among different habitats). Significant differ- plementary information, Note 1). ences in δ13C values have been only found between mor- For mesowear, we measured individuals and provide photypes 1 and 2 of Micromeryx (t = 4.250, p < 0.001) the results for the three morphotypes separately (Add- (Additional file 7: Table S2). Tooth enamel δ18OCO3 and itional file 1: Supplementary information, Note 2). All δ18OPO4 values do not show significant differences Micromeryx morphotypes (Table 1) show occlusal sur- among the three morphotypes (δ18OCO3: F = 0.845, df = faces with predominance of high relief (pH = 95–100%) 2, p = 0.439, and δ18OPO4: F = 0. 562, df = 2, p = 0.577) and sharpened cusps (pS = 69–87%), although there is a (Additional file 7: Table S2). considerable proportion of rounded apices (pR = 13– 31%). Morphotypes do not have any incidence of blunt Relationship between mesowear and δ13C values cusps or, except for Micromeryx morphotype 3 (pL = A scatter plot showing the correlation between mean 5%), low occlusal relief (which relates to a low height MS and mean δ13C (‰ VPDB) among Micromeryx mor- difference between tooth cusps and valleys). Average photypes by localities was constructed (Fig. 2). Niche do- mesowear score (MS) for morphotypes ranges from 0.19 mains are visually presented for each variable. This to 0.31 (Table 1). We do find significant differences with approach (described in Additional file 1: Supplementary the chi-square test (χ2) but marginally non-significant information, Notes 2 and 4) allowed us to contextualise DeMiguel et al. BMC Biology Table 1 Summary of mesowear and isotopic values of Micromeryx from the ACM sequence according to morphotypes and environmental phases (2021) 19:6 Morphotypes # M pS pR pH MS #C δ13C SD δ18OCO3 SD #P δ18OPO4 SD δ18OPO4 Δ δ18OCO3- δ13C δ18OCO3 δ18OPO4 Morphotype 1 8 69.2 30.8 100 0.31 8 − 10.5 0.7 27.9 1.0 6 19.6 1.4 8.4 Morphotype 2 24 80.6 19.4 100 0.19 21 − 11.8 0.8 27.7 1.5 19 19.1 1.9 8.7 Morphotype 3 11 86.7 13.3 95 0.21 6 − 11.0 1.2 27.0 0.6 5 18.5 1.1 8.3 13 18 18 18 13 Environmental #M pS pR pH MS #C δ C SD δ OCO3 SD #P δ OPO4 SD Δ δ OCO3- δ Cdiet, mequ MAPa (mm/ MAPb (mm/ MAT phases δ13C δ18OCO3 δ18OPO4 δ18OPO4 year) year) (°C) Phase III 23 71.9 28.1 97.3 0.32 19 − 11.1 1.2 27.2 1.2 19 18.5 1.6 8.7 − 27.2 801 608 17.1 (11.70–11.60 Ma) Phase II 15 83.3 16.7 100 0.16 13 − 11.4 1.2 27.7 1.2 6 19.8 1.7 8.2 − 27.5 992 765 20.3 (11.90–11.79 Ma) Phase I 7 90.9 9.1 100 0.09 5 − 12.0 0.3 28.9 0.8 5 20.4 0.9 8.5 − 28.1 1190 928 21.8 (12.33–11.95 Ma) #M (number of samples for mesowear); percentage of specimens with sharp (pS) and rounded (pR) cusps; percentage of specimens with high (pH) occlusal relief; mesowear score (MS); #C (number of samples for stable isotope analyses on the carbonate fraction); mean δ13C (‰ VPDB); standard deviation (SD) δ13C (‰ VPDB); mean δ18OCO3 (‰ VSMOW); standard deviation (SD) δ18OCO3 (‰ VSMOW); #P (number of samples for stable isotope analyses on the phosphate fraction); mean δ18OPO4 (‰ VSMOW); standard deviation (SD) δ18OPO4 (‰ VSMOW); Δδ18OCO3 − δ18OPO4, mean δ13Cdiet, mequ (‰ VPDB); inferred mean MAP (mm/year) (from Kohn [21]) without (estimated MAPa) and with (estimated MAPb) altitude and latitude correction; and inferred mean MAT (°C) Page 5 of 15 DeMiguel et al. BMC Biology (2021) 19:6 Page 6 of 15 Fig. 2 Scatter plot of mesowear and δ13C values (‰ VPDB) of Micromeryx. Stippled areas show the transition between C3-dominated diets, mixed C3-C4 diets and C4-dominated diets. Colour informs about Micromeryx morphotype (green for Micromeryx morphotype 1, blue for Micromeryx morphotype 2 and yellow for Micromeryx morphotype 3), and symbol refers to the temporal range within the sedimentary sequence (diamond for 12.38–11.95 Ma, circle for 11.90–11.79 Ma and square for 11.70–11.60 Ma). See Additional file 1: Supplementary information, Notes 2 and 3 for further details the niche occupation per morphotype and locality given of relatively low abrasion, as seen in extant forest- the variables investigated. MS and δ13C values point to a dweller browsers [26]. We rule out a regular consump- frequent ingestion of C3 plants in woodland to mesic C3 tion of fruits and/or seeds in ACM, as teeth show no grasslands. Mean MS of some individuals of morpho- signs of strong rounding or blunt apices—and frugivor- types 1 and 2 from a few localities (those with MS = 0.33 ous taxa have significant percentages of rounded and to 0.5) is also compatible with regular consumption of blunt cusps because of tip-crushing [27]. These results C4 vegetation. C4 plants have never been documented as contrast with some previous data for middle and late an important component of plant communities in the Miocene Micromeryx from elsewhere in Europe, which Iberian Neogene (despite being recorded there since the appear strictly frugivorous [28–30]. However, these re- Oligocene) [24, 25]. However, they may well have been sults are in agreement with the leaf browsing inferred present to some extent in some areas and/or time inter- for other Micromeryx [29, 31]. Therefore, it seems that vals (e.g. ACM/C4-C1 and ACM/C5-C2 at ~ 11.8 Ma). Micromeryx, since its oldest occurrences in the middle Miocene of Eurasia, was capable of feeding alternatively Discussion on fruits, seeds and soft leaves, depending on habitat- Micromeryx diet at ACM specific circumstances (e.g. ecologic niche partitioning Our results indicate that the bulk of Micromeryx diet at or food availability). In the case of ACM, the unusual ACM consisted of foliage with a particular emphasis on secondary crests of the upper molars of Micromeryx are forbs, dicots and woody leaves. The various morpho- compatible with an adaptation for heavier reliance on types generally maintain sharpened, high-relief cusp api- leaves and stems, as seen in other mammalian groups ces and low MS—a signal that informs that foods were [32]—an anatomical trait that is consistent with the DeMiguel et al. BMC Biology (2021) 19:6 Page 7 of 15 folivorous signal retrieved from mesowear analyses. The Temporal patterns in Micromeryx tooth wear and stable Micromeryx from ACM are therefore the only ones in isotopes along the ACM sequence which these features are recognised, probably showing a The primate assemblage recorded at ACM [5] includes regional adaptation associated with the particularity of three great ape (dryopithecine) species from different these environments. Moreover, Micromeryx had a wider genera (Pierolapithecus catalaunicus, Anoiapithecus bre- dietary plasticity than modern Moschus, whose diet com- virostris and Dryopithecus fontani) [5, 10, 16, 34], the prises mainly arboreal lichens (a resource rarely pliopithecoid Pliopithecus canmatensis [35], probably a exploited by other ruminants), forbs and woody leaves second pliopithecoid unassigned to species [36] and a [33]. putative stem hominoid (Pliobates cataloniae) [11], al- Within such a generalised soft, leafy browsing, there ternatively interpreted as a pliopithecoid [37]. The distri- are differences among morphotypes in tooth wear and bution of these taxa is not homogeneous along the ACM isotopic values through time (Additional file 6: Table sequence, with Pliopithecus postdating most great ape S1). In other words, the same morphotype behaves dif- finds but preceding Pliobates [5] (Fig. 3a). ferently when the temporal gradient is considered. The We show mesowear and isotopic data of Micromeryx less sharp and more rounded cusps of Micromeryx mor- teeth in chronological order of the localities for a time photypes 1 and 2 recorded from 11.81 Ma onwards (and span ranging from 12.38 to 11.63 Ma (Fig. 3), to com- a signal of browse-dominated mixed feeding for some pare their chronostratigraphic distribution with palaeo- individuals) reflect a more pronounced abrasion than in environmental changes through time. While remaining older specimens and indicate that more abrasive browse within a browsing dietary category based on C3 plants, and/or some dust/grit-infested foliage was eaten—as ex- differences observed in both mesowear and isotopic tant browsers that feed on leafy, soft foods generally values among Micromeryx specimens provide insight on maintain sharpened/high relief cusps [27, 31] (Additional the environmental changes along the ACM temporal se- file 6: Table S1). That is, abrasive browse and encroach- quence during the latest middle Miocene. This is par- ments by exogenous dust/grit content was only slight or ticularly relevant given that the lack of other even absent in the diet of older Micromeryx, whereas it palaeoenvironmental proxies from this area (such as significantly increased in later forms along the ACM se- pollen or macroplant remains) hinders a more precise quence. δ13C data support dietary inferences based on reconstruction of the vegetation structure and other mesowear, as all Micromeryx morphotypes depict values characteristics of the various habitats occupied by that are within woodland to woodland-mesic C3 grass- primates. land conditions, thereby indicating a consumption of Overall, the data reported here indicate that ACM both soft leaves and more abrasive grasses. habitats became progressively less humid and more Fig. 3 Correlation of mesowear and isotopic values of Micromeryx arranged temporally along the ACM sequence. a Environmental phases, separated by dashed lines, and stratigraphic ranges of the primates recovered at ACM based on occurrence in localities. b Average mesowear scores (MS) by locality for Micromeryx morphotypes. c Micromeryx tooth enamel raw and mean δ13C (‰ VPDB). Calculated average modern equivalent of diet composition (δ13Cdiet, meq, ‰ VPDB) and estimated MAP (mm/year) values (without [MAPa] and with [MAPb] altitude and latitude correction) are given in parentheses. d Micromeryx tooth enamel raw and mean δ18OCO3 and δ18OPO4 (‰ VSMOW) values. Calculated MAT (°C) values are given in parentheses. Locally weighted polynomial lines are fitted to isotopic data. Calculations further include the 95% confidence region. Colour symbols are for raw data and grey symbols are for mean values. See Additional file 1: Supplementary information, Note 4 for further details DeMiguel et al. BMC Biology (2021) 19:6 Page 8 of 15 heterogeneous (or, at least, are characterised by a gradi- genus Moschus (musk deer) [44], which inhabits forest ent toward less dense canopy structure and more open and mountainous parts of Asia [33, 45] and has a brows- patches), as reflected by (i) a trend toward higher values ing diet, although it also has the ability to cope with of mesowear—from the sharpest cusps and low MS of 0 poorer, less nutritious foods when high-quality forage is at 12.38 Ma (ACM/C1-Ee) to more intermediate (more in short supply, such as in winter [33, 46]. Independently rounded) cusp morphologies and higher MS around 0.5 from its dietary behaviour, Moschus has been observed at 11.60 Ma (ACM/C5-D1) (Fig. 3 (b)); (ii) an increase in to drink water on a daily basis (Prikhod’ko, pers. mean Micromeryx δ13C values (from − 12.4‰ in ACM/ comm.); therefore, its tooth enamel δ18O signal will C1-Ee to − 11.1‰ in ACM/C5-D1; Fig. 3 (c)); and (iii) a largely be dependent on drinking water δ18O values. On high variability in the type of vegetation consumed—re- the assumption that the water reliance of Moschus is ap- vealed by a wider range of δ13C values in the youngest plicable to Micromeryx, mean annual temperature localities and fuelled by the coexistence of the three (MAT) values have been estimated based on Micromeryx morphotypes (see a change in standard deviation values tooth enamel δ18OPO4 values. They show a decreasing in Table 1). The increase in mean δ13C values may have trend along the ACM sequence from 21.4 °C in ACM/ been driven by two phenomena: a shift toward drier hab- C1-Ee to 17.4 °C in ACM/C5-D1 (Table 1, Fig. 3 (d), itats, including non-forest patches, or, alternatively, a Additional file 6: Table S1). This trend toward lower greater reliance on fruits through time. A change toward temperatures may be framed within the gradual cooling a more frugivorous diet would have led to a slight in- that started by 14 Ma after the Mid-Miocene Climatic crease in Micromeryx tooth enamel δ13C, as a significant Optimum [47]. In the Iberian Peninsula, this long inter- consumption of fruits ultimately results in higher bioa- val, which coincided with the expansion of mesothermic patite δ13C values [38, 39]. There is, however, little rea- deciduous vegetation and the extinction or significant son to support strong frugivory for ACM Micromeryx, as decrease in abundance of thermophilous evergreen this is contradicted by their attrition-dominated meso- plants [48, 49], witnessed an increase in the diversity of wear patterns. This does not mean that fruits were un- moschids [18]. available at ACM. In fact, all of the primate species recorded relied on frugivory to a large extent, even if Primate assemblage composition in relation to with a different emphasis on hard-object feeding de- palaeoenvironmental changes pending on the species [11, 40, 41]. However, a drop in Our analyses further show a fluctuation in diet compos- the estimated mean annual precipitation (MAP) values ition for Micromeryx individuals, revealing the existence throughout the sequence (from ~ 1395 to ~ 762 mm/ of three distinct environmental phases in ACM (Table 1, year or from ~ 1097 to ~ 575 mm/year with altitude and Fig. 3 (a–d)), with temporal patterns in precipitation, latitude correction; Table 1, Fig. 3 (c), Additional file 6: temperature and aridity that relate to changes in primate Table S1) supports the fact that fruits were preferentially assemblage composition. exploited by arboreal, or at least semiterrestrial, species instead of terrestrial taxa such as Micromeryx. Phase I Although the oxygen isotope composition does not A first phase ranges from the beginning of the sequence vary significantly through time, there is a slight decrease (12.38 Ma, ACM/C1-Ee) to ~ 11.95 Ma (ACM/C4-Cp), in both δ18OCO3 and δ18OPO4 values (from − 28.8‰ and where only Micromeryx morphotype 2 is recorded. 20.3‰ in ACM/C1-Ee to 27.4‰ and 18.7‰ in ACM/ Overall, Micromeryx maintained sharp apices, high-relief C5-D1; Fig. 3 (c, d)). Oxygen isotope composition of car- cusp and low average MS of 0.1, and tooth enamel δ13C bonate (δ18OCO3) and phosphate (δ18OPO4) fractions of values of − 12.0 ± 0.3‰ (VPDB) (Table 1) that point to tooth enamel reflects δ18O of body water (δ18Obw) [42, the consumption of plant resources from relatively dense 43]. Changes associated with δ18Obw value mirror varia- wooded areas (see Additional file 1: Supplementary in- tions in the isotopic composition of ingested water, ei- formation, Note 4 for explanation of the calculated δ13C ther through drinking or plant water (in the case of cut-off values among different habitats). Estimated MAP herbivores). Micromeryx δ18O values are less likely to ranges from 928 to 1190 mm/year (depending on vary according to physiological factors like fractionated whether a correction for altitude and latitude is applied water loss through the lungs or skin, since this is a rela- or not) (Table 1). In phase I, Micromeryx tooth enamel tively mesic environment overall, and the species are δ18OCO3 and δ18OPO4 values are the highest among the closely related over a narrow time window. When con- three environmental phases, with calculated MAT values sidering extinct mammals such as Micromeryx, it is diffi- reaching 21.8 °C (Table 1). According to Whittaker’s cult to assess the type of water economy they may have biome classification [50], estimated MAP and MAT for had, due to the lack of modern analogues. The extant phase I would correspond to those of a tropical seasonal sister group of Micromeryx within the Moschidae is the forest/savanna (Additional file 5: Figure S4). This agrees DeMiguel et al. BMC Biology (2021) 19:6 Page 9 of 15 with the soft-leafy browsing diet inferred from mesowear subtropical desert [50] (Additional file 5: Figure S4). In and indicates a humid climate with rainfall seasonality the light of the fauna recorded at ACM [15], we consider (although not marked) and the development of long- the latter inference as unrealistic and most likely attrib- standing forests with bushy and woody vegetation [23, utable to a preservational bias toward drier ecosystems 31, 51]. This type of environment at the beginning of the [54]. Alternatively, higher CO2 levels during the Miocene ACM series, characterised by humid and warm forests might produce a similar bias in biome reconstructions, with a dense upper canopy, is somewhat more seasonal given their documented relationship not only with than previous inferences for ACM as a whole [52] and higher temperatures but also enhanced water-use effi- would be suitable for the multiple large-bodied homi- ciency and leaf-level productivity [55]. This “forest noids—A. brevirostris, D. fontani and P. catalaunicus— fertilization effect”, resulting from higher CO2 levels in recorded during phase I. The abundance of trees may the Miocene, might have resulted in more forested envi- have allowed hominoids to eat a diverse array of vegeta- ronments than indicated by estimated MAP and TAP tion, ranging from leaves and soft fruits (Anoiapithecus based on current standards. Discerning whether such and Dryopithecus) to harder and brittle fruits (Pierola- potential biases apply uniformly to the whole ACM se- pithecus) [11, 41]. This also fits with the postcranial quence would require a taxonomically broader isotopic morphology of Pierolapithecus, which indicates an or- sampling in selected ACM localities—as averaging values thograde bodyplan with adaptations for arboreal vertical from multiple taxa from the same locality would argu- climbing [10, 21, 34, 53]. Only a single Micromeryx spe- ably provide more robust MAP estimates [54]. Neverthe- cimen from ACM/C2-A3 (IPS29396) displays more less, we consider that the palaeoenviromental changes rounded cusps and slightly higher δ13C and δ18O values. recorded by Micromeryx isotopic values through time This specimen might reflect the exploitation of more are at least valid in relative terms, even if their exact in- abrasive elements probably located along less humid— terpretation in terms of extant biomes should be subject but still forested—patches in specific localities. It is note- to further scrutiny. The development of mosaic environ- worthy that only one pliopithecoid is recorded in this ments (with the earlier forested habitats containing for first phase (Fig. 3 (a)) (Pliopithecoidea indet. from the first time partial clearing as new open patches) in ACM/C3-B2 at 12.06 Ma). This immediately precedes a ACM might have allowed the local evolution of new first short pulse of decreased humidity as documented Micromeryx morphotypes (i.e. species) adapted to more by ACM/C2-A3, which might explain the lack of great open landscapes and with different dietary preferences ape record between their first appearance in the se- (e.g. more abrasive forbs, shrubs and other ligneous quence at 12.4–12.3 and their more abundant record at vegetation rich in phytoliths, and even some grass). Such 12.0–11.9 [5]. This shifting climatic pattern at 11.98 Ma an interpretation is reinforced by the record at ACM/ toward less humid conditions might have also influenced C3-Ak of the bovid Tethytragus, a common faunal elem- (preferred) food availability (as seen in Micromeryx) and ent in the more open and arid palaeoenvironments from impelled hominoids to exploit alternative sources, espe- inner Iberia that is otherwise not documented from the cially as fallback foods [41], not consumed before. Vallès-Penedès Basin [56]. Our results support greater habitat heterogeneity, rather than a complete change in Phase II the palaeoenvironment compared to the previous phase. ACM localities experienced a different environmental On the one hand, Micromeryx morphotype 2 (with affin- phase from ~ 11.90 to 11.79 Ma. There was a rapid in- ity for humid conditions) persists in phase II with little crease in Micromeryx phenotypic diversity and popula- variation in MS and δ13C—with higher values at the end tion abundance after ACM/C3-Ak (11.88 Ma), with the of phase II (at ACM/C4-C1, 11.81 Ma) likely indicating first co-occurrence of all morphotypes (at least three) that more abrasive foods and/or some grit loaded foliage being recorded at ACM/C4-C1 (Fig. 3 (b–d)). Compared was eaten at this time. On the other hand, the new mor- to phase I, from 11.88 Ma onwards the less sharp and photypes 1 and 3 appear for the first time with higher more rounded cusp shapes, higher average MS of 0.16 of mean MS and δ13C values (Fig. 3 (b, c)). Micromeryx and the increase in the mean δ13C (− 11.4 ± These changes toward habitat (canopy) fragmentation, 1.2‰, VPDB) (Table 1, Fig. 3 (b, c)) are consistent with leading to a mosaic of forest patches interrupted by less humid and more open areas during this part of the more open woodlands and maybe even shrublands, ACM sequence. The broader range of δ13C observed in would have represented a challenge for the frugivorous phase II (Fig. 3 (c)) is congruent with a phase of in- and presumably arboreal great apes from ACM—espe- creased habitat heterogeneity. Estimated MAP ranges cially in dietary terms (given the impossibility of main- from 765 to 992 mm/year, whereas estimated MAT is taining a year-round supply of ripe fruits), and perhaps 20.3 °C (Table 1). The biomes of phase II would be in also from a locomotor viewpoint (at least for the highly the domains of tropical seasonal forest/savanna and arboreal Pierolapithecus, given the need to travel across DeMiguel et al. BMC Biology (2021) 19:6 Page 10 of 15 the canopy to exploit such resources). Admittedly, ar- of 0.32), less sharp and more rounded cusps, and—for boreal adaptations are only unambiguously recorded for the first time in the sequence—low occlusal relief the orthograde Pierolapithecus [10, 21, 34, 53], while (Table 1, Fig. 3 (b)). Micromeryx tooth enamel mean there are no postcranials for Anoiapithecus, and Dryo- δ13C (− 11.1 ± 1.2‰, VPDB) (Fig. 3 (c)) is also higher. pithecus appears to have been more pronograde [57]. This evidence lends support to increased opening of Nevertheless, there are no indications that Dryopithecus the canopy, with trees growing in more open condi- was particularly adapted to terrestrial locomotion [57], tions or sparser formations. MS and δ13C ranges are whereas in contrast Pliopithecus has been inferred as a the widest (up to twice that recorded at any other generalised semiarboreal quadruped [58, 59], which previous phase) of the sequence, which is consistent would have allowed the latter genus to better exploit not with greater habitat heterogeneity than in phase II. only the shrub level but also to walk terrestrially among Phase III witnesses the lowest δ18OCO3 and δ18OPO4 different arboreal feeding sources. Furthermore, the values (Table 1, Fig. 3 (d)). Calculated MAP reaches smaller body size of pliopithecoids would have likely the lowest value of the three phases ranging from 608 allowed them to exploit smaller patches of fruiting trees to 901 mm/year, and estimated MAT is also the than in the case of great apes. Overall, the inferred lowest inferred with 17.1 °C—which means a drop of palaeoenvironmental changes might explain why great 3 °C compared to phase II (Table 1, Fig. 3 (c, d)). All apes became even rarer faunal elements and were even- this evidence suggests a woodland/shrubland biome tually replaced by pliopithecoids during this phase. Our [50] (Additional file 5: Figure S4). Nonetheless, the results therefore differ from those of previous analyses record of the aquaphilous tragulid Dorcatherium naui based on hypsodonty [4], according to which pliopithe- [60], flying squirrels [61] and arboreal dormice at coids generally inhabited more humid environments ACM/C5-D1 indicate that permanent water masses than hominoids. This discrepancy is largely attributable (palustrine areas or streams) and forests were also to the different methods used. The work by Sukselainen present in this phase. Accordingly, our results are et al. [4] was devised to detect general patterns over consistent with more open, mosaic environments in broader geographical (all Eurasia) and temporal (most of phase III, although they may have been interspersed the Miocene) scales than our study, which is designed to with restricted gallery forests, which would have pro- test such patterns in a small region with a shorter time vided a denser tree cover and preserved some sub- span (ca. 1 Myr) but with considerably higher temporal tropical plant elements—contrasting with the resolution. The environmental constraints affecting the predominant vegetation of the nearby areas. Such occurrence of the different primate taxa may vary be- kind of habitats would have probably been suitable tween distant geographic regions and at different time for Pliobates (whose small body size would have en- intervals. For example, the findings by Sukselainen et al. abled the exploitation of smaller areas of fruiting [4] are certainly influenced by the inclusion of several trees), which displays a mosaic of primitive and hominoid taxa (such as Kenyapithecus, Ankarapithecus derived postcranial features consistent with cautious and Ouranopithecus) that surely lived in more arid habi- and eclectic arboreal climbing, and even some degree tats than pliopithecoids [3], which may result in the of suspensory behaviours [11]—as opposed to the overestimation of some subtle (but real) differences be- more generalised semiterrestrial quadrupedalism of tween hominoids and pliopithecoids. Accordingly, our Pliopithecus mentioned above. The decrease in esti- conclusions cannot be generalised to Miocene European mated MAP and MAT values throughout the three- primates as a whole. To test this possibility, similar ana- phased sequence proposed in this study may be lyses focused on other pliopithecoid and hominoid spe- tracking the progressive drop in temperatures that cies from elsewhere would be required. took place throughout the middle and late Miocene and that was responsible for biome change in the Phase III Iberian Peninsula, from subtropical to drier and more Our results provide compelling evidence for a distinct open conditions [48, 49]. phase toward the top of the ACM sequence (11.70– 11.60 Ma), roughly coinciding with the middle to late The rare co-occurrence of pliopithecoids and hominoids Miocene boundary. During such a short interval, dryo- Combining tooth wear patterns and stable isotope data pithecines reappear at 11.65 Ma after a gap of > 200 kyr for the moschid Micromeryx allows us to reconstruct for in which only Pliopithecus was recorded, and overlap the first time the local palaeoenvironmental changes that with the small-bodied stem hominoid Pliobates (not pre- took place during the ca. 1 Myr time span recorded by viously recorded), even coexisting in a particular locality the ACM series and provide insight on the factors (ACM/C5-D1; 11.63 Ma [5]). Micromeryx morphotypes underpinning the rare co-occurrence of hominoids and yield MS higher on average than in phase II (average MS pliopithecoids. Our results conclusively show that ACM DeMiguel et al. BMC Biology (2021) 19:6 Page 11 of 15 habitats significantly changed with the passing of time comprising a wider array of taxa would be required to and that the recorded turnovers in the primate faunal provide more refined palaeoenvironmental reconstruc- assemblage are correlated with three main successive en- tions for specific ACM localities. vironmental phases. Humid conditions are only predom- Furthermore, the results of our study illustrate that inantly present during phase I (at the lowermost portion continuous and densely sampled fossiliferous sequences, of the sequence, 12.38–11.95 Ma), where most large- such as that from ACM, are essential for deciphering the bodied and presumably arboreal hominoids are re- complex interplay between biotic and abiotic factors that corded; from then onwards, humid ecosystems are only shape the chronostratigraphic and geographic distribu- intermittently present in the area and only the smaller- tion patterns of past taxa. Even if local, such sequences bodied and presumably semiterrestrial pliopithecoids are enable removing potential biases caused by differences recorded during phase II. Hominoid-bearing localities in geographical situation as well as local physiographic thus emerge as different from those recording pliopithe- and climatic factors among localities in continental-wide coids, with the former corresponding to quite homoge- studies. In other words, densely sampled sequences per- neous habitats with high humidity and continuous mit detailed focus on the temporal component of canopies (i.e. densely wooded environments). In con- palaeoenvironmental change and, hence, arguably have a trast, pliopithecoid-bearing localities correspond to less greater potential to test hypotheses on the correlation humid and more patchy habitats with a reduced canopy between faunal, climate and habitat change at a much cover. Dryopithecines are recorded again, together with finer scale than studies at a more global scale. Extending the small-bodied arboreal hominoid Pliobates, during such local comparisons to other densely sampled se- the third phase (corresponding to the uppermost portion quences covering similar time spans elsewhere in Eurasia of the sequence; 11.70–11.60 Ma), characterised by more would be the logical next step to further investigate the open conditions and increased environmental heterogen- general validity of the conclusions drawn in this paper eity. This might be explained by the existence of narrow for the rare co-occurrence of hominoids and pliopithe- forest settings (with more humid conditions than its sur- coids as a whole. roundings), to which hominoids would have probably been restricted to a large extent. Methods Hypothesis-testing framework Conclusions We proceed with the null hypothesis that the chrono- Our results indicate that the chronostratigraphic distri- stratigraphic distribution of pliopithecoids and homi- bution of pliopithecoids and hominoids, as previously noids through ACM is random relative to the suspected [1, 4, 5], is not merely attributable to sampling palaeoenvironmental conditions inferred from the asso- biases, as the various primate taxa are consistently ciated fauna. If the null hypothesis cannot be rejected, it present in different types of environments. The pur- would follow that hominoids and pliopithecoids might ported preference of pliopithecoids for more humid en- have exploited the same (even if varied) habitats and, vironments as compared to large-bodied hominoids, hence, the possibility that their rare co-occurrence is previously inferred based on hypsodonty data [4], is not merely a statistical artefact due to insufficient sampling supported in the case of ACM. More generally, our re- could not be discounted. If the null hypothesis is sults clearly show that hominoid and pliopithecoids rejected, because one group is consistently recorded in exploited different environments, and support that they different types of environments, then it would follow displayed differences in behavioural ecology, largely that their rare co-occurrence is likely attributable to dif- “tracking” the most suitable habitats for each group— ferent ecological preferences. which would explain why they only co-occurred at a few localities throughout the European Miocene. Our results Data set and ACM stratigraphic series further support the view that ACM hominoids, like Micromeryx (Additional file 1: Supplementary informa- other Miocene apes as a whole, inhabited more seasonal tion, Note 1 and Additional file 2: Figure S1) specimens and arid environments than extant apes [3]. Indeed, iso- analysed in this paper are housed in the Institut Català topic data for the Siwaliks (Pakistan) based on some par- de Paleontologia Miquel Crusafont (ICP) in Sabadell ticular herbivorous mammals (not including (Spain). They come from 11 fossil localities in the 250- Micromeryx) [62, 63] and rodents show higher d13C m-thick ACM composite stratigraphic sequence [5, 15], values than those of the Vallès-Penedès [64], which evi- in els Hostalets de Pierola (Vallès-Penedès Basin, Catalo- dences that Sivapithecus, which is present at some of nia, Spain; Fig. 1a–c). These localities are restricted areas those sites, might have inhabited somewhat drier envi- corresponding to a single layer (or stratigraphic hori- ronments. However, given the potential biases of the zon), usually between 0.2 and 1.0 m in thickness, from record, additional mesowear and isotopic analyses which fossil remains were collected (see Alba et al. [5] DeMiguel et al. BMC Biology (2021) 19:6 Page 12 of 15 for further details). The sequence includes more than immediate long-term dietary signal compared to the in- 200 micro- and/or macromammal localities and spans dividual’s lifespan and is a very powerful tool for docu- more than 1 Myr (12.6 to 11.4 Ma; late Aragonian, mid- menting changes in fossil habitats and climate [27, 51]. dle to late Miocene, MN7+8) [5]. The genus Micro- Mesowear for 45 teeth, each one belonging to a different meryx, widely dispersed across Europe and ubiquitous individual, was recorded. The young and old adult indi- [18], is common throughout the whole ACM sequence, viduals were identified by dental wear and discarded thus allowing for a stratigraphically continuous sam- from analyses [67]. Preference was given to upper cheek pling. The oldest occurrence of Micromeryx sp. in the teeth (usually second molars) but lower teeth were also sequence corresponds to the locality ACM/C1-Ee, with evaluated when necessary following Fraser et al. [68]. In- an estimated age of 12.38 Ma, whereas the youngest oc- dividual cusp shape and occlusal relief were scored on currence corresponds to ACM/C5-D1 (11.63 Ma) (Fig. the buccal cups. Cusp shape was categorised as either 1c). The remaining investigated ACM localities cluster sharp = pS, rounded = pR, or blunt = pB, according to in an interval spanning from 11.68 to 12.23 Ma (Fig. 1c). the degree of facet development. Occlusal relief was We analyse patterns of tooth wear and enamel stable categorised as high = pH or low= pL, depending on the isotope composition—including carbon (δ13C) and oxy- relative difference in height between the tip of the cusp gen (δ18OCO3 and δ18OPO4)—of three different Micro- and the intercusp valley. A mesowear score (MS) for meryx morphotypes (termed morphotypes 1, 2 and 3). each ACM locality was obtained in order to trace tem- They represent three distinct species, although we re- poral evolution following a five-point scoring system frain from formally naming them (see more details in [69]: 0 = high relief and sharp cusps; 1 = high relief and Additional file 1: Supplementary information, Note 1, rounded cusps; 2 = low relief and rounded cusps; 2.5 = Additional file 3: Figure S2 and Additional file 4: Figure low relief and sharp cusps; and 3 = low relief and blunt S3). cusps. We do not use the more recent scoring system by The use of a single moschid genus in palaeoecological Mihlbachler et al. [70] because it was specifically devised analysis constitutes a novelty of this study, as this group for the analysis of equid mesowear. Although the has not previously been explored from the standpoint of method could potentially work with other taxa, we pre- mesowear and stable isotopes in tandem, due to the fer to rely on the scoring system [69] customarily used small size of its teeth and the consequent difficulty in to score ruminant mesowear. See Additional file 1: Sup- gathering the data. Combining both approaches allows plementary information, Note 2 for further details on us to make robust inferences not only about the feeding this method. ecology of Micromeryx, but also in estimating climatic and environmental parameters through the ACM se- Stable isotope analysis quence. Mesowear and stable isotopes represent differ- Stable carbon and oxygen isotope compositions (δ13C ent time spans of the animal’s life and operate over and δ18O) of mammalian tooth enamel reflect diet, pri- different temporal scales—with the former recording an mary production and animal and landscape water use, average of the diet during the last years before death [25, which permit to reconstruct palaeoecological preferences 65] and the latter corresponding to the time period of and, in turn, past environmental and climatic conditions tooth mineralisation, which usually records earlier years [24, 66, 71, 72]. For herbivorous terrestrial mammals, of the animal’s ontogeny [66]. When inferring palaeo- tooth enamel δ13C values are largely controlled by the environmental change though time (in relation to tem- photosynthetic pathway followed by the ingested plants, poral changes in the composition of the primate as well as by plant water economy, and inform about the assemblage), focusing the study on Micromeryx mini- vegetation cover present in an area [73]. The δ18O values mises the biases that would be introduced by analysing a of mammalian bioapatite—both the carbonate fraction wider array of taxa, which are much less ubiquitously (δ18OCO3) and the phosphate fraction (δ18OPO4)—record and continuously recorded across the ACM sequence. the δ18O value of body water, which in turn reflects the Indeed, adding more taxa could contribute to refining fluxes and δ18O value of oxygen entering and exiting the palaeoenvironmental reconstructions for particular sites body. Within an obligate-drinking taxon, variations in and stratigraphic bins, but it would also introduce biases body water δ18O values can be interpreted as chiefly caused by the sparser record of less abundant taxa along reflecting changes in the isotopic composition of the ACM sequence. ingested water, which varies with mean annual temperature and aridity [24, 66, 71, 72]. Long-term patterns of tooth mesowear A total of 37 samples were analysed for δ13C and Mesowear is a method for determining diet (browse vs. δ OCO3, whereas 30 tooth enamel samples were ana- 18 graze) and abrasiveness of foods and exogenous items lysed for δ18OPO4. Tooth enamel was sampled using a based on gross dental wear [26]. It provides an rotary drill with a diamond-tipped dental burr under a DeMiguel et al. BMC Biology (2021) 19:6 Page 13 of 15 microscope to recover enamel from an area of the tooth Spain). D Life reconstruction of a Micromeryx male. Art by Mauricio Antón, as large as possible to avoid seasonal bias in the time of reproduced with permission of the author. mineralisation. Sampling protocol and chemical treat- Additional file 3: Figure S2. ACM Micromeryx, examples of the lower ment follows Domingo et al. [24] for carbonate and dentition of the morphotypes with remarks on their most conspicuous features. A Morphotype 1, specimens IPS90767 and IPS43689. B phosphate in bioapatite. Morphotype 2, specimen IPS44457. C Morphotype 3, specimens IPS43907 When performing isotopic analyses on fossil material, and IPS57264. diagenetic effects must be addressed before attempting Additional file 4: Figure S3. ACM Micromeryx, examples of the upper to extract palaeoecological, palaeoenvironmental and/or dentition of the morphotypes with remarks on their most conspicuous features. A Morphotype 1, specimens IPS44358 and IPS45524. B palaeoclimatic information. When both δ18OCO3 and Morphotype 2, specimens IPS45552, IPS48103 and IPS29766. C δ18OPO4 values have been determined, post mortem al- Morphotype 3, specimen IPS45382. Anatomical definitions: protocone-T, teration can be checked by calculating the difference be- T-shaped fold of enamel formed at the terminal end of the post- protocrista; metaconule-T, T-shaped fold of enamel formed at the ter- tween δ18OCO3 and δ18OPO4 (Δ18OCO3-PO4 = δ18OCO3 − minal end of the pre-metaconulecrista. δ18OPO4). This follows the premise that bioapatite in Additional file 5: Figure S4. Whittaker’s biome classification plot modern mammals have a restricted difference of ~ 8.6– showing terrestrial biome-types aas function of mean annual precipita- 9.1‰ [22, 23]. Δ18OCO3-PO4 mean value for the whole tion (MAP; in cm) and temperature (MAT; in °C). ACM localities from the three distinct environmental phases are plotted according to calculated dataset of Micromeryx samples is 8.6 ± 0.8‰, and there- values of estimated MAP and MAT using average stable carbon and oxy- fore, it is within the expected range for unaltered gen isotope values of Micromeryx enamel apatite. For each locality we bioapatite. show estimated MAP before (circles) and after (squares) correcting for palaeolatitude and palaeoaltitude. Note that these corrections result in Mean annual precipitation (MAP) from Micromeryx lower estimates. Diagram modified from Whittaker [42]. tooth enamel δ13C was estimated following Kohn [54], Additional file 6: Table S1. Individual values of mesowear and stable both with and without altitude and latitude correction. isotopes of Micromeryx specimens from the Abocador de Can Mata Estimated mean annual temperature (MAT) was also stratigraphic sequence (Vallès-Penedès Basin). calculated following a two-stepped method and using Additional file 7: Table S2. ANOVA and Tukey post hoc tests for δ13C (‰ VPDB), δ18OCO3 (‰ VSMOW), and δ18OPO4 (‰ VSMOW) values of published equations [74, 75]. See Additional file 1: Sup- Micromeryx from Abocador de Can Mata according to morphotypes and plementary information, Notes 3 and 4, for a full de- environmental phases. Values in bold are statistically significant (p < 0.05). scription of sample processing and chemical treatment, and estimated MAP and MAT calculations. Acknowledgements We thank P. L. Koch and D. Andreasen (University of California, Santa Cruz) for access to laboratory and assistance with stable isotope analyses; V. Statistical analyses Prikhod’ko (Russian Academy of Sciences) for information about water intake Chi-square (χ2) and Fisher’s exact tests with level of sig- and for the image of extant Moschus; and M. Antón for permission to reproduce his illustration of Micromeryx. We also thank AE Vitor Sousa, nificance p < 0.05 were performed based on individual Florent Rivals (Institut Català de Paleoecologia Humana i Evolució Social), mesowear score for testing for significant differences in Daniel Green (Harvard University) and three anonymous reviewers for their mesowear between the three Micromeryx morphotypes. suggestions and constructive comments on a previous version of this paper. Shapiro-Wilk tests and Q-Q plots revealed normal dis- Authors’ contributions tributions for isotopic data. Parametric analysis of vari- D.DM. and D.M.A. designed the study; D.DM. and L.D. collected and analysed ance (ANOVA) and Turkey’s post hoc pairwise the main data, while I.C.V., I.M.S. and J.M.R. provided additional contextual data; D.DM., L.D., I.C.V. and D.M.A. interpreted the data, with input from the comparisons were performed to test for significant mean remaining authors; D.DM., D.M.A. and L.D. drafted and wrote the paper. All differences at p < 0.05 between all possible pairs using a authors read and approved the final manuscript. Studentised range distribution. Locally weighted polyno- Funding mial lines were fitted to the isotopic data. Calculations This research was supported by the Agencia Estatal de Investigación (AEI) for isotopic data further include a 95% confidence re- from Spain and the European Regional Development Fund of the European gion. Descriptive and inferential statistics were con- Union (projects CGL2016-79334-P, CGL2017-82654-P and PGC2018-094955-A- I00), the Ministerio de Economía y Competitividad (RYC-2013-12470 to I.C.V), ducted using R software version 3.5.1. the Fundación Ibercaja and Universidad de Zaragoza (Spain), and CERCA Programme (Generalitat de Catalunya). This work has been performed by re- searchers from the consolidated research groups “Neogene and Quaternary Supplementary Information Vertebrate Paleobiodiversity (NQVP)” (2017 SGR 116 GRC) and “Paleoecology The online version contains supplementary material available at https://doi. and Evolutionary Biology (PEB)” (2017 SGR 960 GRC) of the Generalitat de Ca- org/10.1186/s12915-020-00939-5. talunya, and the group “Evolution of Cenozoic Mammals and Continental Paleoenvironments” (UCM 910607) of the Universidad Complutense de Additional file 1. Additional information on the moschid Micromeryx, Madrid. Fieldwork at ACM was funded by CESPA Gestión de Residuos, S.A.U. tooth-wear patterns and stable isotope analysis. Availability of data and materials Additional file 2: Figure S1. Moschus and Micromeryx. A Moschus All data generated or analysed during this study are included in this moschiferus adult male. Image by Vladimir Prikhod’ko, reproduced with published article and its additional information files. permission of the author. B Skull of M. moschiferus showing the enlarged upper canines of males. C MPZ 2006/413, skull of a juvenile male Micromeryx azanzae from Toril-3 (middle Miocene, MN7+8, Zaragoza, Ethics approval and consent to participate Not applicable. DeMiguel et al. BMC Biology (2021) 19:6 Page 14 of 15 Consent for publication 20. Sánchez IM, et al. New Hispanomeryx (Mammalia, Ruminantia, Moschidae) Not applicable. from Spain and a reassessment of the systematics and paleobiology of the genus Hispanomeryx Morales, Moyà-Solà, and Soria, 1981. J Vertebr Paleontol. 2019;39:e1602536. Competing interests 21. Alba DM, Almécija S, Moyà-Solà S. Locomotor inferences in Pierolapithecus The authors declare no competing interests. and Hispanopithecus: reply to Deane and Begun. J Hum Evol. 2010;59:143–9. 22. Bryant JD, Koch PL, Froelich PN, Showers WJ, Genna BJ. Oxygen isotope Author details partitioning between phosphate and carbonate in mammalian apatite. 1 ARAID foundation / Universidad de Zaragoza, Departamento de Ciencias de Geochim Cosmochim Acta. 1996;60:5145–8. la Tierra, and Instituto Universitario de Investigación en Ciencias Ambientales 23. Iacumin P, Bocherens H, Mariotti A, Longinelli A. Oxygen isotope analyses of de Aragón (IUCA), Pedro Cerbuna 12, 50009 Zaragoza, Spain. 2Institut Català coexisting carbonate and phosphate in biogenic apatite: a way to monitor de Paleontologia Miquel Crusafont, Universitat Autònoma de Barcelona, diagenetic alteration of bone phosphate? Earth Planet Sci Lett. 1996;142:1–6. Edifici ICTA-ICP, C/ Columnes s/n, Campus de la UAB, 08193 Cerdanyola del 24. Domingo L, et al. Late Neogene and Early Quaternary paleoenvironmental Vallès, Barcelona, Spain. 3Departamento de Geodinámica, Estratigrafía y and paleoclimatic conditions in Southwestern Europe: isotopic analyses on Paleontología Facultad de Ciencias Geológicas, Universidad Complutense de ungulate taxa. PLoS One. 2013;8:e63739. Madrid, José Antonio Novais 12, 28040 Madrid, Spain. 4Earth and Planetary 25. Urban MA, Nelson DM, Jiménez-Moreno G, Châteauneuf J-J, Pearson A, Hu Sciences Department, University of California Santa Cruz, 1156 Hight Street, FS. Isotopic evidence of C4 grasses in southwestern Europe during the early Santa Cruz, CA 95064, USA. Oligocene–middle Miocene. Geology. 2010;38:1091–4. 26. Fortelius M, Solounias N. Functional characterization of ungulate molars Received: 1 July 2020 Accepted: 10 December 2020 using the abrasion-attrition wear gradient: a new method for reconstructing paleodiets. Am Mus Novit. 2000;3301:1–35. 27. DeMiguel D. Disentangling adaptive evolutionary radiations and the role of diet in promoting diversification on islands. Sci Rep. 2016;6:29803. References 28. Merceron G. The early Vallesian vertebrates of Atzelsdorf (Late Miocene, 1. Andrews P, Harrison T, Delson E, Bernor RL, Martin L. The evolution of Austria) 13. Dental wear patterns of herbivorous ungulates as ecological western Eurasian Neogene mammal faunas. New York: Columbia University indicators. Ann Nat Hist Mus Wien. 2009;111:647–60. Press; 1996. p. 168–207. 2. Eronen JT, Rook L. The Mio-Pliocene European primate fossil record: 29. Aiglstorfer M, Semprebon GM. Hungry for fruit? – a case study on the ecology of middle Miocene Moschidae (Mammalia, Ruminantia). dynamics and habitat tracking. J Hum Evol. 2004;47:323–41. 3. Andrews P. An Ape’s View of Human Evolution. Cambridge: Cambridge Geodiversitas. 2019;41:385–99. University Press; 2015. 30. Merceron G, Schultz E, Kordos L, Kaiser TM. Paleoenvironment of Dryopithecus 4. Sukselainen L, Fortelius M, Harrison T. Co-occurrence of pliopithecoid and brancoi at Rudabánya, Hungary: evidence from dental meso- and micro-wear hominoid primates in the fossil record: an ecometric analysis. J Hum Evol. analyses of large vegetarian mammals. J Hum Evol. 2007;53:331–49. 2015;84:25–41. 31. Kaiser TM, Rössner GE. Dietary resource partitioning in ruminant communities 5. Alba DM, et al. Ten years in the dump: an updated review of the Miocene of Miocene wetland and karst palaeoenvironments in Southern Germany. primate-bearing localities from Abocador de Can Mata (NE Iberian Palaeogeogr Palaeoclimatol Palaeoecol. 2007;252:424–39. Peninsula). J Hum Evol. 2017;102:12–20. 32. Ungar PS, Kay RS. The dietary adaptations of European Miocene catarrhines. 6. Alba DM, et al. Bio- and magnetostratigraphic correlation of the Miocene Proc Natl Acad Sci U S A. 1995;92:5479–81 Groves, CP. in Handbook of the primate-bearing site of Castell de Barberà to the earliest Vallesian. J Hum Mammals of the World. Volume 2: Hoofed Mammals 2011:336–348 (Lynx Evol. 2019;132:32–46. Edicions, Barcelona). 7. Begun DR. The primate fossil record. Cambridge: Cambridge University 33. Groves CP. Handbook of the Mammals of the World. Volume 2: Hoofed Press; 2002. p. 221–40. Mammals 740. Barcelona: Lynx Edicions; 2011. p. 336–48. 8. Begun DR. The primate fossil record. Cambridge: Cambridge University 34. Moyà-Solà S, et al. Pierolapithecus catalaunicus, a new Middle Miocene great Press; 2002. p. 339–68. ape from Spain. Science. 2004;306:1339–44. 9. Begun DR. Handbook of Palaeoanthropology. Vol. 2: Primate Evolution and 35. Alba DM, et al. A new species of Pliopithecus Gervais, 1849 (Primates: Human Origins. Heidelberg: Springer; 2015. p. 1261–332. Pliopithecidae) from the Middle Miocene (MN8) of Abocador de can Mata 10. Alba DM. Fossil apes from the Vallès-Penedès Basin. Evol Anthropol. 2012;21: (els Hostalets de Pierola, Catalonia, Spain). American J Phys Anthropol. 2010; 254–69. 141:52–75. 11. Alba DM, et al. Miocene small-bodied ape from Eurasia sheds light on 36. Alba DM, et al. Brief communication: The oldest pliopithecid record in the hominoid evolution. Science. 2015;350:aab2625. Iberian Peninsula based on new material from the Vallès-Penedès Basin. 12. Begun D, Nargolwalla MC, Kordos L. New catarrhine fossils from Rudabánya American J Phys Anthropol. 2012;147:135–40. (Hungary): evidence for sympatric primates in a late Miocene swamp forest. 37. Nengo I, et al. New infant cranium from the African Miocene sheds light on Am J Phys Anthropol. 2010;141(S50):64. ape evolution. Nature. 2017;548:169–74. 13. Hardin G. The competitive exclusion principle. Science. 1960;131:1292–7. 38. Codron D, Codron J, Lee-Thorp JA, Sponheimer M, De Ruiter D. Animal 14. Casanovas-Vilar I, et al. The Miocene mammal record of the Vallès-Penedès diets in the Waterberg based on stable isotopic composition of faeces. Basin (Catalonia). C R Palevol. 2016;15:791–812. South Afr J Wildl Res. 2005;35:43–52. 15. Alba DM, et al. Los vertebrados fósiles del Abocador de Can Mata (els 39. Aiglstorfer M, Bocherens B, Böhmne M. Large mammal ecology in the late Hostalets de Pierola, l’Anoia, Cataluña), una sucesión de localidades del Middle Miocene Gratkorn locality (Austria). Palaeobiodivers Palaeoenviron. Aragoniense superior (MN6 y MN7+8) de la cuenca del Vallès-Penedès. 2014;94:189–213. Campañas 2002-2003, 2004 y 2005. Est Geol. 2006;62:295–312. 40. DeMiguel D, Alba DM, Moyà-Solà S. European pliopithecid diets revised in 16. Moyà-Solà S, et al. First partial face and upper dentition of the Middle the light of dental microwear in Pliopithecus canmatensis and Miocene hominoid Dryopithecus fontani from Abocador de Can Mata Barberapithecus huerzeleri. Am J Phys Anthropol. 2013;151:573–82. (Vallès-Penedès Basin, Catalonia, NE Spain): taxonomic and phylogenetic 41. DeMiguel D, Alba DM, Moyà-Solà S. Dietary specialization during the implications. Am J Phys Anthropol. 2009;139:126–45. evolution of Western Eurasian hominoids and the extinction of European 17. Casanovas-Vilar I, et al. An updated biostratigraphy for the late Aragonian Great Apes. PLoS One. 2014;9:e97442. and the Vallesian of the Vallès-Penedès Basin (Catalonia). Geol Acta. 2016;14: 42. Bryant JD, Froelich PN. A model of oxygen isotope fractionation in body 195–217. water of large mammals. Geochim Cosmochim Acta. 1995;59:4523–37. 18. Sánchez IM, Morales J. Distribución biocronológica de los Moschidae 43. Kohn MJ. Predicting animal δ18O: accounting for diet and physiological (Mammalia; Ruminantia) en España. Estud Geol. 2006;62:533–46. adaptation. Geochim Cosmochim Acta. 1996;60:4811–29. 19. Sánchez IM, Morales J. Micromeryx azanzae sp. nov. (Ruminantia: Moschidae) 44. Sánchez IM, Cantalapiedra JL, Ríos M, Quiralte V, Morales J. Systematics and from the middle-upper Miocene of Spain, and the first description of the evolution of the Miocene three-horned palaeomerycid ruminants cranium of Micromeryx. J Vertebr Paleontol. 2008;28:873–85. (Mammalia, Cetartiodactyla). PLoS One. 2015;10:e0143034. DeMiguel et al. BMC Biology (2021) 19:6 Page 15 of 15 45. Prikhod'ko VI, Deer M. Origins, taxomomy, ecology, behaviour and 72. Domingo L, Domingo MS, Koch PL, Morales J, Alberdi MT. Carnivoran communication. Moscow: GEOS; 2003. resource and habitat use in the context of a Late Miocene faunal turnover 46. Liu Z, Sheng H. Effect of habitat fragmentation and isolation on the episode. Palaeontology. 2017;60:461–83. population of Alpine musk deer. Russ J Ecol. 2002;33:121–4. 73. Ehleringer JR, Monson RK. Evolutionary and ecological aspects of 47. Zachos JC, Dickens GR, Zeebe RE. An early Cenozoic perspective on photosynthetic pathway variation. Ann Rev Ecol Evol System. 1993;24:411–39. greenhouse warming and carbon-cycle dynamics. Nature. 2008;451:279–83. 74. D'Angela D, Longinelli A. Oxygen isotopes in living mammal’s bone 48. Postigo-Mijarra JM, Barrón E, Gómez Manzaneque F, Morla C. Floristic phosphate: further results. Chem Geol. 1990;86:75–82. changes in the Iberian Peninsula and Balearic Islands (south-west Europe) 75. Rozanski K, Araguás-Araguás L, Gonfiantini R. Climate change in continental during the Cenozoic. J Biogeogr. 2009;36:2025–43. isotopic records: American Geophysical Union; 1993. p. 1–36. 49. Jiménez-Moreno G, Fauquette S, Suc J-P. Miocene to Pliocene vegetation reconstruction and climate estimates in the Iberian Peninsula from pollen data. Rev Palaeobot Palynology. 2010;162:403–15. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in 50. Whittaker RH. Classification of natural communities. Bot Rev. 1962;28:1–239. published maps and institutional affiliations. 51. DeMiguel D, Rook L. Understanding climate’s influence on the extinction of Oreopithecus (late Miocene, Tusco-Sardinian paleobioprovince, Italy). J Hum Evol. 2018;116:14–26. 52. Casanovas-Vilar I, et al. Biochronological, taphonomical and paleoenvironmental background of the fossil great ape Pierolapithecus catalaunicus (Primates, Hominidae). J Hum Evol. 2008;55:589–603. 53. Almécija S, Alba DM, Moyà-Solà S. Pierolapithecus and the functional morphology of Miocene ape hand phalanges: paleobiological and evolutionary implications. J Hum Evol. 2009;57:284–97. 54. Kohn MJ. Carbon isotope compositions of terrestrial C3 plants as indicators of (paleo) ecology and (paleo)climate. Proc Natl Acad Sci U S A. 2010;107:19691–5. 55. Richgelt T, et al. Elevated CO2, increased leaf-level productivity, and water- use efficiency during the early Miocene. Clim Past. 2020;16:1509–21. 56. DeMiguel D, et al. First evidence of Tethytragus Azanza and Morales, 1994 (Ruminantia, Bovidae), in the Miocene of the Vallès-Penedès Basin (Spain). J Vertebr Paleontol. 2012;32:1457–62. 57. Pina M, Alba DM, Moyà-Solà S, Almécija S. Femoral neck cortical bone distribution of dryopithecin apes and the evolution of hominid locomotion. J Hum Evol. 2019;136:102651. 58. Senut B. Les restes post-crâniens des Pliopithecidae (Primates) de Sansan. Mém Mus Nat Hist Nat. 2012;203:535–58. 59. Arias-Martorell J, et al. Morphological affinities of the proximal humerus of Epipliopithecus vindobonensis and Pliopithecus antiquus: suspensory inferences based on a 3D geometric morphometrics approach. J Human Evol. 2015;80:83–95. 60. Alba DM, et al. Middle Miocene tragulid remains from Abocador de can Mata: the earliest record of Dorcatherium naui from Western Europe. Geobios. 2011;44:135–50. 61. Casanovas-Vilar I, et al. Oldest skeleton of a fossil flying squirrel casts new light on the phylogeny of the group. eLife. 2018;7:e39270. 62. Nelson SV. Isotopic reconstructions of habitat change surrounding the extinction of. Sivapithecus, a Miocene hominoid, in the Siwalik Group of Pakistan. Paleogeogr Paleoclimatol Paleoecol. 2007;243:204–22. 63. Badgley C, Barry JC, Morgan ME, Nelson SV, Behrensmeyer AK, Cerling TE, Pilbeam D. Ecological changes in Miocene mammalian record show impact of prolonged climatic forcing. Proc Natl Acad Sci U S A. 2008;105:12145–9. 64. Casanovas-Vilar I, Kimura Y, Flynn LJ, Pilbeam D, Moyà-Solà S, Alba DM. Rodent stable isotopes provide new data on Miocene ape environments in the Siwaliks and the Vallès-Penedès Basin [Abstract]. Am J Phys Anthropol. 2020;171:46. 65. Ackermans N, et al. Mesowear represents a lifetime signal in sheep (Ovis aries) within a long-term feeding experiment. Palaeogeogr Palaeoclimatol Palaeoecol. 2020;553:109793. 66. Koch PL. Stable Isotopes in Ecology and Environmental Science, vol. 99. Boston: Blackwell Publishing; 2007. p. 154. 67. Rivals F, Mihlbachler MC, Solounias N. Effect of ontogenetic-age distribution in fossil samples on the interpretation of ungulate paleodiets using the mesowear method. J Vertebr Paleontol. 2007;27:763–7. 68. Fraser D, Zybutz T, Lightner E, Theodor JM. Ruminant mandibular tooth mesowear: a new scheme for increasing paleoecological sample sizes. J Zool. 2014;294:41–8. 69. Rivals F, Schulz E, Kaiser TM. Late and middle Pleistocene ungulates dietary diversity in Western Europe indicate variations of Neanderthal paleoenvironments through time and space. Quat Sci Rev. 2009;28:3388–400. 70. Mihlbachler MC, Rivals F, Solounias N, Semprebon GM. Dietary change and evolution of horses in North America. Science. 2011;331:1178–81. 71. Domingo L, Koch PL, Grimes ST, Morales J, López-Martínez N. Isotopic paleoecology of mammals and the Middle Miocene Cooling event in the Madrid Basin (Spain). Palaeogeogr Palaeoclimatol Palaeoecol. 2012;339-341:98–113.