Research Article

Miocene small-bodied ape from Eurasia sheds light on hominoid evolution

See allHide authors and affiliations

Science  30 Oct 2015:
Vol. 350, Issue 6260, aab2625
DOI: 10.1126/science.aab2625

Meet your gibbon cousin

Apes are divided into two groups: larger-bodied apes, or hominoids, such as humans, chimps, and gorillas; and smaller-bodied hylobatids, such as gibbons. These two lineages are thought to have diverged rather cleanly, sharing few similarities after the emergence of crown hominoids. Alba et al. describe a new ape from the Miocene era that contains characteristics from both hominoids and small-bodied apes (see the Perspective by Benefit and McCrossin). Thus, early small-bodied apes may have contributed more to the evolution of the hominoid lineage than previously assumed.

Science, this issue p. 10.1126/science.aab2625; see also p. 515

Structured Abstract


Reconstructing the ancestral morphotype from which extant hominoids (apes and humans) evolved is complicated by the mosaic nature of ape evolution, the confounding effects of independently evolved features (homoplasy), and the virtual lack of hylobatids (gibbons and siamangs) in the Miocene fossil record. For several decades, small-bodied anthropoid primates from Africa and Eurasia have not played an important role in this debate, because they generally lack the shared derived features of extant catarrhines (hominoids and Old World monkeys) and are thus considered to precede their divergence. Even some small-bodied catarrhines from Africa (dendropithecids), considered to be stem hominoids by some authors, are viewed as more primitive than the larger-bodied stem ape Proconsul. This has led to the assumption that hylobatids are a dwarfed lineage that evolved from a larger-bodied and more great ape–like common ancestor with hominids (great apes and humans).


Here we describe a new genus of small-bodied (4 to 5 kg) ape from the Miocene (11.6 Ma), discovered in the Abocador de Can Mata stratigraphic series (Vallès-Penedès Basin, northeast Iberian Peninsula), that challenges current views on the last common ancestor of extant hominoids. This genus is based on a partial skeleton that enables a reliable reconstruction of cranial morphology and a detailed assessment of elbow and wrist anatomy. It exhibits a mosaic of primitive (stem catarrhine–like) and derived (extant hominoid–like) features that forces us to reevaluate the role played by small-bodied catarrhines in ape evolution.


The new genus retains some features that are suggestive of generalized above-branch quadrupedalism, but it possesses more extensive hominoid-like postcranial features (mostly related to enhanced forearm rotation and ulnar deviation capabilities) than those convergently displayed by atelids. Its overall body plan is more compatible with an emphasis on cautious and eclectic climbing, combined with some degree of below-branch forelimb-dominated suspension (although less acrobatic than in extant gibbons). Its relative brain size implies a monkey-like degree of encephalization (similar to that of hylobatids but below that of great apes), and dental microwear indicates a frugivorous diet. From a phylogenetic viewpoint, the new genus combines craniodental and postcranial primitive features (similar to those of dendropithecids) with multiple derived cranial and postcranial features shared with extant hominoids. Some cranial similarities with gibbons would support a closer phylogenetic link between the new genus and hylobatids. However, this possibility is not supported by the total evidence. A cladistic analysis based on more than 300 craniodental and postcranial features reveals that the new genus is a stem hominoid (preceding the divergence between hylobatids and hominids), although more derived than previously known small catarrhines and Proconsul.


As the first known Miocene small-bodied catarrhine to share abundant derived features with extant hominoids, the new genus indicates a greater morphological diversity than previously recognized among this heterogeneous group, and it provides key insight into the last common ancestor of hylobatids and hominids. Our cladistic results, coupled with the chronology and location of the new genus, suggest that it represents a late-surviving offshoot of a small African stem hominoid that is more closely related to crown hominoids than Proconsul is. These results suggest that, at least in size and cranial morphology, the last common ancestor of extant hominoids might have been more gibbon-like (less great ape–like) than generally assumed.

Cranial reconstruction and life appearance.

Artist’s representation of the cranial reconstruction (in frontal view) and of the life appearance (in lateral oblique view) of the new genus of small-bodied ape from the Iberian Miocene. [Artwork by M. Palmero]


Miocene small-bodied anthropoid primates from Africa and Eurasia are generally considered to precede the divergence between the two groups of extant catarrhines—hominoids (apes and humans) and Old World monkeys—and are thus viewed as more primitive than the stem ape Proconsul. Here we describe Pliobates cataloniae gen. et sp. nov., a small-bodied (4 to 5 kilograms) primate from the Iberian Miocene (11.6 million years ago) that displays a mosaic of primitive characteristics coupled with multiple cranial and postcranial shared derived features of extant hominoids. Our cladistic analyses show that Pliobates is a stem hominoid that is more derived than previously described small catarrhines and Proconsul. This forces us to reevaluate the role played by small-bodied catarrhines in ape evolution and provides key insight into the last common ancestor of hylobatids (gibbons) and hominids (great apes and humans).

Apes and humans (hominoids) diverged from Old World monkeys (cercopithecoids) by the Oligocene [≥25 million years ago (Ma)] (1, 2) and subsequently diversified in both Africa and Eurasia during the Miocene (23 to ~5 Ma) (3). They are currently represented by crown hominoids (4)—i.e., the small-bodied hylobatids (gibbons and siamangs) and the larger-bodied hominids (great apes and humans), which diverged from one another by the early Miocene (~17 Ma) (1). Reconstructing the ancestral morphotype from which extant apes and humans evolved is a challenging task (57), given the mosaic nature of evolution (8), the confounding effects of independently evolved features (homoplasy) (5, 9), the conflicting evidence provided by Miocene great apes (6, 9), and the incomplete and fragmentary nature of the hominoid fossil record [in particular, the virtual lack of fossil gibbons until at least the latest Miocene (10, 11)]. Thus, although extant hominoids share numerous derived features, particularly in the trunk and forelimb, it is uncertain to what extent these characteristics were inherited from their last common ancestor, whose morphotype is still under discussion (5, 6, 11, 12).

The earliest hypotheses postulated a small-bodied gibbon-like ancestor (13), and for many years, fossil small-bodied catarrhines (anthropoid primates from Africa and Eurasia) were considered to be broadly ancestral to gibbons (14, 15). Today, hylobatids are generally considered to be a specialized and probably dwarfed lineage, evolved from a larger and more great ape–like last common ancestor with hominids (6, 16). This is because known small-bodied extinct catarrhines, such as the African dendropithecids (including at least Dendropithecus, Simiolus, and Micropithecus) and the Eurasian pliopithecoids (Pliopithecus, Epipliopithecus, and allied genera), lack most of the synapomorphies (shared derived features) of crown catarrhines (1721). Even dendropithecids, which are more derived than pliopithecoids and are currently interpreted by some authors as stem hominoids (preceding the hylobatid-hominid divergence) (3, 8, 22), are considered to be more basal than the stem ape Proconsul (8, 2123). Here we describe a new Miocene small-bodied ape from Spain, Pliobates cataloniae gen. et sp. nov. (24). In some primitive features, this new primate resembles previously known small-bodied catarrhines such as dendropithecids, but it differs from them and from Proconsul by displaying multiple crown-hominoid derived features. This mosaic provides key insight into ape evolution by forcing us to reevaluate the role played by small-bodied catarrhines in the emergence of crown hominoids.


The new taxon is described on the basis of a partial skeleton (IPS58443; table S1) that is permanently housed in the collections of the Institut Català de Paleontologia Miquel Crusafont (ICP) in Sabadell, Spain. The main cranial fragments were initially discovered on 3 January 2011 by a team of paleontologists from the company FOSSILIA Serveis Paleontològics i Geològics, directed by one of the authors (J.M.R.), while overseeing the excavation works performed by digging machines at the Can Mata landfill (els Hostalets de Pierola, Catalonia, Spain). In the following days, postcranial elements were found at the field during excavation of the same stratigraphic level, and the remaining postcranial small bones and other small fragments were subsequently recovered by screen-washing the excavated sediments. All the fossils therefore come from a single horizon, which was labeled ACM/C8-A4 (Abocador de Can Mata, Cell 8, sector A, locality 4).

The ACM local stratigraphic composite series (2527) is located in the Vallès-Penedès Basin (northeast Iberian Peninsula), a half-graben trending north-northeast–south-southwest and limited by the Catalan Coastal Ranges; this structure was generated by the rifting of the northwest Mediterranean during the Neogene (28, 29). The basin infill mostly consists of marginal alluvial fan sediments that have provided a rich fossil record of Miocene terrestrial vertebrates (30). The area of els Hostalets de Pierola has thick middle to late Miocene sequences that were deposited in distal-to-marginal inter-fan zones of coalescing alluvial fan systems (27). The ACM composite series, about 250 m in thickness, includes more than 200 formally defined localities that can be accurately dated based on lithostratigraphic, magnetostratigraphic, and biostratigraphic correlation (26, 27, 31). The whole series is late Aragonian, and, based on updated chron boundaries (32), it spans from ~12.6 to 11.5 Ma. Locality ACM/C8-A4 is correlated to chron C5r.2n (11.657 to 11.592 Ma) with an interpolated age of 11.628 Ma, which is close to the middle/late Miocene boundary, as defined by the base of the Tortonian and currently dated to 11.625 Ma (32).


Cranial reconstruction

The cranium was preserved in a main piece (IPS58443.1) with parts of the neurocranium, basicranium, muzzle, and right maxilla (Fig. 1, A to C); a medium-sized fragment with the left maxilla (IPS58443.2); and other smaller fragments that were found in close spatial association or recovered by screen-washing the surrounding sediments. These other fragments include part of the occipital and right parietal (IPS58443.3), the right glenoid region (IPS58443.4), the right occipital condyle (IPS58443.5), a fragment of the right orbital margin (IPS58443.6), the left orbital margin and temporal process of the zygomatic (IPS58443.12), a parietal fragment (IPS58443.11), and several other minor fragments whose location cannot be determined (IPS58443.7 to IPS58443.10, IPS58443.13, and IPS58443.37). The mandible was not preserved, except for a fragment of the right ramus with the condyle (IPS58443.14). The main fragment (IPS58443.1) is composed of several bone fragments that are crushed against each other and somewhat displaced from their anatomical position, but (like the remaining specimens) not plastically distorted.

Fig. 1 Cranium and dentition.

(A to C) Cranium of the holotype (IPS58443) of Pliobates cataloniae gen. et sp. nov. The main cranial fragments, including the basicranium and the right palate, are shown in basal view (A); details of the right palatal fragment are shown in left-lateral (B) and right-lateral (C) views. (D) Detail of the right postcanine teeth, in occlusal view (mesial is to the right).

Several bone fragments of IPS58443.1 were individualized through careful mechanical preparation, but for many other bone fragments, their fragility and state of preservation precluded a complete isolation from each other and/or from the embedding matrix. Therefore, for conservational reasons, no complete preparation of the specimen was performed, and a virtual three-dimensional (3D) reconstruction was undertaken instead. The larger specimens (IPS58443.1 and IPS58443.2) were scanned at the American Museum of Natural History (New York) with a high-resolution computed tomograph (CT) system (Phoenix v|tome|x s180, General Electric Measurement & Control Solutions, Hanover, Germany), using a nanofocus x-ray tube. Different protocols were used for these two cranial fragments: 160 kV voltage, 1.4 mA current, 0.2 mm Cu filter, and magnification of 2.10013075, obtaining 1600 slices (virtual cross-sectional images) of 0.2 mm in thickness and a pixel size of 0.09523217 mm (IPS58443.1); and 145 kV voltage, 1.3 mA current, 0.1 mm Cu filter, and a magnification of 2.93159482, obtaining 1500 slices of 0.2 mm in thickness and a pixel size of 0.06822225 mm (IPS58443.2). Raw CT data were imported into VGStudio Max 2.1 and exported (as a stack of TIFF files) to Avizo 7.0 and Geomagic 2012 for segmentation, repositioning, mirroring, and/or visualization. CT-scanned bone fragments were segmented slice by slice by digitally removing the matrix with the aid of differential bone and sediment densities, using the semiautomatic thresholding tools of Avizo 7.0. Additional small isolated bone fragments not surrounded by matrix were scanned with a 3D desktop laser scanner (NextEngine) at high definition and with a dimensional accuracy of 0.13 mm (Macro Mode); the resulting 3D models were exported to Geomagic 2012 to align and repair the meshes. The preserved bone fragments generally had clean fractures that allowed them to be easily matched with other fragments. Up to 39 3D virtual models of bone fragments, including pieces of the premaxillae and maxillae with teeth, lacrimals, zygomatics, frontal, parietals, temporals, occipital, and pterygoids, were thus digitally assembled and repositioned using Avizo 7.0 and Geomagic 2012 on the basis of preserved morphology, fracture congruence, and bilateral symmetry. Once the preserved fragments were adequately positioned, areas preserved only in one side of the cranium were mirrored. The individual models were then merged to obtain the definitive 3D virtual model, which was also 3D-printed with a ZPrinter450 at the Universitat Autònoma de Barcelona for visualization and comparative purposes.

To reconstruct the muzzle area, we primarily relied on IPS58443.1, in which the right premaxilla (with the alveoli for I1 to C1) and the maxilla (with P3 to M3, the lower portion of the nasal aperture, most of the palatine process, and the intermaxillary suture) are almost completely preserved. The left maxillary specimen (IPS58443.2) is less complete and only includes the M2 to M3 series and a smaller portion of the intermaxillary suture, as well as the zygomatic root. Palate width and shape were thus reconstructed by mirroring the right fragment onto the left one, then further adjusting it based on the fit onto the left portion of the intermaxillary suture and the alignment with the left molars. With regard to the orbital area, although the orbits are not completely preserved on either side, their profile can be reconstructed with reasonable accuracy based on the available fragments: IPS58443.1 includes, in several fragments, most of the interorbital area of the frontal and the superior orbital rims, the maxillary portion of the right inferior orbital rim, a zygomatic portion of the right lateral orbital rim, and both lacrimals; IPS58443.12 includes a larger lateral portion of the left orbit that encompasses part of the temporal process of the zygomatic arch and the frontomaxillary suture. Although the zygomaticomaxillary suture is not preserved on either side, the missing portion of the zygomatic process of the maxilla on the left side is minimal. Therefore, it is possible to situate the left orbit relative to the muzzle, on the basis of the orientation between the zygomatic process of IPS58443.12 and the zygomatic root preserved in IPS58443.2. Similarly, although the frontozygomatic suture is not preserved at the end of the left zygomatic process of IPS58443.1, only a very small portion is missing; thus, the preservation of the suture in the zygomatic portion of the lateral orbital rim in IPS58443.12 enables a reliable reconstruction of the superolateral aspect of the orbit, based on the curvature of these two fragments. The inferior orbital rim can be also reconstructed using its right maxillary portion, which is preserved in IPS58443.1, by mirroring onto the left side. Moreover, the preservation of both superior orbital rims in the frontal fragments of IPS58443.1 further allows the right orbit to be positioned relative to the muzzle, in spite of the fact that the right zygomaticomaxillary suture is not preserved. Overall, although the missing fragments of the orbital rims introduce some degree of uncertainty, the curvature of the various preserved portions enables a confident reconstruction of orbital size and shape. The preserved portions of the zygomatic and maxilla further enable reliable positioning of the orbits relative to the muzzle. Because the missing portions are very small, other possible reconstructions do not differ substantially from the one provided here.

To reconstruct the neurocranium, we relied on the relative spatial position of the braincase fragments in the crushed main specimen (IPS58443.1). Because these fragments are not plastically deformed, they show the original curvature of the braincase, which considerably facilitated the reconstruction of its overall shape. The frontals are preserved in six fragments that enable the reconstruction of most of the braincases morphology (including the whole supraorbital area and the connection with the parietals on the left side). The parietals are preserved in 19 fragments, which have well-preserved sutures and neat fractures. The latter fact, together with the spatial association among the various fragments, enables the reconstruction of the posterior contour of the cranial vault. Large portions of both temporals, including the tympanic bullae, are also preserved in IPS58443.1. The right temporal is more completely preserved than the left one; it includes not only the glenoid fossa and the temporal process of the zygomatic arch, but also the suture with the parietal. These preserved pieces, mirrored onto the left side, enabled the reconstruction of the lateral aspects of the neurocranium. The occipitals are less completely preserved, being restricted to three basioccipital fragments and one occipital fragment, which include the occipital condyles and enable a reliable orientation of the preserved profile of the foramen magnum. One of these fragments connects with the left temporal, thereby enabling the reconstruction of the basicranium. Only minor displacements, caused by diagenesis, had to be corrected; these included displacements between the left bulla and the left occipital condyle and basioccipital, and the artifactual separation of the occipital-temporal suture. Once these elements were slightly repositioned into their correct anatomical place, they fit congruently with the surrounding elements. Last, the pterygoid wings are also preserved on the right side, attached to the palatine but slightly crushed, so that some realignment of this bone was necessary. The connection of the braincase with the superior portion of the face is possible through the frontal fragments, which on the left side are continuous with the parietal fragments. The connection between the braincase and the lower portion of the face is possible, not only because of the preservation of the superior orbital rims in the frontal and of a maxillary portion of the right inferior orbital rim, but also because part of the right pterygoid wing (anatomically adjacent to the palatine) was preserved in its original anatomical position, attached to one of the bone fragments of IPS58443.1 (which includes a large portion of the basicranium and temporal). This pterygoid fragment enables a reliable reconstruction of the relative position of the muzzle and the neurocranium via the basicranium, and it further confirms the correctness of the reconstruction of orbital shape and position. The position of the muzzle was further tested on the basis of the congruent alignment on the right side between the temporal process of the zygomatic in IPS58443.12 and the zygomatic process of the temporal in IPS58443.4.

Dental measurements and body mass estimation

Standard dental measurements of mesiodistal length (MD) and maximum buccolingual breadth (BL) were taken (in millimeters) to assess dental size and proportions by means of comparative bivariate plots of log-transformed BL versus MD for each of the upper molars. Data for extant and extinct small-bodied catarrhines were taken from the literature (17, 3346) or measured by one of the authors (D.M.A.). Based on MD and BL measurements, tooth square area (A, in square millimeters) was computed for postcanine teeth to estimate the body mass (BM, in kilograms) of Pliobates, using allometric equations of BM versus A (47). BM of Pliobates and, for comparative purposes, of Epipliopithecus vindobonensis was also estimated based on allometric equations, using postcranial estimators (48) separately for samples of hominoids and catarrhines (i.e., hominoids plus cercopithecoids), based on sex-species mean data. Six postcranial BM estimators were used (49), including three surface areas (for the tibia, humerus, and radius) and three linear measurements (for the humerus and radius; tables S2 and S3).

Humeral torsion and arm angle

Humeral torsion, or the orientation of the humeral head relative to the mediolateral axis of the distal humerus (50), cannot be directly computed in the described skeleton because the humeral head is missing. Accordingly, humeral torsion was kindly estimated by S. Larson, following her methodology (50) for incomplete humeri lacking the proximal end (based on the bisector of the bicipital groove, indicating the orientation of the humeral head), as well as using the posterior buttress for the humeral head (as a reference for the head axis). Measurements were obtained from a cast of the two humeral diaphyseal fragments of the holotype specimen. This humerus was originally preserved as a single specimen, with a crack filled with sediment at about midshaft level, the two fragments being slightly crushed against each other in the proximodistal direction. After manual separation of the matrix, the original shape of the diaphysis was reconstructed by correctly aligning casts of the two fragments.

The arm angle (or carrying angle at the elbow joint), which is the angle between the long axes of the humerus and ulna (arm angles <0° imply medial deviation of the ulna) (51), was computed from a photograph of the rearticulated original specimens.

Postcranial proportions

The degree of elongation of the forearm, the arm, and the forelimb as a whole were assessed by means of allometric regressions of radius, humerus, and radius-plus-humerus length (in millimeters), respectively, relative to BM (in kilograms). Allometric equations (table S4) were derived based on data taken from the literature (48, 5254) or kindly provided by E. Sarmiento to S.M.S. Regressions for anthropoid primates were based on n = 54 sex-species means, corresponding to 31 species from 17 genera; hylobatids, orangutans, Ateles, and Brachyteles were excluded from the regressions because they are clear outliers. Allometric residuals were computed for the studied sample as well as for Pliobates and Epipliopithecus; humeral length in the former, given the lack of the humeral head, was estimated by taking into account the proportions of the humerus of Epipliopithecus.

To assess the size of the triquetrum relative to the hamate, the size of each bone was computed as the geometric mean of three measurements representing their maximum dimensions: maximum mediolateral breadth, dorsopalmar height, and proximodistal length for the triquetrum; and maximum proximodistal length, dorsopalmar height, and mediolateral breadth for the hamate. Comparative data were kindly provided by T. Kivell; her metrics and sample (including 28 anthropoid species from 16 genera) are described in the literature (55, 56).

Cranial capacity and encephalization

For preservational reasons, it was not possible to compute cranial capacity (CC, in cubic centimeters) from the virtual endocast. Therefore, CC was estimated using published allometric equations based on various external neurocranial measurements (57): maximum width of the braincase base (CW, in millimeters); vertical height of braincase (CH, in millimeters); chord of the midline suture through occipitals, parietals, and frontals (CL, in millimeters); modulus of the above-mentioned linear measurements (CO, in millimeters), computed as CO = CW + CH + CL; the product of the above-mentioned linear dimensions (PR, in cubic millimeters), computed as PR = CW × CH × CL; and foramen magnum area (FMA, in square centimeters), computed as FMA = (π/4) × FMW × FML, where FMW is foramen magnum width, and FML is foramen magnum length (both in millimeters). To assess encephalization, we relied on lower taxonomic–level metrics of relative brain size, which are significantly correlated with general domain cognitive abilities in primates (58, 59). Encephalization residuals (ER) were computed as ER = ln CCobserved – ln CCpredicted, by using the ordinary least-squares cercopithecoid allometric regression (ln CC = 0.4778 ln BM + 3.457), whereas encephalization constants (EC) were computed as EC = CC / BM0.28 (58, 59). Sex-species mean data for extant species were taken from the literature (58). In addition to the estimates derived for Pliobates in this study, published data for various extinct catarrhines (female Aegyptopithecus zeuxis, male Victoriapithecus macinnesi, female Proconsul nyanzae, male Oreopithecus bambolii, and female Hispanopithecus hungaricus) were also included in the analyses (37, 58, 6062).

Dental microwear

Paleodietary reconstruction was based on dental microwear analysis (63). The M1 of Pliobates was selected, because it exhibits clearer and larger phase II crushing and grinding facets than the remaining molars. Occlusal surfaces were examined through an environmental scanning electron microscope at ×500 magnification in secondary emissions mode and at 20 kV, following established procedures (63). An area of standardized size, corresponding to 0.02 mm2 on the original facet (64), was analyzed using the custom software package Microware 4.02 (65). Three microscopic variables were measured: percentage of pits, breadth of striations, and breadth of pits. Striations and pits were categorized by following an arbitrarily set length-to-width ratio of 4:1 (63, 6668). We compared our results with those previously derived for a sample of 11 extant anthropoids with well-known diets (63, 69, 70), partitioned into three distinct dietary categories (66), as well as with those reported for extinct catarrhines (64, 6668, 71), including both pliopithecoids and hominoids. Microwear data were analyzed by means of canonical variates analysis, using SPSS Statistics 19 software.

Phylogenetic analysis

A cladistic analysis based on maximum parsimony was performed with PAUP* (Phylogenetic Analysis Using Parsimony) version 4.0 for Unix (72), with the search command “branch-and-bound,” based on a taxon-character data matrix of 319 characters and 20 taxa (tables S5 and S6). This matrix was coded anew by the authors, although it was partially based on character statements taken from the literature (8, 22, 7376). All but 10 characters were treated as unordered, whereas inapplicable characters were treated as missing data. Clade robusticity was assessed by means of bootstrap analysis (10,000 replicates) and Bremer support indices. For the most-parsimonious cladograms, the following metrics were computed: consistency index (excluding uninformative characters), retention index, and rescaled consistency index. Character polarity was determined using the outgroup method, with the stem catarrhine Aegyptopithecus being used as such. Ingroup taxa include the stem catarrhine Saadanius, two cercopithecoids (the extant Macaca and the extinct Victoriapithecus), extant hylobatids, extant (Pongo, Gorilla, and Pan) and extinct (Pierolapithecus and Hispanopithecus) great apes, and a wide representation of small-bodied fossil catarrhines from Africa [two dendropithecids, including Micropithecus and Dendropithecus-plus-Simiolus (coded simultaneously)] and Eurasia (six pliopithecoids, including Pliopithecus, Epipliopithecus, Dionysopithecus, Barberapithecus, Anapithecus, and Plesiopliopithecus). The phylogenetic placement of Saadanius, Micropithecus, and most pliopithecoids (except Epipliopithecus) should be considered with caution, because they have a large proportion of missing data. For this reason, the analysis was also performed with these taxa removed.

Systematic paleontology

Order Primates Linnaeus, 1758. Suborder Haplorrhini Pocock, 1918. Infraorder Anthropoidea Mivart, 1864. Parvorder Catarrhini É. Geoffroy Saint-Hilaire, 1812. Superfamily Hominoidea Gray, 1821. Family Pliobatidae fam. nov. Type genus: Pliobates gen. nov., whose diagnosis is as for its type (and only) species, described below.

Pliobates cataloniae gen. et sp. nov.

Holotype: IPS58443, a partial skeleton with an associated skull (Fig. 1 and movie S1), housed at the ICP. It is composed of 70 bones and bone fragments (table S1) found in close spatial association, which, given the lack of repeated elements, are attributed to a single adult female individual (based on the small canine alveolus), with an estimated body mass of 4 to 5 kg (tables S7 and S8). It includes large portions of the cranium with postcanine maxillary teeth (Table 1), a mandibular fragment, a partial left forelimb (nearly complete humerus, radius, partial ulna, carpals, and bones of the manual rays), more fragmentary elements of the right forelimb, and bones from the hind limb.

Table 1 Dental measurements.

Standard dental measurements to assess dental size and proportions were taken to the nearest 0.1 mm in the holotype (IPS58443) of Pliobates cataloniae gen. et sp. nov. MD, mesiodistal length (in millimeters); BL, buccolingual width (in millimeters); BLI, breadth/length index (in percent), computed as BL/MD × 100. Dashes indicate lack of data due to incomplete preservation.

View this table:

Type locality: ACM/C8-A4 (els Hostalets de Pierola, Catalonia, Spain), in the ACM stratigraphic series (Vallès-Penedès Basin, northeast Iberian Peninsula).

Age, stratigraphic position, and distribution: Only known from the type locality, which has an estimated age of 11.6 Ma (middle/late Miocene boundary) and is thus somewhat younger than all other ACM hominoid- and pliopithecoid-bearing localities (9, 31, 33, 77, 78), the latter of which have been dated to 11.7 to 11.9 Ma [updated from (77, 78)].

Etymology: Genus name from the Latin plio- (itself from the Greek, meaning “greater in extent”) and from the Greek bates (meaning “the one that walks or haunts”). The name is a contraction of the genus names Pliopithecus (“more ape”) and Hylobates (“the one that walks in the woods or in the trees”), in allusion to the small body size and the mosaic of primitive (stem catarrhine–like) and derived (crown hominoid) features displayed by the new taxon. The species epithet is the genitive of the female substantive “Catalonia,” the Latin name of Catalunya (in which the type locality is situated).


Small-bodied catarrhine primate (estimated female BM of 4 to 5 kg). Dental formula Female upper canines moderately compressed. Upper cheek teeth low-crowned and with subpyramidal, moderately peripheral, and inflated cusps. Upper premolars relatively broad and ovoid, P4 smaller than P3, both with heteromorphic cusps, a markedly convex lingual contour and a distinct lingual cingulum (more developed in the P4), a distinct transverse crest separating the restricted mesial fovea from the extensive trigon basin, and the postparacrista forming an abrupt angle with the distal marginal ridge. Upper molars only moderately broader than long, with markedly convex lingual profiles; buccal cusps quite peripheral and buccal cingula discontinuous; lingual cingula relatively well developed, shelf-like, and C-shaped, but not surrounding the hypocone (which is distinct and more peripheral than the protocone); mesial fovea restricted, with an obliquely directed preprotocrista, and trigon basin extensive, being separated by a continuous crista obliqua from the slightly smaller distal fovea, which displays no hypocone-metacone crest. M2 slightly larger than the M1, and M3 shorter and trapezoidal (due to the oblique buccal margin, with a centrally situated metacone and a rudimentary hypocone).

Face small but with a distinct snout, the anterior portion of the nasals being almost parallel to the palate. Maxillary sinus large and frontal sinus present but small. Nasal aperture narrow. Nasoalveolar clivus short, with an open palatine fenestra. Anteriorly slightly narrow palate with somewhat convergent upper tooth rows. Zygomatic root moderately high. Orbits subcircular, large, and frontated, with telescopic orbital rims located over the P4. Estimated cranial capacity (69 to 75 cm3) indicating a monkey-like degree of encephalization. External auditory meatus tubular but short and not completely ossified, with a V-shaped end and its anterior portion fused with the postglenoid process. Carotid foramen perforating the bulla posterodistally, and carotid canal horizontally and anteriorly oriented. Spinosum and postglenoid foramina absent. Jugular foramen large and ventrally visible.

Humerus without entepicondylar foramen and capitular tail, with a well-developed capitulum, and a narrow and deep zona conoidea. Radial head rounded and not very tilted, with a markedly beveled surface for articulation with the humeral zona conoidea, the articular surface for the ulnar radial notch extending along a large portion of the radial head, and a laterally facing bicipital tuberosity. Distal radioulnar joint fully diarthrodial, with an expanded and two-faceted semilunar articulation on the ulnar head, and a partially developed ulnar fovea. Ulnar styloid process with reduced girth and not articulating with the short pisiform. Triquetrum small and with a reduced articular surface for the ulnar styloid process. Hamate relatively long proximodistally, with a steep triquetrum facet, a relatively large head and a distally projecting hamulus. Capitate with a relatively small and oblong head and a divided facet for the second metacarpal on its radial side.

Differential diagnosis

The new taxon differs from pliopithecoids and dendropithecids in its lack of a humeral capitular tail, its hominoid-like proximal radial morphology, its expanded ulnar head with a two-faceted semilunar articulation, and its partially developed ulnar fovea. It further differs from these taxa and proconsulids in its more hominoid-like carpal morphology (including the lack of a pisiform facet for the styloid process, a capitate facet for the second metacarpal divided by a deep ligamentary notch, and a distally projecting hamulus in the hamate), and particularly from pliopithecoids in its overall larger muzzle, more horizontal nasals anteriorly, some details of the upper molars, and (at least compared with Epipliopithecus) the lack of an entepicondylar foramen in the humerus. It also differs from all of the above-mentioned taxa in its fused ectotympanic and postglenoid process, and from these taxa and hominids in its horizontal and anteriorly oriented carotid canal. Last, it differs from crown hominoids (hylobatids and hominids) in its incompletely ossified ectotympanic and in its more primitive dentition and forelimb morphology (particularly in the humeroulnar articulation).

Description, comparisons, and paleobiology

Dental morphology and diet

Although the lack of lower dentition precludes comparisons with some taxa, the upper cheek teeth of Pliobates (Fig. 1D and Table 1) generally resemble those of other small-bodied Miocene catarrhines in both occlusal morphology and proportions (figs. S1 and S2). In contrast, they display a more primitive morphology than those of extant hominoids, including the similarly sized gibbons. Hylobatids possess more elongated cheek teeth with more peripheralized cusps, less developed cingula, and a much more extensive central fovea. Compared with Miocene small-bodied catarrhines from Eurasia and Africa (fig. S1), the upper molars of Pliobates more closely resemble those of the dendropithecid Micropithecus (21, 34, 35, 79) in several features, such as the markedly convex lingual profiles and moderately developed buccal cingula (albeit to a lesser extent than in Micropithecus), the C-shaped lingual cingulum that is mostly restricted to the protocone (not surrounding the hypocone), the well-developed and lingually situated hypocone, and the relatively narrow M1 and M2. Nevertheless, Pliobates differs in several features from Micropithecus, which has more restricted buccal cingula, a hypocone-metacone crest, and a relatively longer and less trapezoidal M3. The dentition of Pliobates more clearly differs from Epipliopithecus and other pliopithecoids (fig. S1), including from Barberapithecus [also recorded at the Vallès-Penedès Basin (36)] and Pliopithecus [previously recorded at ACM (33)] in several features, such as the more convex lingual profile, the more peripheral buccal cusps, the less developed cingula of the molars, the narrower M1 and M2, and the M3 occlusal morphology and proportions (fig. S2).

With regard to microwear features, the M1 displays a pitting percentage of 30.0%, a pit breadth of 5.67 μm, and a striation breadth of 1.98 μm. Based on pitting incidence (Fig. 2, A and B), which is the most useful metric for distinguishing among dietary categories (70), Pliobates closely resembles extant frugivores (Pan troglodytes) and eclectic feeders (Papio cynocephalus) that largely rely on ripe fruit. In contrast, the pitting incidence of Pliobates is higher than in extant folivores and much lower than in extant hard-object feeders (including orangutans). Compared with other extinct catarrhines from Western Europe, the pitting percentage of Pliobates is somewhat lower than in most pliopithecoids and hominoids, for which some degree of sclerocarpy has been inferred (66, 68). This low pitting incidence is consistent with the pit- and striation-breadth measurements (which show no sign of extreme folivory or specialized hard-object feeding) and most closely approaches that of the fossil hominoids Anoiapithecus brevirostris and Hispanopithecus laietanus, previously interpreted as soft frugivores (68). These results are confirmed by a multivariate analysis that simultaneously examined the three microwear variables (Fig. 2C and tables S9 to S11), in which Pliobates falls closer to the extant frugivorous–mixed-feeder centroid for the first and second canonical axes and is classified as a frugivore. Dental microwear analyses therefore indicate a mainly frugivorous diet for Pliobates, compatible with a high consumption of ripe fruit and a low sclerocarpic component.

Fig. 2 Results of the dental microwear analyses.

(A) Pitting incidence (%) of Pliobates, the extant comparative sample, and pliopithecoids and extinct hominoids from Europe and Turkey. (B) Bivariate plot of striation breadth versus pitting incidence. (C) Bivariate plot of the first two canonical axes delivered by the canonical variates analysis, based on three distinct, broad dietary groups: folivores, mixed feeders and frugivores, and hard-object feeders. Colored polygons in (B) and ellipses in (C) illustrate the variability of extant dietary categories. Small black symbols denote the comparative sample of extant anthropoids, whereas large black symbols represent the centroids of each dietary category. Different symbols are employed to distinguish the various extinct species; results for Iberian hominoids and pliopithecoids are shown in red and blue, respectively, whereas those from other localities are shown in yellow and green, respectively.

Body mass

Dental BM estimates for the female holotype of Pliobates (table S7) range from 2.9 to 4.8 kg, with an average BM estimate of 3.9 kg and an uncertainty degree (based on the combined 95% confidence intervals for each dental locus) of 2.5 to 5.7 kg. Postcranial BM estimates (table S8) are on average 4.8 kg (range: 4.0 to 5.6 kg), based on catarrhine regressions, and 4.3 kg (range: 2.6 to 6.3 kg), based on hominoid regressions (estimates for each postcranial estimator and their confidence intervals are given in table S8). Given that the size of Pliobates is in the lower range for extant hominoids, the catarrhine regressions probably yield more accurate estimates, although the hominoid-based estimates are closer to the dental ones. Overall, the body mass of the holotype of Pliobates cataloniae can be estimated at ~4 to 5 kg (much lower than that estimated for Epipliopithecus vindobonensis, ~11 to 12 kg; table S8).

Cranial morphology and encephalization

A 3D virtual reconstruction of the cranium, based on the preserved specimens, is shown in fig. S3, whereas the final reconstruction (including mirrored portions) is shown in Fig. 3 and movie S1. Based on this reconstruction, the cranium of Pliobates differs from the primitive catarrhine condition (22, 37, 60) by being short, wide, and high. However, the tubular ectotympanic is short and incompletely ossified—i.e., less developed than in Saadanius and extant crown catarrhines (2022). The maxillary sinus is extensive, as in stem catarrhines and hominoids (22, 80), and there is also a small frontal sinus, as in stem hominoids but unlike in stem catarrhines, cercopithecoids, hylobatids, and pongines (22, 37, 80). The face is short and displays anteriorly situated orbits, as in hylobatids, colobines, and some extinct small-bodied catarrhines such as Epipliopithecus, Micropithecus, and Lomorupithecus (19, 21, 38, 76, 79). However, Pliobates differs from these taxa (and more closely resembles hylobatids) by displaying a more well-defined muzzle (especially compared with Epipliopithecus) with long and more horizontal nasals, a higher zygomatic root (moderately high as in hylobatids, but less so than in hominids), an interorbital pillar nearly orthogonal to the frontal squama (as in hylobatids and chimpanzees), a high degree of orbital convergence and frontation (as in all extant hominoids), and thin and anteriorly projecting (telescopic) orbital rims [to a greater extent than in Epipliopithecus (38), and thus most closely resembling hylobatids and, as far as it can be ascertained with incomplete preservation, Micropithecus (79)]. Pliobates also displays derived hominoid features in the basicranium (Fig. 4A), including the absence of a postglenoid foramen with a large and ventrally visible jugular foramen (as in all extant hominoids), the foramen ovale situated anteriorly and laterally to the Eustachian aperture (as in hylobatids and African apes), the fusion between the auditory meatus and the postglenoid process (as in hylobatids and African apes), and the horizontal and anteriorly directed carotid canal in the petrosal bone (as in hylobatids).

Fig. 3 Cranial reconstruction.

Virtual reconstruction of the holotype (IPS58443) cranium of Pliobates cataloniae gen. et sp. nov., including mirrored fragments, in frontal (A), lateral (B), posterior (C), basal (D), and superior (E) views. Further details are given in fig. S3 and the methods in the text.

Fig. 4 Basicranial morphology.

(A) Drawing of the left basicranium of the holotype (IPS58443) of Pliobates cataloniae gen. et sp. nov., as preserved in ventral view. The jugular foramen appears artifactually larger because of the displacement of the temporal and occipital portions along the occipitotemporal suture (corrected in the reconstruction in Fig. 3). The course of the carotid canal is shown with a dashed line, based on CT images. AE, articular eminence; CAF, carotid foramen; COF, condylar fossa; EA, Eustachian aperture; EAM, external auditory meatus; EP, Eustachian process; ET, ectotympanic; FM, foramen magnum; FO, foramen ovale; GF, glenoid fossa; JF, jugular foramen; OC, occipital condyle; OTS, occipitotemporal suture; PGP, postglenoid process. (B to D) Drawings of comparable views (not to scale) in Hylobates sp. (B), Proconsul heseloni KNM RU 2036 [(C), reversed], and Victoriapithecus macinnesi KNM MB 29100a (D) (KNM, Kenyon National Museums; RU, Rusinga; MB, Maboko). Arrows denote the V-shaped, incompletely ossified ventral terminal tip of the tubular ectotympanic in the extinct taxa. [Artwork by M. Palmero]

Braincase measurements yield an average cranial capacity estimate of 69.0 cm3 (range: 41.4 to 110.7 cm3; table S12), which is close to the estimates of 60.1 and 65.3 cm3 delivered by the two most reliable estimators (57) and only slightly lower than the estimate of 75.1 cm3 obtained from foramen magnum area (table S12). According to our estimates of body mass (4.5 kg) and cranial capacity [72 cm3 (average of cranial and foramen magnum estimates)], Pliobates would display a monkey-like degree of encephalization extensively overlapping with extant cercopithecoids (fig. S4 and table S13), being much more encephalized than the stem catarrhine Aegyptopithecus, slightly more so than the stem cercopithecoid Victoriapithecus, and only slightly less so than hylobatids and the extinct hominoids Proconsul and Oreopithecus. All these taxa, like cercopithecoids, are less encephalized than the extinct hominoid Hispanopithecus (Rudapithecus) and the extant great apes. Although humans are outliers in brain size–body size allometric regressions, great apes further display an allometric grade shift compared with hylobatids (and Pliobates), which are only slightly more encephalized on average than cercopithecoids (58).

Postcranial morphology and locomotion

The humerus (Fig. 5) resembles that of extant crown catarrhines, proconsulids, and dendropithecids by lacking (unlike Epipliopithecus) an entepicondylar foramen (17, 20, 21, 38, 79, 81, 82). Pliobates more closely resembles extant hominoids in the laterally facing bicipital tuberosity in the radius (81, 83) (Fig. 5), as well as in the configuration of the humeroradial articulation (81, 82, 84) (Fig. 6), including: in the humerus, the lack of capitular tail [present in Epipliopithecus, dendropithecids, and cercopithecoids (84)] and the moderately globulous (although not posterolaterally expanded) capitulum with a well-developed zona conoidea [lacking in Epipliopithecus and dendropithecids (8184)]; and, in the radius, the only slightly tilted and almost circular radial head with a small and flat area, a reduced lateral lip, and a beveled surface for the humeral zona conoidea. Pliobates also has a hominoid-like diarthrodial distal radioulnar joint (8587), with a two-faceted expanded semilunar articulation in the ulnar head (Fig. 6). In this regard, Pliobates departs from Epipliopithecus, dendropithecids, and cercopithecoids (17, 38, 82) and more closely resembles Proconsul (88), although the ulnar head is less extensive than in extant hominoids. In contrast with these derived features, the humeral shaft and humeroulnar joint are plesiomorphic: The former (Fig. 5) is anteriorly straight and somewhat proximally retroflexed; the latter (Fig. 6) lacks the stabilizing features of extant hominoids (83, 89), as shown by the narrow ulnar trochlear notch without a median keel (in agreement with the absence of spooling and the poorly defined trochlear lateral keel in the humerus of Pliobates).

Fig. 5 Forelimb long bones.

Shown are the humerus, radius, and ulna of the holotype (IPS58443) of Pliobates cataloniae gen. et sp. nov. (A to E) Partial left humerus in medial (A), posterior (B), lateral (C), anterior (D), and distal (E) views. (F to K) Left radius in medial (F), posterior (G), lateral (H), anterior (I), proximal (J), and distal (K) views. (L to O) Proximal half of the left ulna in medial (L), posterior (M), lateral (N), and anterior (O) views. (P to T) Distal fragment of the left ulna in medial (P), posterior (Q), lateral (R), anterior (S), and distal (T) views.

Fig. 6 Elbow and wrist morphology.

The most diagnostic features of the elbow and wrist joints of Pliobates cataloniae gen. et sp. nov. (IPS58443), denoted by arrows in drawings of the distal humerus, proximal radius, and distal ulna, are shown with those of selected extant and extinct anthropoids for comparison. (A to D) Anterior (top) and distal (bottom) views of the distal humerus in P. cataloniae (A), Epipliopithecus vindobonensis Individual I [(B), reversed], Dendropithecus? sp. KNM MO 17022A (C) (MO, Moruorot), and Hylobates moloch (D). (E to H) Views perpendicular to the radial tuberosity (top) and proximal view (bottom) of the proximal radius in P. cataloniae (E), E. vindobonensis Individual I (F), Simiolus enjiessi KNM MO 63 [(G), reversed)], and H. moloch (H). (I to M) Medial (top) and distal (bottom) views of the distal ulna in P. cataloniae (I), E. vindobonensis Individual I (J), H. moloch (K), Ateles paniscus (L), and Cercopithecus aethiops (M). 1, absence of entepicondylar foramen; 2, absence of capitular tail; 3, lack of spool-shaped trochlea; 4, well-developed beveled surface for the zona conoidea; 5, small and flat area in the radial head; 6, ulnar fovea; 7, two-faceted, expanded semilunar articular surface in the ulnar head. Specimens are shown as if from the left side and are not to scale. [Artwork by M. Palmero]

Humeral torsion in Pliobates is estimated at 101°, irrespective of the method employed (based on the posterior buttress for the humeral head or the bisector of the bicipital groove), with a confidence interval spanning 95.7° to 106.3° [based on the prediction error (5.23%) for the bicipital groove method (50)]. This degree of torsion is moderate, higher than that estimated for Proconsul heseloni (92°), but comparable to estimates for Dryopithecus fontani (102°) and Dendropithecus macinnesi (103.5°), and only slightly below the value estimated for Epipliopithecus vindobonensis (109°). The humeral torsion of Pliobates is thus most comparable to that of non-atelid platyrrhines and lower than that of Ateles and extant hominoids, especially African great apes and humans [although the high degree of humeral torsion of extant hominoids is related to increased mobility at the glenohumeral joint, the higher values of great apes and humans appear related to knuckle-walking and enhanced manipulation, respectively, rather than suspensory behaviors (50)]. In contrast to the moderate humeral torsion, the forelimb of Pliobates appears somewhat elongated relative to its body size (fig. S5). Allometric computations of relative forelimb length in fossils (residuals are given in table S14) must be considered with caution, because they are dependent on the accuracy of body-size estimates. However, the forelimb of Pliobates (based on our BM estimate of 4.5 kg) appears more elongated than that of Epipliopithecus (based on our estimate of 11.5 kg). The latter taxon, contrary to previous assertions (81, 83), has the generalized proportions of quadrupedal monkeys. Pliobates, in contrast, has a forelimb elongation similar to that of female orangutans and Brachyteles, although it is less extreme than in Ateles and especially than in hylobatids. The same pattern holds when the humerus and radius are analyzed separately, although in Pliobates, relative length is somewhat higher for the radius than for the humerus. Pliobates further displays a high arm angle (8°), which is considerably greater than the average in most anthropoids, except Hylobates (9.8°), Pongo (6.3°), and Ateles (6.5°) (51).

The ulnocarpal articulation of Pliobates is completely different from that of Epipliopithecus and dendropithecids (17, 38, 90), including a partially developed ulnar fovea (Fig. 6), which, in extant hominoids, is the attachment area of the triangular disc ligament and the intra-articular meniscus (8587). The ulnar styloid process is relatively long and slender, with no discernible articular surfaces for the pisiform or triquetrum. This agrees with the lack of an articular facet for the styloid process on the pisiform, like extant hominoids but unlike monkeys and Proconsul (91). However, in contrast to Pierolapithecus (92), the triquetrum of Pliobates shows a proximal articular facet, which is more developed than that present in hylobatids and sometimes Pan (51, 87) but less developed than in monkeys. This suggests that ulnotriquetral contact might have been reduced by some kind of intra-articular tissue, similarly to some Ateles species (93). Moreover, as in apes, the triquetrum of Pliobates is small relative to hamate size (fig. S6), indicating a reduced loading on the ulnar side of the wrist. However, as in monkeys and Proconsul, the triquetrum of Pliobates differs from that of extant apes and Pierolapithecus (92) by possessing a proximally protruding beak-like process (Fig. 7). The hamate of Pliobates is “Miocene ape–like” (91, 92), although it more closely resembles that of hylobatids by possessing a dorsopalmarly narrow and proximodistally long triquetral articular surface that is proximally globular, as well as a distally projecting hamulus (Fig. 7). Pliobates further resembles hylobatids and Ateles by having an oblong and mediolaterally narrow capitate head that, like the facet for the hamate, is proximodistally aligned. This morphology contrasts with the more globulous, wider, and ulnarly inclined capitate head of other catarrhines, including Proconsul (88, 91) and Pierolapithecus (92); in Pliobates, though, it is not radially inclined, as it is in hylobatids. Moreover, the capitate facet for the second metacarpal is divided by a deep ligamentary notch (Fig. 7), as in extant hominoids and Pierolapithecus (92) but not in other catarrhines (including Proconsul), in which the facet for the second metacarpal is dorsopalmarly continuous and occupies the whole lateral aspect of the capitate (Fig. 7) (91). Pliobates has a complex articulation between the third metacarpal and capitate, as in extant apes but not in Proconsul (91); however, as in Pierolapithecus (92), the capitate of Pliobates lacks a hook-like process.

Fig. 7 Carpal bones.

Line drawings of carpal bones in Pliobates cataloniae gen. et sp. nov. (IPS58443) are shown with those of selected anthropoid genera for comparison. (A to E) Left capitate, in radial (top) and proximal (bottom) views, of Cercopithecus aethiops (A), Ateles paniscus (B), Pierolapithecus catalaunicus (C), Hylobates lar (D), and P. cataloniae (E); gray shading denotes articular areas for the second metacarpals, and cross-hatching denotes those for the third metacarpal. (F to J) Left hamate, in radial (top) and ulnar (bottom) views, of C. aethiops (F), A. paniscus (G), Pi. catalaunicus (H), H. lar (I), and P. cataloniae (J). (K to O) Left triquetrum, in proximomedial (top) and distal (bottom) views, of C. aethiops (K), A. paniscus (L), Pi. catalaunicus (M), H. lar (N), and P. cataloniae (O). Drawings are not to scale.

Regarding positional behaviors, although the primitive morphology of the proximal humerus of Pliobates is suggestive of generalized above-branch quadrupedalism (94), its overall postcranial body plan is more compatible with a locomotor repertoire that includes a large amount of cautious and eclectic climbing (87, 95). This inference is supported by the emphasis on pronation and supination capabilities, the reduced compressive forces transferred across the ulnar side of the wrist, and the important ulnar deviation and rotatory capabilities. It agrees with previous hypotheses on the original locomotor adaptations of hominoids (95, 96) and with recent interpretations of Proconsul that similarly depict this taxon as an arboreal quadruped with adaptations for cautious climbing and clambering (12, 88), including an incipient distal radioulnar diarthrosis that (unlike in Pliobates) is still associated to a nonretreated ulnar styloid process (88). The reduced ulnocarpal articulation of Pliobates thus more closely foreshadows the condition of extant hominoids, although to a lesser extent than in the stem great ape Pierolapithecus (92), indicating a decreased emphasis on forearm use under weight-bearing conditions relative to Proconsul (88). Several characteristics of Pliobates (particularly the elongated forearm, the high arm angle, and the laterally facing bicipital tuberosity) further suggest some degree of below-branch forelimb-dominated suspensory behaviors (51). However, the lack of hominoid-like elbow-stabilizing features in the humeroulnar joint (83, 89), the generalized metacarpophalangeal proportions, and the lack of marked phalangeal curvature suggest that Pliobates was not specifically adapted to perform the acrobatic suspensory behaviors (ricochetal brachiation) displayed by extant gibbons.

Phylogeny and evolutionary implications

Our cladistic analysis, based on both craniodental and postcranial characters, recovers a single most-parsimonious tree (Fig. 8 and table S15), indicating that Pliobates is more closely related to crown hominoids than other Miocene small-bodied catarrhines and Proconsul are. With moderate support (bootstrap 71%, Bremer index 3), our results contradict the view of some authors that all of these taxa are stem catarrhines (preceding the divergence between hominoids and Old World monkeys) (17, 21) and concur instead with some previous cladistic analyses indicating a hominoid status for both Proconsul (2, 8, 22) and, at least, dendropithecids (8, 22). Our analysis is inconsistent with the current consensus that Epipliopithecus is a pliopithecoid (20, 66, 76, 78); however, the internal phylogeny of pliopithecoids and dendropithecids is not settled by our results (it is recovered by the most-parsimonious tree but not by the bootstrap 50%-majority-rule consensus). Similarly, the phylogenetic relationships between the analyzed dryopithecines (Hispanopithecus and Pierolapithecus) are not well resolved (the closer link between Pierolapithecus and crown hominids has a Bremer index of 1 and bootstrap support of 56%). In contrast, our analysis recovers, with very high support (bootstrap 93 to 100%, Bremer indices 7 to 11), both the molecular phylogeny of extant hominoids (1, 6) and the stem hominid status of Pierolapithecus and Hispanopithecus (9, 92). The position of Pliobates as a stem hominoid more derived than Proconsul is relatively well supported (bootstrap 78%, Bremer index 2). When the analysis is repeated excluding all fossil taxa with large amounts of missing data (fig. S7), the position of Pliobates as a stem hominoid more derived than Proconsul is much better supported (bootstrap 100%, Bremer index 20); it is even more robust than the monophyly of crown hominoids (bootstrap 98%, Bremer index 11) and than the great-ape status of Pierolapithecus and Hispanopithecus (bootstrap 100%, Bremer index 13).

Fig. 8 Results of the cladistic analysis.

Single most-parsimonious tree of 645 steps, based on a taxon-character data matrix of 319 characters and 20 taxa (tables S5 and S6). Consistency index = 0.5912 (excluding uninformative characters); retention index = 0.6897; Rescaled consistency index = 0.4213. Numbers below nodes are Bremer indices, and numbers above nodes are bootstrap support percentages (only shown when ≥50%). Node numbers refer to clades in the list of apomorphies in table S15.

Given that our analyses support Pliobates as a stem hominoid more derived than Proconsul, the mosaic of primitive and derived features displayed by the former taxon is of utmost relevance for interpreting the evolution of several key features among catarrhine primates. Many authors agree that homoplasy has played an important role in hominoid postcranial evolution (5, 9, 12, 92), but Pliobates shows a mosaic of primitive and derived features both in the cranium and the postcranium. For the cranium, this is best illustrated by the short tubular (but incompletely ossified) external auditory meatus with a V-shaped end, which would imply the independent acquisition of a more completely ossified ectotympanic in cercopithecoids and hominoids. This has previously been posited by some authors (20), and it is suggested to some extent by the stem catarrhine Saadanius (22), the stem Old World monkey Victoriapithecus (60), and the stem ape Proconsul (97); in these taxa, the ectotympanic, albeit slightly more developed, is still short, does not laterally exceed the postglenoid process, and lacks a completely closed terminal ventral tip (Fig. 4). In contrast, several other cranial features of Pliobates are derived toward the crown-hominoid condition, generally more closely resembling hylobatids than hominids. Some of the similarities with gibbons (e.g., short face with a distinct muzzle and anteriorly situated telescopic orbits) may be size-related to a large extent (98), but others (horizontal and anteriorly directed carotid canal) are otherwise only known in hylobatids. Coupled with the fact that Pliobates chronologically fits within the long ghost lineage of hylobatids, from their divergence from hominids at ~17 Ma (1) until their putative oldest record (Yuanmoupithecus) at 8 to 7 Ma (10, 39), these similarities raise the possibility that Pliobates might be a stem hylobatid. This hypothesis is not favored by our total (craniodental plus postcranial) evidence–based cladistic analyses, which support instead a stem hominoid status for Pliobates, mostly because of the lack of various crown-hominoid postcranial synapomorphies. However, its hylobatid cranial features and small body size suggest that, at least in some respects, the last common ancestor of crown hominoids might have been more gibbon-like (or less great ape–like) than generally assumed (6, 16).

Furthermore, Pliobates supports the view that some small-bodied catarrhines played a more important role in the emergence of crown hominoids than has generally been assumed over the past decades. The postcranial evidence provided by Miocene great apes such as Pierolapithecus and Sivapithecus (9, 92, 99) indicates that the last common ancestor of crown hominoids must have been postcranially more primitive than it would be inferred to be exclusively on the basis of extant forms. This supports some degree of parallel evolution in the postcranium between extant hylobatids and hominids (5), although not to such a great extent as if Pliobates was interpreted as a crown hominoid. As a stem ape, Pliobates cannot resolve whether many of the postcranial derived features shared by extant hylobatids and hominids are homologous (6) or homoplastic (5), although Pliobates does suggest that a suite of features in the humeroradial and wrist joints might be homologous. In these anatomical regions, Pliobates displays much more extensive postcranial synapomorphies of crown hominoids than those convergently displayed by atelids, including a diarthrodial radioulnar joint, an expanded ulnar head, an incipient ulnar fovea, a long and thin styloid process with reduced contact with the relatively small triquetrum, and a distally projecting hamulus on the hamate. Pliobates also displays incipient suspensory adaptations, although, based on currently available evidence, it is not possible to conclusively ascertain whether these were inherited by the last common ancestor of crown hominoids (and later secondarily lost in some fossil great apes such as Sivapithecus and Pierolapithecus) or whether they merely represent an independent acquisition of Pliobates’s.


Pliobates provides the first evidence of crown-hominoid postcranial synapomorphies in a Miocene small-bodied catarrhine, thus demonstrating a greater diversity in postcranial morphology and positional behaviors than previously recognized among this paraphyletic assemblage of taxa. Three decades ago, the degree of parallel evolution required to evolve hylobatids from small-bodied catarrhines such as dendropithecids, albeit conceivable, was considered unlikely in light of the available evidence (18), because of the numerous parallelisms that would be required in crown-hominoid (and even crown-catarrhine) features between hylobatids and hominids. Although the evidence provided by Pliobates reduces this morphological gap, this taxon still falls short of being supported as a hylobatid by our cladistic analyses, which strongly favor instead the monophyly of extant hominoids with all fossil small-bodied catarrhines excluded. However, unlike dendropithecids, which are currently interpreted as stem catarrhines (21) or hominoids more basal than Proconsul (22), Pliobates is unambiguously interpreted as more closely related to crown hominoids. Given its chronology and geographic location, as well as the retention of plesiomorphic dental and some postcranial features that resemble those of small-bodied catarrhines such as dendropithecids, Pliobates is likely to be a late-surviving offshoot of a small African stem hominoid more closely related to crown hominoids than Proconsul is. This has important implications for reconstructing the ancestral morphotype from which extant hominoids evolved: It suggests that some small-bodied catarrhines could have played a much more remarkable role in ape evolution than previously thought, and that the last common ancestor of crown hominoids was not necessarily great ape–like.

Supplementary Materials

Figs. S1 to S7

Tables S1 to S15

Movie S1


  1. In the systematic scheme used here (9), hominoids are defined as including both the crown group (extant hominids and hylobatids, plus extinct taxa closely related to either) and the stem lineage (extinct taxa more closely related to crown hominoids than to Old World monkeys, but preceding the hylobatid-hominid divergence).
  2. The genus Proconsul was recently split into two distinct genera (100), but throughout the paper, this genus name is used in a broad sense (including both Proconsul s.s. and Ekembo).
  3. The new names published here are nomenclaturally available according to the requirements of the amended International Code of Zoological Nomenclature, including registration of the work in ZooBank ( with the following Life Science Identifier:
  4. ACKNOWLEDGMENTS: This work has been supported by the Spanish Ministerio de Economía y Competitividad (projects CGL2014-54373-P, CGL2011-27343, and CGL0211-28681; contracts RYC-2009-04533 to D.M.A. and JCI-2011-11697 to D.DM.; and grant BES-2009-020612 to M.P.R.), the Spanish Ministerio de Educación (grant AP2010-4579 to M.P.), the Generalitat de Catalunya (research groups 2014 SGR 416 GRC and 2014/100609 grant 2011 BE-DGR 00310 to D.M.A.), NSF (grant 1316947 to S.A.), and the AAPA Professional Development Grant (to S.A.). We further acknowledge the support of the Servei d’Arqueologia i Paleontologia of the Generalitat de Catalunya. Fieldwork was defrayed by CESPA Gestión de Residuos. The authors thank S. Larson for the estimation of humeral torsion; J. Arias-Martorell for discussion on the fossil material; M. Palmero for scientific artwork; E. Delson, D. Pilbeam, L. Costeur, U. Gölich, and E. Mbua for access to casts and original fossil specimens; I. Sucarrats for assistance in data gathering; E. Sarmiento and T. Kivell for sharing measurements; the staff of the Preparation Division of the ICP for the preparation of the specimens; J. Thostenson and M. Hill for assistance with using the Microscopy and Imaging Facility of the American Museum of Natural History; S. Llàcer for assistance in processing the movie; S. Grau for philological advice; M. Köhler for comments on the manuscript; and three anonymous reviewers for helpful and constructive suggestions. We are also grateful to the staff of the following museums for access to comparative specimens: American Museum of Natural History, New York; Museum of Comparative Zoology, Harvard; Kenyan National Museums; Muséum National d’Histoire Naturelle, Paris; Naturhistorisches Museum Wien; and Naturhistorisches Museum Basel. Author contributions were as follows: D.M.A. and S.M.-S. designed the study and wrote the paper with input from other authors; D.M.A., S.A., M.P.R., M.P., and S.M.-S. performed the descriptions and comparisons; J.F., M.P.R., and S.M.-S. made the virtual cranial reconstruction; D.DM. carried out the microwear analyses; J.M.R. directed the fieldwork and performed the stratigraphic correlations; D.M.A., S.A., M.P.R., M.P., and S.M.-S. assembled the cladistic data matrix; and D.M.A. and J.F. performed the phylogenetic analyses. All authors approved the final manuscript and declare that all data reported here are fully and freely available from the date of publication. The new fossils described in this paper are curated at the ICP with catalog number IPS58443. Supporting data for this paper are presented in the supplementary materials.
View Abstract

Stay Connected to Science

Navigate This Article