Morphometric data collection

Fossil measurements were taken using a digital calliper and recorded to the nearest 0.1 mm. CN specimens were tomographically scanned using a Siemens Somatom at the Hospital de Sant Pau I Santa Tecla (Tarragona, Spain). Measurements on 3D digital models were obtained using MeshLab [25]. Statistical analyses were performed in PAST v.5.0.2 [26].

Anatomical and metric standards

The anatomical terminology and linear measurements follow Cherin et al. [27] and Sorbelli et al. [28], with minor modifications (see Fig 4). Comparative data were either collected directly or sourced from the literature (full references in figures and table captions).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Osteology and in-life reconstruction of Parabos tigneresi based on the specimens from Camp dels Ninots.

Cranium (A–C), dentition (D–I), skeleton (J) and life restoration of an adult male with colours based on the extant bongo (Tragelaphus euryceros) (K), with the position of the most important anatomical elements mentioned in the text. Drawings not to scale. Legend: AK, anterior keel; Ba, basioccipital; c, canine; FoMg, foramen magnum; Fr, frontals; FrFm, frontal foramina; HCo, horncore; He, hemimandible; i1, first incisor; i2, second incisor; i3, third incisor; JuP, jugal process; Lc, lacrimal; M1, first upper molar; m1, first lower molar; M2, second upper molar; m2, second lower molar; M3, third upper molar; m3, third lower molar; Ma, mastoid; Mx, maxilla; Na, nasals; OcS, occipital squama; Ot, otion; P2, second upper premolar; p2, second lower premolar; P3, third upper premolar; p3, third lower premolar; P4, fourth upper premolar; p4, fourth lower premolar; Pa, parietal; PaCr, parietal crest; PLK, posterolateral keel; Pm, premaxilla; PMK, posteromedial keel; Soc, supraoccipital; Sq, squamosal; TeFo, temporal fossa; Zy, zygomatic. Artwork by LS.

https://doi.org/10.1371/journal.pone.0340256.g004

Morphometric analyses

The bovid from Camp dels Ninots was compared to several extant and fossil Bovidae from Eurasia and Africa by means of the following elements:

1. Horncore size and proportion: assessed by plotting the anteroposterior diameter against the transverse diameters, both taken at the base (above the pedicle) and with the boxplot of the transverse-to-anteroposterior diameters ratio *100.
2. Lower dental row size and proportion: evaluated using molar length versus the premolar length biplot, both measured at the base and the boxplot of the ratio molar/premolar lengths *100.
3. Supraoccipital size and proportion: assessed by plotting the transverse diameter versus the anteroposterior diameter, both taken at the narrowest points, and by means of boxplot of the ratio between the two measures *100.
4. Basioccipital tuberosities size and proportion: analysed via a bivariate scatter plot comparing the posterior tuberosities transverse diameter (=width) against the anterior tuberosity width and with the boxplot of these two *100.
5. Teeth (P4, M1, M3, p3, p4, m3) size and proportion: Analysed by plotting tooth length against width, both taken at the base.
6. Limb bones proportions: examined using log-transformed stylopodium length (humerus and femur) versus autopodium (radius/tibia) and metapodial (metacarpal and metatarsal) lengths. Data points and scatter clouds were generated based on values from individual specimens, with each point representing the two corresponding bone lengths of a single limb. This approach was not applicable to Parabos cordieri, Alephis lyrix, Leptobos etruscus, and Bison spp., as it was not possible to assign specific bones to individual specimens. In these cases, the points are represented by the minimum, maximum and mean values for each element.

Hypsodonty index

The hypsodonty degree was calculated following Fortelius et al. [29] in which the height of the tooth taken at its highest point is divided by its basal length. The sample includes only unworn or early wear stage teeth.

Body mass estimation

Body mass was estimated using equations from Scott [30] and Janis [31], based on average extant species masses and 11 dental and postcranial variables (S31 Table). Body mass was estimated for nine of the most complete CN individuals and for other taxa from direct observation or using measurements from the literature (S32 Table). Each individual’s body mass was calculated as the average of the mean values obtained separately from all available variables, including both left and right elements where applicable.

Taxonomic considerations

The fossil relatives of extant boselaphines (i.e., Boselaphus and Tetracerus) are largely unknown and most, if not the totality, of the Miocene boselaphine-like taxa such as Miotragocerus, Tragoportax and allies (e.g., Austroportax, Strepsiportax, Protragocerus, Helicoportax, among others), which have been historically referred to Boselaphini, might be part of a distinct clade. Already in the 50s several authors proposed to unify these forms under the name of Tragocerina/Tragocerini, which, however, cannot be considered valid due to nomenclatural issues [32, 33, 34). Bibi et al. [5] proposed the term Tragoportacini to define the Miotragocerus-Tragoportax complex. This nomenclature had received alternate use in literature being formally accepted [35–37] or not [38–40]. Although, at the current state of the art, the clade is not well defined and lacks a proper diagnosis we do recognize that the Miotragocerus-Tragoportax complex differs substantially from Boselaphini enough to be considered a separate group. Therefore, in this paper, we adopt the nomenclature proposed by Bibi et al. [5] due to its relatively wide use in literature and to avoid informal and verbose definition such as Miotragocerus-Tragoportax complex. Other taxonomical nomenclature that requires specification are hereby listed: nomenclature referred to Bison and Leptobos follows Azzarà et al. [41] and Sorbelli et al. [28], nomenclature referred to Miotragocerus and Tragoportax follows Kostopoulos [22,42], nomenclature referred to Ugandax follows Gentry [43], nomenclature referred to Proamphibos follows Nishioka et al. [44].

Institutional abbreviations

IPHES, Institut Català de Paleoecologia Humana i Evolució Social, Tarragona (Spain); LGPUT, Museum of Geology, Palaeontology and Palaeoanthropology of the Aristotle University of Thessaloniki (Greece); AMNH, American Museum of Natural History, New York (USA); RIN, Rajabhat Institute of Nakhon Ratchasima, Nakhon Ratchasima (Thailand); UFR FSL, Faculté des Sciences, Université Claude Bernard Lyon 1, Lyon (France); MNHNP, Museum national d’Histoire naturele, Paris (France); NMBS, Naturhistorisches Museum Basel, Basel (Switzerland); IGF, Sezione di Paleontologia del Museo di Storia Naturale, Università di Firenze, Florence (Italy); IGPM, Institute für Geologie und Paläontologie Universität Münster, Münster (Germany); GSI, Geological Survey of India, Calcutta (India); MFN, Museum für Naturkunde, Berlin (Germany); ICP IPS, Institut Català de Paleontologia Miquel Crusafont, Sabadell (Spain).

Systematic palaeontology

Unranked hierarchy

1. Artiodactyla Owen 1848
2. Ruminantia Linnaeus 1758
3. Bovidae Gray, 1821
4. Bovinae Gray, 1821
5. Parabos Arambourg and Piveteau, 1929a [16]

Selected synonymy list

1. Antilope de Christol, 1832 [45]
2. Antilope Gervais, 1848–1852 [46]
3. in pars. Antilope Gervais, 1867-1869 [47]
4. in pars. Palaeoryx Depéret, 1885 [48]
5. Parabos nov. gen, Arambourg and Piveteau 1929 [16]
6. in pars. Parabos, Pilgrim, 1939 [18]
7. in pars. Parabos, Merla 1949 [49]
8. in pars. Parabos, Samson and Radulescu, 1963 [50]
9. in pars. Parabos, Gromolard, 1980 [15]
10. in pars. Parabos, Gromolard, 1981 [20]
11. in pars. Parabos, Gromolard and Guérin, 1980 [21]
12. in pars. Parabos, Morales, 1984 [51]
13. 1991: Parabos, Michaux et al., 1991 [52]
14. 2006: Parabos, Montoya et al., 2006 [53]
15. 2009: Parabos, Bibi, 2009 [24]
16. 2017: Parabos, Crégut-Bonnoure and Tsoukala 2017 [54]
17. 2022: Parabos, Kostopoulos, 2022 [22]

Diagnosis

Medium to large sized Bovinae diagnosed by the following combination of characters: horns inserted above the posterior edge of the orbits, well-separated from each other, emerging parallel or slightly diverging, straight or medially curved, with weak heteronymous torsion. Horns with very weak or no mediolateral compression and faint or no surface grooving. Basal cross-section triangular, with three keels, or rounded, without evident keels. Keels, when present, with various degrees of development but anterior keel always stronger than the other two. Anterior keel never stepped. Frontal sinuses extended into the pedicle and slightly into the horn core bases. Pedicle sinus not strutted. Neurocranium shortened with parietal crests elevated above the parietals, supraoccipital elongated and occipital squama relatively wide. Supraoccipital flexed toward the occipital so the angle created by the supraoccipital and the parietal surfaces is obtuse. Skull roof behind the horncores depressed with smooth or weakly sculpted surface. Supraorbital foramina located at the base of the pedicles, flush on the surface without a surrounding fossa. Ethmoidal fenestra and lachrymal fossa absent at least in one species but unknown in the others. Basioccipital triangular with posterior tuberosities almost double the width of the anterior tuberosities. Premaxillae in contact with the nasals. Teeth large and mesodont with wrinkled buccal surface and scarce or no cement. Presence of incipient basal pillars on molars. Premolar series length reduced relatively to the molars (premolar length ca. 60% of the molar length). Fourth lower premolar with prominent metaconid, expanded both lingually and mesiodistally, often in contact with the paraconid and closing the anterior lingual valley. Goat fold often presents in lower molars, especially in the third molar. Limb bones elongated but showing a slight reduction in the length of the metapodials relative to the stylopodium (emended from [15]).

Differential diagnosis

Parabos differs from Alephis in the straighter, less mediolaterally compressed and less twisted horncores, the more developed posterior keels of the horns so their base has a clearer triangular cross-section (this last character is not valid for Parabos savelisi), the larger supraoccipital and stronger parietal crests. It differs from crown Bovini in the same characters mentioned for Alephis as well as in a longer neurocranium with horncores positioned more frontward, supraorbital foramina placed more posteriorly, not located within elongated grooves, and less hypsodont dentition with smaller basal pillars. It differs from Miotragocerus, Tragoportax and allies (i.e., Tragoportacini sensu [5]) in larger size, horncores lacking strong mediolateral compression, weaker frontal keel lacking steps, shorter and wider neurocranium, smaller supraoccipital, weaker temporal crests, lack of sculpting on the parietals and posterior frontals, more reduced premolars relative to the molars, strengthening of the molar pillars and ribs. It differs from extant Boselaphini in the longer and more posteriorly oriented horns, lack of sculpting on the parietals and posterior frontals, shorter and taller neurocranium, and supraoccipital more strongly flexed against the skull roof.

Type species

Parabos cordieri (de Christol, 1832 [45]), as designated by Gromolard (1980 [15]).

Diagnosis

The following combination of characters distinguishes this species from all known species of Parabos and Alephis. Horns weakly diverging, weakly curving posteriorly and medially, with faint heteronymous torsion; more so than in any other Parabos species but less than in Alephis. Horn core basal cross-section sub-triangular, slightly compressed mediolaterally, with three keels evident. Prominent anteromedial keel and weaker posterolateral and posteromedial keels. Medial and posterior surfaces of the horn flattened. Medial surface covered by shallow grooves (absent in P. cordieri and P. savelisi and deeper in A. lyrix). Frontals and pedicles pneumatized with sinuses reaching the base of the horns. Braincase relatively short and high with relatively small occipital squama. Large supraoccipital, shorter than in any other Parabos species but longer than in Alephis. Ethmoidal fenestra absent. Dentition very large. Pillars and ribs relatively well-developed, similar to the condition shown in Alephis lyrix. Lower p4 as in P. cordieri and Tragoportacini in which the metaconid is inflated, almost reaching the paraconid and often closing the anterior valley. Cement present but thin. Size larger than any other Parabos species and comparable with large-sized extant Bovini (ca. 500 kg). Limb bones more massive than in Boselaphini, Tragoportacini and any other Parabos species but slenderer than crown Bovini (emended from [52]).

Type locality

Baho Section (Perpignan Basin, Roussillon, France), Pliocene.

Stratigraphic and Geographic distribution—Early Pliocene (MN14) of western Europe (France, Spain).

Referred specimens

The Camp dels Ninots bovid sample studied in this work includes 1,779 specimens for a total of a minimum of 14 individuals. The sample is figured in Figs 3; 5–16. Measurements are provided in Tables 1, 2, S2–S30. The complete list of the specimens divided by individuals is provided in S1 Table.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Skull of of Parabos tigneresi.

IPHES.CN’11-B4-1 in anterodorsal (A), left lateral (B), posterior (C), left lateral (D), and dorsolateral left (E) views and close-ups on the basioccipital (F), horncore bases and frontals (G) and temporal fossa (H). Scale bar: 100 mm.

https://doi.org/10.1371/journal.pone.0340256.g005

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Skulls of Parabos tigneresi.

A–B, IPHES.CN’19-B15-1 in right (A) and left lateral (B) views; C–D, skull IPHES.CN’17-B14-1 in right (C) and left lateral (D) views. Scale bar: 100 mm.

https://doi.org/10.1371/journal.pone.0340256.g006

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Skulls of Parabos tigneresi.

A, skull IPHES.CN’12-B10-1 in left lateral view; B, skull IPHES.CN’06-B3-1 in right lateral view; C–D, IPHES.CN’12-B7-1 in left (C) and right (D) lateral view. Scale bar: 100 mm.

https://doi.org/10.1371/journal.pone.0340256.g007

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. Horncores and neurocranium of Parabos tigneresi.

A–D, right horncore IPHES.CN’05-B2-4 in anterior (A), posterior (B), medial (C) and lateral (D) views; E–H, left horncore IPHES.CN’05-B2-5 in anterior (E), posterior (F), medial (G) and lateral (H) views; I–M, left horncore IPHES.CN’04-B1-4 in anterior (I), posterior (J), proximal (K), medial (L), and lateral (M) views; N–Q, CT images IPHES.CN’11-B4-1 with transverse cross-sections at the pedicles (N) and horncore bases (O), and sagittal cross-section through the right horncore (P) and close-up of the horncore sinus in the same view (Q). Scale bars: 50 mm.

https://doi.org/10.1371/journal.pone.0340256.g008

Download:

• PNG
  larger image
• TIFF
  original image

Fig 9. Upper dentition of Parabos tigneresi.

A–C, left maxilla with P2-M3 IPHES.CN’05-B2-1 in occlusal (A), labial (B) and lingual (C) views; D–F, right P2-M3 IPHES.CN’04-B1-14–17,19, in occlusal (D), labial (E) and lingual (F) views; G–I, left maxilla with P2-M3 IPHES.CN’04-B1-10 in in occlusal (G), lingual (H) and labial (I) views; J–L, left and right maxillae with left P3-M3, R P2-M2 CN’12-B9-1 in left (J), right lateral (K) and occlusal (L) views. Scale bar: 50 mm.

https://doi.org/10.1371/journal.pone.0340256.g009

Download:

• PNG
  larger image
• TIFF
  original image

Fig 10. Lower dentition of Parabos tigneresi.

A–C, right hemimandible with p3-m3, IPHES.CN’04-B1-8 in lingual (A), labial (B) and occlusal (C) views; D–F, left hemimandible with p3-m3, IPHES.CN’05-B2-3, in occlusal (D), lingual (E) and labial (F) views; G, right hemimandible with p2-m3, IPHES.CN’05-B2-2, in occlusal view; H, upper and lower left tooth rows of IPHES.CN’17-B14-1 in labial view; I, p2–p4 of IPHES.CN’11-B4-1 in occlusolabial view. Scale bars: 50 mm.

https://doi.org/10.1371/journal.pone.0340256.g010

Download:

• PNG
  larger image
• TIFF
  original image

Fig 11. Vertebrae and partial skeletons of Parabos tigneresi.

A–D, atlas IPHES.CN’04-B1-125 in dorsal (A), ventral (B), anterior (C) and posterior (D) views; axis IPHES.CN’04-B1-127 in anterior (E), lateral (F) and posterior (G) views; H, partial axial (rib cage, trunk vertebrae) and appendicular (scapula, humerus, radius) skeletal elements of CN’17-B14 in left lateral view; I, partial axial (skull, vertebral column, rib cage) and appendicular (scapula, humerus, radius and ulna, pelvis and femur) skeletal elements of CN’19-B15 in right lateral view. Scale bars: 50 mm (A–G), 100 mm (H–I).

https://doi.org/10.1371/journal.pone.0340256.g011

Download:

• PNG
  larger image
• TIFF
  original image

Fig 12. Humerus, radius and ulna of Parabos tigneresi.

A–D, left humerus IPHES.CN’04-B1-345 in anterior (A), posterior (B), medial (C) and lateral (D) views; E–H, right humerus IPHES.CN’11-B4-300 in anterior (E), posterior (F), medial (G) and lateral (H) views; I, left ulna IPHES.CN’05-B2-337 in lateral view; J–M, right radius IPHES.CN’04-B1-303 in proximal (J), distal (K), anterior (L) and posterior (M) views; N–Q, left radius IPHES.CN’11-B4-303 in proximal (N), distal (O), anterior (P) and posterior (Q) views. Scale bars: 50 mm.

https://doi.org/10.1371/journal.pone.0340256.g012

Download:

• PNG
  larger image
• TIFF
  original image

Fig 13. Carpal and metacarpal bones of Parabos tigneresi.

A–B, right pyramidal IPHES.CN’05-B2-326 in lateral (A) and medial (B) views; C–D, right semilunar IPHES.CN’05-B2-428 in proximal (C) and distal (D) views; E–F, right scaphoid IPHES.CN’05-B2-313 in lateral (E) and medial (F) views; G–H, right unciform IPHES.CN’05-B2-311 in proximal (G) and distal (H) views; I–J, right magnum-trapezoid IPHES.CN’05-B2-312 in proximal (I) and distal (J) views; K–L, right pisiform IPHES.CN’05-B2-325 in lateral (K) and medial (L) views; M–P, right metacarpal IPHES.CN’04-B1-344 in anterior (M), posterior (N), proximal (O) and distal (P) views; Q–T, right metacarpal IPHES.CN’11-B4-306 in anterior (Q), posterior (R), proximal (S) and distal (T) views; U–X, left metacarpal IPHES.CN’19-B15-432 in anterior (U), posterior (V), proximal (W) and distal (X) views; Y–AB, left metacarpal IPHES.CN’17-B14-316 in anterior (Y), posterior (Z), proximal (AA) and distal (AB) views. Scale bar: 50 mm.

https://doi.org/10.1371/journal.pone.0340256.g013

Download:

• PNG
  larger image
• TIFF
  original image

Fig 14. Femur and tibia of Parabos tigneresi.

A–D, left femur IPHES.CN’11-B4-401 in anterior (A), posterior (B), lateral (C) and medial (D) views; E–J, left tibia IPHES.CN’11-B4-403 in anterior (E), lateral (F), posterior (G), medial (H), proximal (I) and distal (J) views. Scale bar: 50 mm.

https://doi.org/10.1371/journal.pone.0340256.g014

Download:

• PNG
  larger image
• TIFF
  original image

Fig 15. Tarsal and metatarsal bones of Parabos tigneresi.

A–D, right astragalus IPHES.CN’05-B2-402 in anterior (A), posterior (B), medial (C) and lateral (D) views; E–F, right malleolus IPHES.CN’05-B2-409 in lateral (E) and medial (F) views; G–H right cuneiform IPHES.CN’05-B2-410 in proximal (G) and distal (H) views; I–K, right calcaneum IPHES.CN’19-B15-308 in medial (I), anterior (J) and lateral (K) views; L–M, left cubonavicular IPHES.CN’05-B2-400 in proximal (L) and distal (M) views; N–Q, right metatarsal IPHES.CN’17-B14-447 in proximal (N), distal (O), anterior (P) and posterior (Q) views; R–S, left metatarsal IPHES.CN’19-B15-300 in proximal (R), distal (S), anterior (T) and posterior (U) views; V–Y, left metatarsal IPHES.CN’11-B4-404 in proximal (V), distal (W), anterior (X) and posterior (Y) views. Scale bar: 50 mm.

https://doi.org/10.1371/journal.pone.0340256.g015

Download:

• PNG
  larger image
• TIFF
  original image

Fig 16. Phalanges of Parabos tigneresi.

A, C, F, H, right proximal phalanx of the right forelimb IPHES.CN’05-B2-320 in anterior (A), abaxial (B), posterior (C) and proximal (D) views; B, D, E, G, left proximal phalanx of the right forelimb IPHES.CN’05-B2-333 in anterior (B), abaxial (D), posterior (E) and proximal (G) views; I, K, N, P, right intermediate phalanx of the right forelimb IPHES.CN’05-B2-332 in anterior (I), abaxial (K), posterior (N) and proximal (P) views; J, L, M, O, left intermediate phalanx of the right forelimb IPHES.CN’05-B2-323 in anterior (J), abaxial (L), posterior (M) and proximal (O) views; Q, S, V, X, right distal phalanx of the right forelimb IPHES.CN’05-B2-322 in anterior (Q), abaxial (S), posterior (V) and proximal (X) views; R, T, U, W, left distal phalanx of the right forelimb IPHES.CN’05-B2-425 in anterior (R), abaxial (T), posterior (U) and proximal (W) views; Y, AA, AD, AF, right proximal phalanx of the right hindlimb IPHES.CN’05-B2-329 in anterior (Y), abaxial (AA), posterior (AD) and proximal (AF) views; Z, AB, AC, AE, left proximal phalanx of the right hindlimb IPHES.CN’05-B2-432 in anterior (Z), abaxial (AB), posterior (AC) and proximal (AE) views; AG, AI, AL, AN, right intermediate phalanx of the right hindlimb IPHES.CN’05-B2-435 in anterior (AG), abaxial (AI), posterior (AL) and proximal (AN) views; AH, AJ, AK, AM, left intermediate phalanx of the right hindlimb IPHES.CN’05-B2-429 in anterior (AG), abaxial (AI), posterior (AL) and proximal (AN) views; AO, AQ, AT, AV, right distal phalanx of the right hindlimb IPHES.CN’05-B2-426 in anterior (AO), abaxial (AQ), posterior (AT) and proximal (AV) views; AP, AR, AS, AU, left distal phalanx of the right hindlimb IPHES.CN’05-B2-427 in anterior (AP), abaxial (AR), posterior (AS) and proximal (AU) views. Scale bar: 50 mm.

https://doi.org/10.1371/journal.pone.0340256.g016

Description

Comparisons

The bovid from Camp dels Ninots belongs to the Bovinae subfamily given its quadrangular braincase (i.e., ‘box-shaped’ neurocranium), low facial-cranial flexion, keeled horn cores lacking transverse ridges, and simple cranial sutures [22,24]. Extant Bovinae is composed of the tribes Tragelaphini, Boselaphini and Bovini in addition to the saola (Pseudoryx nghtinhensis), which has been placed in its own tribe (Pseudorygini) or within Bovini [e.g., 60, 61]. Most Miocene Bovinae have traditionally been referred to the Boselaphini, even though workers have long suspected these to include the ancestors of Bovini (see [5] and references therein). Therefore, it has been proposed that species potentially on the line to Bovini be recognized as stem bovins, those on the line to Boselaphini as stem boselaphins, and numerous species of Tragoportax, Miotragocerus, and related boselaphine-like bovids treated as a separate tribe, Tragoportacini (e.g., [5], see Materials and Methods). Among Neogene Bovinae, the Camp dels Ninots bovid shows cranial and dental characters similar to both Tragoportacini and early (stem) Bovini from the Miocene and Pliocene of Eurasia and Africa. Consequently, our comparisons focus primarily on these groups and extant tribes of Bovidae (Figs 17–24). Due to the deformation of the cranial elements, complete metrics could not be obtained; thus, morphological and morphometric comparisons were largely confined to the most intact and undeformed elements, including horns, dentition, and occipital and basioccipital bones.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 17. Metric comparison of cranial and dental elements in Parabos tigneresi and other bovids from the Eurasian Neogene and Quaternary.

A, basal horn core transversal and anteroposterior diameters; B, lower molar and premolar row lengths. HAPD, anteroposterior diameter of the horncore; HTD, transverse diameter of the horncore; ML, lower molar row length; PML, lower premolar row length. Superscript denotes data source: a, this study; 1, [62]; 2, [18]; 3, [63]; 4, [20]; 5, [51]; 6, [57]; 7, [64]; 8, [65]; 9, [42]; 10, [66]; 11, [67]; 12, [68]; 13, [69]; 14, [70]; 15, [71]; 16, [72]; 17, [73]; 18, [38]; 19, [74]; 20, [75]; 21, [76]; 22, [44]; 23, [77]; 24, [39]; 25, [36]; 26, [28]. The full dataset is provided in S33 and S34 Tables.

https://doi.org/10.1371/journal.pone.0340256.g017

Download:

• PNG
  larger image
• TIFF
  original image

Fig 18. Metric comparison of cranial and dental elements in Parabos tigneresi and other bovids from the Eurasian Neogene and Quaternary.

A, basioccipital tuberosities transversal diameters; B, supraoccipital length and width; C, boxplot of the horncore base compression index: D, boxplot of the lower premolar index; E, boxplot of the supraoccipital compression index; F, boxplot of the basioccipital tuberosities index. Abbreviations: BATW, basioccipital width at the anterior tuberosities; BPTW, basioccipital width at the posterior tuberosities; SOL, supraoccipital length; SOW, supraoccipital width. Superscript denotes data source: a, this study; 1, [63]; 2, [78]; 3, [57]; 4, [64]; 5, [65]; 6, [42]; 7, [67]; 8, [74]; 9, [27]; 10, [44]; 11, [77]); 12, [39]; 13, [36]; 14, [28]. The full dataset is provided in S35 and S36 Tables.

https://doi.org/10.1371/journal.pone.0340256.g018

Download:

• PNG
  larger image
• TIFF
  original image

Fig 19. Metric comparison of dental elements in Parabos tigneresi and other bovids from the Eurasian Neogene and Quaternary.

A, length vs width of P4; B, length vs width of M1; C, length vs width of M3.Superscript denotes data source: a, this study; 1, [79]; 2, [18]; 3, [20]; 4, [57]; 5, [52]; 6, [80]; 7, [66]; 8, [67]; 9, [53]; 10, [81]; 11, [73]; 12, [82]; 13, [38]; 14, [74]; 15, [27]. The full dataset is provided in S37 Table.

https://doi.org/10.1371/journal.pone.0340256.g019

Download:

• PNG
  larger image
• TIFF
  original image

Fig 20. Metric comparison of dental elements in Parabos tigneresi and other bovids from the Eurasian Neogene and Quaternary.

A, length vs width of p3; B, length vs width of p4; C, length vs width of m3. Superscript denotes data source: a, this study; 1 [79]; 2, [20]; 3, [57]; 4, [52]; 5, [80]; 6, [66]; 7, [53]; 8, [81]; 9, [73]; 10, [82]; 11, [38]; 12, [74]. The full dataset is provided in S38 Table.

https://doi.org/10.1371/journal.pone.0340256.g020

Download:

• PNG
  larger image
• TIFF
  original image

Fig 21. Boxplot of the hypsodonty index in lower molars of Parabos tigneresi specimen IPHES.

CN’12-B7-1 and other bovids tribes. A, m1; B, m2; C, m3. Hypsodonty index is unworn crown height divided by crown length. Star and red line indicate values for IPHES. CN’12-B7-1. Coloured columns represent fossil Bovidae. Number of specimens shown above each sample. Data taken from [83] and references therein.

https://doi.org/10.1371/journal.pone.0340256.g021

Download:

• PNG
  larger image
• TIFF
  original image

Fig 22. Metric comparison of front and hind limb element length in Parabos tigneresi from Camp dels Ninots and other bovids from the Eurasian Neogene and Quaternary.

A, humerus vs radius lengths; B, humerus vs metacarpal lengths; C, femur vs tibia lengths; D, femur vs metatarsal lengths. Axes in log10. Superscript denotes data source: a, this study; 1, [84]; 2, [85]; 3, [86]; 4, [87]; 5, [88]; 6, [89]; 7, [90]; 8, [91]; 9, [56]; 10, [52]; 11, [92]; 12, [93]; 13, [94]; 14, [95]; 15, [96]; 16, [28].

https://doi.org/10.1371/journal.pone.0340256.g022

Download:

• PNG
  larger image
• TIFF
  original image

Fig 23. Crania of various Eurasian Miocene, Pliocene and Pleistocene bovids compared with Parabos tigneresi.

A–C, Miotragocerus macedoniensis (LGPUT DKO-53) from Ditiko (Greece) in dorsal (A) and posterodorsal (B) views, basioccipital (C); D–F, Tragoportax amalthea (LGPUT NIK-1688) from Nikiti (Greece) in dorsal (D) and posterodorsal (E) views, basioccipital (F); G–I, Selenoportax vexillarius (AMNH 19748, holotype and RIN 316) from Hasnot (Pakistan) and Tha Chang Sandpits (RIN 316, Thailand) in frontal (G, AMNH 19748) and posterodorsal (H, RIN 316) views, basioccipital (I, RIN 316); J–L, Parabos savelisi (LGPUT GAS9, holotype) from Gephyra (Greece) in dorsal (J) and posterodorsal (K) views, basioccipital (L); M–O, Parabos cordieri (UFR FSL 40 024) from Sables de Montpellier (France) in dorsal (M) and posterodorsal (N) views, basioccipital (O); P–R, Parabos tigneresi (IPHES.CN’17-B14-1) from Camp dels Ninots (Spain) in dorsal (P) and posterodorsal (Q) views, basioccipital (R); S–U, Alephis lyrix (MNHNP, Figured in in Depéret, 1890, pl. V, fig. 4 (8), holotype) from Serrat d’en Vaquer in the Perpignan/Roussillon basin (France) in dorsal (S) and posterodorsal (T) views, basioccipital (U); V–X Ugandax coryndonae (AL 194−1, holotype) from Hadar formation (Ethiopia) in dorsal (V) and posterodorsal (W) views, basioccipital (X); (Y–Y, Leptobos etruscus (NMBS Se792 and IGF 612, lectotype) from Senéze (France) and Upper Valdarno (Italy) in dorsal (Y, NMBS Se792) and posterodorsal (Z, IGF 612) views, basioccipital (AA, IGF612). Scale bar: 100 mm.

https://doi.org/10.1371/journal.pone.0340256.g023

Download:

• PNG
  larger image
• TIFF
  original image

Fig 24. Right lateral views of crania and basal horn core cross-sections of various bovids from the Eurasian Miocene, Pliocene and Pleistocene.

A, Miotragocerus valenciennesi (ICP IPS 5169) from Piera (Spain); B, Tragoportax amalthea (LGPUT NIK-1688) from Nikiti (Greece); C, “Parabos” athanasiui (No ID, mirrored) from Malusteni (Romania); D, Samokoerus minotaurus (IGPM PIM 99, holotype) from Samos (Greece); E, Parabos savelisi (LGPUT GAS 9, holotype, mirrored) from Gephyra (Greece); F, Parabos cordieri (UFR FSL 40 024) from Sables de Montpellier (France); G, Parabos tigneresi (IPHES.CN’17-B14-1) from Camp dels Ninots (Spain); H, Grevenobos antiquus (LGPUT MIL 401a, holotype) from Milia (Greece); I, Selenoportax falconeri (PRY 204) from central Myanmar; J, Alephis lyrix (MNHNP, Figured in in Depéret, 1890, pl. V, fig. 4 (8), holotype) from from Serrat d’en Vaquer in the Perpignan/Roussillon basin (France); K, Ugandax coryndonae (AL 194−1, holotype) from the Hadar formation (Ethiopia); L, Proamphibos lachrymans (GSI No. B. 561) from Jammu (India); M, Leptobos etruscus (IGF 612, lectotype) from Upper Valdarno (Italy); N, Boselaphus tragocamelus (MFN 108046) from South Asia. Scale bar: 100 mm.

https://doi.org/10.1371/journal.pone.0340256.g024

The horns of the CN bovid slightly bend posteromedially with a faint heteronymous torsion and are characterized by one prominent frontal keel and two weaker posterior keels and a triangular cross-section (Figs 23P; 24G). These features are shared with the Baho Alephis tigneresi, distinguishing both the Spanish and French samples from the younger Alephis lyrix, which, exhibits longer horn cores with more pronounced heteronymous torsion and medial curvature of the cores (Figs 23S; 24J; Table 3). In addition, A. lyrix differs from the CN bovid in having a wider divergence angle of the horns, deeper grooves on the medial surface, and an ovoidal cross-section with a less prominent posterolateral keel (Figs 23S; 24J). The horns of A. tigneresi and Parabos cordieri show strong similarities. Nevertheless, P. cordieri horns differ from CN and Baho samples in the following aspects: smaller absolute size, less divergent horn bases, weaker medial bending and torsion, more prominent lateral keel and absence of surface grooves (Figs 23M; 24F; Table 3). Parabos savelisi from Gephyra (Greece), differs from the bovid from CN in the same characters as P. cordieri (Figs 23J; 24E; Table 3). Additionally, P. savelisi is marked by a strong reduction of the three keels, resulting in a sub-ovoidal cross-section (Fig 24E, [54]), distinguished from the triangular cross-sections of P. cordieri and A. tigneresi. Due to the extremely limited material available for Parabos soriae, Parabos macedoniae, and Parabos athanasioui, detailed cranial comparisons are not possible. However, the horncores of both P. soriae and P. athanasioui are notably smaller than those of the CN bovid (Fig 17A). A commentary on the diagnostic traits and taxonomic status of these species is provided in the discussion (see next chapter). The horns of the studied sample differ from the typical Tragoportacini morphology which presents, apart from the overall smaller size, a stronger mediolateral compression and, often (as it is the case of Miotragocerus), only two well-defined keels (Figs 17A; 18C; 23A, D; 24A, B). In addition, some Tragoportacini show a stepped anterior margin on the anterior keel which is completely absent in the sample from CN. The Late Miocene Samokeros minotaurus from Greece and Iran differs from CN specimens in smaller horns, with oval basal cross-section, mediolaterally flattened distal cross-section, stronger medial curvature, and lacking keels (Figs. 17A; 24D). The bovins Grevenobos antiquus (Late Pliocene, Balkans) and Leptobos spp. (Plio-Pleistocene, Eurasia), although having similar basal horn size, show some derived Bovini features, namely marked grooves, deep pneumatization that extends into the horn core and wide horn divergence, which allow us to separate them from the CN sample (Figs 23Y; 24H, M). The horns of putative stem Bovini/early Bovini such as Selenoportax, Pachyportax, and Proamphibos (South Asia) can be differentiated from both A. tigneresi and the CN bovids by their greater mediolateral compression and more developed anterior keel (Selenoportax spp. and Pachyportax spp.), stronger torsion (Selenoportax spp. and Pachyportax spp.), and more pronounced medial curvature (Selenoportax vexillarius, Pachyportax giganteus, and Proamphibos spp.) (Figs 17A; 18C; 23G; 24I, L). Lastly, species of Ugandax (Pliocene Africa) differ from CN bovid in keels less evident, more pronounced medial curvature, deeper surface grooving, and greater divergence, all indicating that this genus belongs in the Bovini (Fig 23V). Horn core dimensions in CN specimens are similar to those in Parabos, Alephis, Grevenobos, Ugandax, Proamphibos and Leptobos (Fig 17A). These forms all have horn bases lacking significant mediolateral compression, unlike the moderate compression in Pachyportax and Selenoportax and the extreme compression in Miotragocerus and Tragoportax (Fig 18C). In particular, the studied sample fall within the cluster represented by the large-sized bovids with weakly compressed horns such as A. lyrix, A. tigneresi, and Parabos spp. (i.e., Parabos-Alephis cluster), also partially overlapping with the crown bovini group formed by Leptobos, Proamphibos and Ugandax spp (Figs 17A; 18C).

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Comparative cranial measurements (mm) of selected Parabos and Alephis samples. Parabos cordieri from the type locality of Sables de Montpellier, Parabos savelisi from the type locality of Gephyra, Parabos tigneresi from the type locality of Baho, Alephis lyrix from the type locality in Perpignan. Measures as in Table 1. Sources: a, this study, 1, [15]; 2, [52].

https://doi.org/10.1371/journal.pone.0340256.t003

The morphology of the braincase in the CN bovid aligns it with the earliest stem Bovini. While the mastoid region is larger than that of Miocene Tragoportacini, the braincase relative length is shorter. Similar proportions can be observed in more evolved forms such as Parabos, Alephis, Selenoportax and Ugandax. The two convergent temporal crests are elevated above the parietals and the supraoccipital constriction, despite being well-developed is not as long and narrow as it is in Miotragocerus or Tragoportax (Fig 23B, E, Q). The CN bovid supraoccipital is similar to that in Selenoportax, Parabos, Alephis lyrix and the earliest forms of Leptobos ex gr. SEM (i.e., L. stenometopon, L. elatus, L. merlai) but longer and narrower than in later Leptobos ex gr. EV (i.e., L. etruscus, L. vallisarni) (Fig 23H, K, N, Q, T, Z). The supraoccipital in IPHES.CN’11-B4-1 falls within the variation of Leptobos ex gr. SEM and next to the cluster formed by A. lyrix, P. cordieri and P. savelisi, which have a narrow but relatively short supraoccipital (Fig 18B). Tragoportacini are characterized by an extremely elongated and narrow supraoccipital. On the other side of the spectrum are derived Bovini such as Bison, which has a very anteroposteriorly compressed and wide supraoccipital. This element does not form a part of the skull roof and migrates on the occipital plane and became indistinguishable from the nuchal crest. This character is clearly visible in Fig 18E where the CN bovid and the Parabos-Alephis group have intermediate values of compression index between the crown Bovini and Tragoportacini. The rugosities covering the cranial roof, typical of Boselaphini and most Miotragocerus and Tragoportax species seems to be extremely reduced in the CN bovid. The basioccipital of the CN bovid is characterized by a triangular shape, with narrow, elongated anterior tuberosities and wide, bulging posterior tuberosities separated by a sagittal groove (Figs 5F; 23R). Bovini are distinguished from Tragoportacini by greater width across the posterior tuberosities (i.e., greater size), with the CN specimens intermediate between the two tribes, partially overlapping with latter group and the early Bovini (Fig 18A). The index given by the ratio between the anterior and posterior tuberosities indicates strong homogeneity within the comparative sample (Fig 18F). This result suggests that the relative widths of the anterior and posterior tuberosities do not distinctly differentiate the two tribes, although the most basal tragoportacines (e.g., Miotragocerus) tend to have more squared and smaller basioccipital. The sagittal groove of the basioccipital is quite narrow but marked by steep and clear walls, strongly resembling the condition in some specimens of Tragoportax and Selenoportax and differing from most of the other bovids (Fig 23I, R).

The dentition of the CN specimens is large (Table 2) with teeth larger than Tragoportax and Miotragocerus and in the size range of bovines such as Selenoportax, Pachyportax, Parabos, Alephis, and Leptobos (Figs 17B, 18D, 19, 20; Table 4). The premolar row is reduced compared to the molar row, a typical trait in Bovini and differing from the long premolars in Tragoportacini (Figs 16B; 18D). Fig 18D shows the CN specimens are of similar premolar reduction as crown Bovini and extant Boselaphus whereas the only specimen of Alephis tigneresi from Baho and early Bovini possess slightly longer premolar row and Tragoportacini has the highest values for this element.

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Comparative dental measurements (mm) of selected Parabos and Alephis specimens. Parabos cordieri from the type locality of Sables de Montpellier, Parabos tigneresi from the type locality of Baho, Alephis boodon from the type locality of Alcoy, Alephis lyrix from the type locality in the Perpignan basin. Measures as in Table 2. Sources: a, this study; 1, [15]; 2, [52]; 3, [53].

https://doi.org/10.1371/journal.pone.0340256.t004

The overall tooth morphology of the CN bovid resembles that of Late Miocene Asian forms such as Selenoportax and Pachyportax, as well as Pliocene European genera including Parabos and Alephis. Features such as sparse cementum, relatively weak styles/stylids, small goat folds on m3, a mesiodistally expanded metastyle on p4, and rugose enamel distinguish the CN sample from more advanced Bovini, such as Leptobos, Ugandax, and Proamphibos [63,69,79, among others]. Conversely, the presence of incipient labial and lingual basal pillars, although primitive, sets the CN bovid apart from the typical Tragoportacini of the Eurasian Miocene in which this feature is less developed. In both upper and lower molars, the CN remains have stronger ribs, larger styles and more complex internal enamel islets than in Miotragocerus and Tragoportax. Additionally, traces of cementum and the labially shifted hypoconulid on m3 distinguish the CN remains from Miotragocerus, Tragoportax and Parabos cordieri. Despite these differences, the studied sample are associated to the aforementioned genera by a p4 with a mesiodistally expanded metaconid and consequently reduced 2^nd valley. In younger individuals, the metaconid is well-separated from the paraconid and projects distally, creating a wide 2^nd valley. In more worn p4s, the metaconid expands mesially and gradually acquires the typical Tragoportacini morphology. The specimens characterized by very late wear stages show a narrowing of the 2^nd valley, caused by the nearing of the paraconid and metaconid. This character is shared with P. cordieri but not with A. lyrix, in which the metaconid does not expand mesially and the 2^nd valley rests open with wear. The CN bovid, however, is similar to A. lyrix in the closure of the 3^rd valley in mid wear which, on the contrary, is always open in P. cordieri [20]. The hypsodonty index calculated on the least worn lower molars (specimen IPHES.CN’12-B7-1) shows that the CN bovid presents a mesodont condition, fitting between the more hypsodont tribes (e.g., Aepycerotini, Alcelaphini, Bovini) and the brachiodont Boselaphini, Neotragini and Tragelaphini (Fig 21). Interestingly, the CN individual, when compared with the two fossil groups analysed, namely Late Miocene stem Bovini from Asia and Tragoportacini (i.e., Tragoportax and Miotragocerus from Pikermi), shows values that are between the two groups in m1 and m2. On the other hand, the m3 value is slightly above the results obtained for the Selenoportax-Pachyportax complex (Fig 21C). In summary, the dentition of the CN bovid is more primitive than advanced Bovini and more derived than Tragoportacini, and most similar to stem or early bovins like Parabos, Alephis, Pachyportax and Selenoportax.

The limb bones of the CN bovid have similar size and proportions to Alephis and Leptobos, being larger than both Tragoportacini and P. cordieri (Fig 22). The metapodials are relatively elongated and lack the reduction typical of fossil and extant Bovini but do not show the extreme elongation typical of Antilopini (Fig 22B, D). On the contrary, relative metacarpal and metatarsal lengths are similar to those in Tragelaphini, Boselaphini, and Hippotragini, suggesting a plesiomorphic bovid condition. The relative lengths of the radius and tibia largely confirm this, though here the differences among bovid tribes are not so evident (Fig 22A, C).

The CN bovid exhibits a distinctive combination of Tragoportacini and Bovini traits, similar to the Pliocene genera Parabos and Alephis. As extensively discussed above, the large size, the reduction of premolars in relation to the molars, the higher hypsodonty, the shortening of the postcornual section of the skull and the morphology of the horncores (e.g., presence of three well-developed keels without stepped anterior keel, weak transversal compression), differentiate the studied sample from Tragoportax, Miotragocerus and most of their Miocene tragoportacin allies [24,65,97]. At the same time the CN bovid has less derived dentition and neurocranium than Alephis lyrix, Grevenobos, Ugandax, Proamphibos and Leptobos, among others. The comparison with Parabos species reveals notable similarities which are, however, contrasted by some features derived in the direction of Alephis, namely the larger size, the incipient terminal horn torsion, the more divergent and slightly medially bending cores, the faint grooving of the medial surface, the larger and more massive dentition, among others. These characters ally the Camp dels Ninots with the species Alephis tigneresi from the Pliocene French locality of Baho confirming the preliminary attribution provided by Gómez de Soler et al. [9]. The record from CN therefore allows us to better define the morphological characters of this otherwise poorly known species. This study evidences several characters, shared with the genus Parabos, which clearly segregate A. tigneresi from Alephis and its type species A. lyrix. These characters include the weak medial bending nor significant twisting of the horn, the presence of three evident keels (i.e., clear triangular cross section of the core), the lack of deep grooves on the horncores’ surface, the anteroposteriorly elongated supraoccipital and the strong parietal crests with depressed frontotemporal surface.

Given the multiple differences existing between A. lyrix and A. tigneresi and the similarities found between Parabos and A. tigneresi we suggest the relocation of A. tigneresi, hereinafter named Parabos tigneresi, from Alephis to Parabos.

Reappraised taxonomy of Parabos and Alephis

Arambourg and Piveteau [16] erected the genus Parabos to accommodate the abundant remains of the bovine-like antelopes from the Pliocene localities of Sables de Montpellier (France) and Alcoy (Spain) which were firstly described as Antilope cordieri [45] and Antilope boodon [98] respectively. Both reallocated to Palaeoryx by Depéret [48], who also referred the Pliocene remains from the Roussillon/Perpignan Basin sites (France) to Palaeoryx boodon. Following this, Arambourg and Piveteau [16] erected the genus Parabos to accommodate both these species. Finally, Gromolard [15] reassigned the Roussilon P. boodon (but not the Alcoy sample) to a new genus and species, Alephis lyrix. Due to their similar morphology and geochronological distribution, Parabos and Alephis are often considered strictly, if not directly, related [15, 19, 24, 53, 55, 63, among others]. Further species have also since been assigned to these genera. Currently, Parabos includes the type species Parabos cordieri (de Christol, 1832) [45], Parabos savelisi Crégut-Bonnoure and Tsoukala 2017 [54], Parabos athanasiui Simionescu 1922 [99], Parabos macedoniae Arambourg and Piveteau 1929 [17] and Parabos soriae Morales 1984 [51]. Alephis includes the type species Alephis lyrix Gromolard 1980 [15] and Alephis tigneresi Michaux et al. 1991 [52]. ‘Antilope’? boodon Gervais, 1853 [98] which has been referred alternatively to Parabos [e.g., 15, 21, 51] or Alephis [e.g., 24, 53, 74] (Fig 2). In this study, we argue for the reassignment of Alephis tigneresi to Parabos. To help make this clearer, we next review the record of these genera.

Parabos cordieri occurs in many MN14 localities across Western Europe (Fig 2) [21]. The main traits that characterize this form include straight/slightly diverging, untwisted horncores with triangular cross-section at the base, strong anterior keel, large frontal foramina placed anteriorly to the horn pedicles, weakly raised parietal crests, elongated supraoccipital and dentition with reduced basal pillars and hypsodonty [15,18,24]. The analyses performed in this study evidence that the morphology, size and proportions of P. cordieri horns teeth and neurocranium differentiate it from all the known Tragoportacini and crown Bovini (previous chapter and Figs 17–20; 23; 24). The teeth show primitive features such as small stylids and low crowns, however, their size is consistently larger than all Miotragocerus and Tragoportax comparative sample, allying them with early Bovini (e.g., Proamphibos, Leptobos; Figs 19; 20). The horns are larger and less laterally compressed than the ones observed for Tragoportacini. Despite that, the presence of three well-developed keels testifies a primitive morphology which clearly separates P. cordieri also from crown Bovini (Figs 17A; 18C; 24F). At the same time, the postcornual section of the skull (e.g., supraoccipital area) is less elongated than in Miotragocerus-Tragoportax complex but still does not feature the typical squared and anteroposterior compressed morphology of later bovines (Figs 18E; 23). This unique mixture of traits makes difficult the positioning of this species (and of the whole genus) within Bovinae. Despite most of the authors refer the clade to Boselaphini or Tragoportacini its position within this group is not fully supported (see also below).

Parabos savelisi was first described by Crégut-Bonnoure and Tsoukala [54] on the basis of two partial neurocraniums (one male and one female) with horncores and an isolated, fragmentary metatarsal from the Greek locality of Gephyra (MN16, Fig 2). This small species of Parabos, is characterized by horncores with an ovoid cross-section, lacking the prominent keels that are distinctive of both P. cordieri and P. tigneresi (Fig 24E, F, G). This species is also distinguished by the least divergent and straightest horns of the genus (Fig 23J). The supraoccipital is elongated and narrow and the parietal roof stretches posteriorly so the postcornual portion of the cranium is reminiscing of the largest Tragoportacini such as Tragoportax (Fig 23K). The frontal foramina are located posteriorly on the internal side of the pedicle and lack of clear grooves for their accommodation. The frontal sinuses stop right after the pedicles and extend into the horncores only in one specimen (GAS 23). All these characters indicate a plesiomorphic stage for this species of Parabos, linking it with Boselaphini/Tragoportacini. Trifonov et al. [100] referred a well-preserved cranium from Erikdere terrace V (Turkey, MN15–16) to P. cf. savelisi, although Kostopoulos [22] suggested that this specimen could be more closely related to Tragoportax than to Parabos. Based on the position and morphology of the horns—specifically, the mediolaterally compressed base and the faint step on the anterior keel—we are inclined to agree with the latter author, though a more comprehensive description of the fossil is needed for further clarification.

Parabos soriae from Venta del Moro (MN13, Spain) is a poorly understood species, known from a few isolated teeth, four incomplete horncores, and postcranial elements (Fig 2) [51]. This taxon appears to be smaller and less derived than the type species of the genus [51]. Morales [51], describing the species, noted the primitive aspects of the teeth which are more reminiscent of the Tragoportacini (Boselaphini in [51]) condition (i.e., weak hypsodonty, small ribs). The two most complete horns from Venta del Moro, fall within the variability of Parabos spp., next to the specimens referred to P. cordieri and P. savelisi but far from the gigantic P. tigneresi of Baho and Camp dels Ninots (Fig 17A). The triangular section of the horncores described by Morales [51] differs from the ovoid cores of P. savelisi but closely resembles those of P. cordieri and P. tigneresi. Their small size allied them better to the type species of the genus. Notwithstanding these similarities, the evident short and stocky appearance of P. soriae clearly differentiate it from the Parabos from Montpellier. At the present state of the art P. soriae represents the oldest and the most primitive member of the genus. The paucity of remains referred to this species, however, severely hampered its diagnosis and does not allow to infer more on its relationship with other members of the genus.

Parabos athanasiui is represented by an almost complete cranium, accompanied by a few postcranial bones from Malusteni (Romania, MN 15a), and sparse remains from other Romanian localities (Fig 2) [20,62,101–104]. Based on the illustrations provided by Simionescu [62], the Malusteni skull exhibits notable differences from the typical Parabos morphology. These include extremely reduced horn cores, their position relative to the orbit, their curvature, and the absence of a prominent keel (Fig 24C). Despite these distinctions, the dentition and horn size are comparable, though not entirely overlapping, with those of Parabos species (Fig 17A, B). This combination of conflicting features has been recognized by Radulescu et al. [103,104], who placed the generic attribution of “Parabos” athanasiui in quotation marks. Given its pronounced divergence from the observed morphology of Parabos, we recommend excluding this taxon from the genus Parabos until a thorough reassessment of the material assigned to this taxon is conducted.

Parabos macedoniae, firstly described from the Pliocene of Greece by Arambourg and Piveteau [17], is based on few teeth and postcranial bones of a large-sized bovid found in the “Falaise de Karabouroun” (= Megalo Emvolon; Greece, MN15; Fig 2). The lack of cranial material severely affected the definition of P. macedoniae which has been considered a dubious member of the genus [17,20,22,54]. Arambourg and Piveteau [17] provide a basic description of the teeth that appear to be relatively primitive and resembling to the ones of P. cordieri (e.g., rugose enamel, absence of cement, poorly developed ribs and pillars). The size and proportions of the available limb bones do not vary from the variability observed in other Pliocene large sized bovids such as A. lyrix, P. tigneresi and G. antiquus [17,20,74]. The extremely scanty material referred to this species does not allow to properly define any diagnostic features and hence its taxonomic status. The nomen dubium status for the taxon is hereby proposed. The Megalo Emvolon remains are here referred to Bovinae indet.

Alephis lyrix is a large bovid from the MN14–15 of Western Europe (Fig 2), characterized by several derived features which suggest its belonging to an early lineage of Bovini [15,19,22–24,53]. Its resemblance to the Asian genera Pachyportax and Selenoportax is evident in the strong anterior keel, and spiralization of the horns although the European genus displays more advanced traits, such as deeper grooves, weaker mediolateral compression of the cores, and a less depressed postcornual skull roof (Figs 17A; 23; 24) [15,24,44,77]. Apart from Parabos, the Late Miocene Samokeros minotaurus from Greece and Iran shares the closest morphological similarities with A. lyrix. Both taxa exhibit massive, medially curved horncores and a reduced, box-shaped neurocranium [97]. However, Samokeros differs from Alephis for the keeless, distally flattened horncores, narrower angle of horn emergence, posteriorly placed, unsunken frontal foramina and more primitive dentition (Fig 24D, J) [36,73,97]. The Asian and African counterparts of A. lyrix could be considered Proamphibos lachrymans and Ugandax demissum respectively (Figs 23S–X; 24J–L). These three species are indeed sharing numerous Bovini apomorphies characterizing the size and the morphology of the horncores and dentition, although both the Asian and African taxa differ from A. lyrix by their slightly more advanced features [18,63]. Their presence during a timespan contemporaneous with or even preceding A. lyrix challenges any direct ancestor-descendant relationship [24,69]. Currently, the classification of Alephis within the Bovini is widely accepted. However, its relatively primitive morphology places it among the most basal forms within the tribe [2, 15, 23, 24, among others]. Its relationship with Parabos and Samokeros is still unclear although the affinity with the former genus seems stronger (see below).

Alephis boodon was first described by Gervais [98] as Antilope? boodon, based on scant dental and postcranial remains from Alcoy-Mina (MN14–15, Spain, Fig 2). The attribution to Antilope, although uncertain, was given due to the similarities found between the Spanish and the French remains ascribed to Antilope cordieri by de Christol [45]. Both taxa were subsequently moved in Palaeoryx [48] and then lumped in Parabos [16]. Gromolard [15], in her reappraisal of Parabos established a new genus (i.e., Alephis) and separated the Roussillon specimens from the type material of P.? boodon, creating a new taxon which became the type species of the genus, namely Alephis lyrix. The Iberian remains, however, were left in open nomenclature and largely forgotten by scholars, including Michaux et al. [52], who revisited the taxonomy of the Parabos-Alephis group. Montoya et al. [53] reassigned Parabos? boodon to Alephis on the basis of its similarities with A. lyrix. The authors however remark that the extremely scant type material hinder any tentative of a proper taxonomic attribution. We here confirm that the dentition of A. boodon are similar in morphology and size to A. lyrix as well as P. tigneresi and lack important distinctive features that could facilitate differentiation. In the absence of horn cores or other more diagnostic materials, we propose that A. boodon is a nomen dubium and refer the Alcoy Mina material to Bovinae indet. cf. Parabos vel Alephis.

Alephis tigneresi was erected by Michaux et al. [52] on the basis of few remains of a large bovid from Baho, a village in the Roussillon/Perpignan Basin (MN 14, France, Fig 2). The material, likely referrable to a single individual, includes two horncores, a fragment of the skull (i.e., portion of the right frontals, parietals and temporals), several teeth and postcranial bones. The remains of Camp dels Ninots, already attributed to this taxon by Gómez de Soler et al. [9] and hereby confirmed, represents the richest and most complete record for this species and the whole genus. The analyses performed allowed us to recognize several plesiomorphic traits which ally P. tigneresi to Parabos (see previous chapter) and therefore compel the moving of the species from Alephis to the latter genus. At the same time, both the Iberian and French record, referred to P. tigneresi are characterized by intermediate morphology between Parabos cordieri –“primitive” condition– and Alephis lyrix –“derived” condition–(Figs 23P–U; 24G, J). The intermediate characters between P. cordieri and A. lyrix of P. tigneresi are hereby provided. Parabos tigneresi features posteriorly directed horncores with smaller diverging angle compared to Alephis but larger than the one shown by the subparallel cores of P. cordieri. The incipient distal torsion and the weak medial furrows of P. tigneresi cornual appendages separate it from both A. lyrix (strongly spiralized and grooved horns) and P. cordieri (straight, untwisted and smooth horns). Although the horn keels of P. cordieri resemble the ones of P. tigneresi, in the latter taxon the posterolateral keel is blunter. This character is amplified in A. lyrix, so much that this keel disappears. The development of the supraoccipital in P. tigneresi appears to be intermediate between the ones of P. cordieri (longer) and A. lyrix (shorter). Lastly P. tigneresi present a dentition similar to the one of Alephis (large teeth, strong basal pillars, presence of cement) which separates it from the more primitive Parabos cordieri. This evidence suggests that P. tigneresi may either represent an intermediate form between P. cordieri and A. lyrix as part of a monophyletic group or reflect independent convergences toward Bovini morphology. A phylogenetic analysis is needed to test whether Alephis could have originated from a derived Parabos similar to P. tigneresi.

The position of Parabos and Alephis within Bovinae

It is generally accepted that Bovini evolved from Miocene ‘Boselaphini’ [5,18,23,79, among others]. Features such as large size, divergent horns, and high crowned teeth with large basal pillars, among others, discriminate taxa such as Selenoportax and Pachyportax from Miotragocerus, Tragoportax and allies, traditionally placed within ‘Boselaphini’, and place them on the line of Bovini [18,78,79]. Many of these characters are also present in Parabos and Alephis, which could represent either stem Bovini, or ‘boselaphins’ that independently converged on Bovini-like morphology.

Gromolard [15], presenting an emended diagnosis of Parabos, emphasized the primitive aspects of the genus and referred it to Boselaphini (Tragoportacini in this work). Over time, most of the scholars have supported this classification [e.g., 20,24,54]. Indeed, the type species Parabos cordieri differs notably from the contemporaneous African taxa which had already developed clear derived features in the direction of Bovini [24,69]. For example, Ugandax sp. from the Late Miocene of the Middle Awash (Ethiopia) and Ugandax demissum from the Early Pliocene of Langebaanweg (South Africa) show advanced dentition, stouter and more divergent horn cores, reduced neurocranium and a broader occipital surface, than Parabos [24,63,105]. At the same time, this work shows that Parabos does show characters that clearly distinguish it from fossil ‘Boselaphini’ like Tragoportax and Miotragocerus, and suggest, instead, affinity to (or convergence with) purported stem bovins from Miocene of Asia like Selenoportax and Pachyportax (Figs 17–24) These include the absence of sculpturing on the frontoparietal surface, the reduction of the premolar row relative to the molar row, the strong ribs and basal pillars on molars the incipient hypsodonty, the shortening of the neurocranium and the large size. Therefore, while Parabos might not belong in Bovini, it also does not seem to belong in Tragoportacini or ‘Boselaphini’, as already partially postulated by Montoya et al. [53]. In a phylogenetic analysis, Geraads [23] recovered P. cordieri either outside or at the very base of Bovini, confirming the similarities we and others have noted and suggesting, however, that Parabos might be part of ‘Boselaphini’ (here Tragoportacini); a view shared also by Gentry [2] and Bibi [24]. If this is true, these taxa would have independently and parallelly developed Bovini traits during the latest stages of Neogene. Nonetheless, some results in the phylogenetic analysis performed by Bibi [24] points toward the inclusion of both Parabos and Alephis in stem Bovini. Finally, Montoya and colleagues [53] suggest a more cautious approach evidencing that referring these genera to either one of these tribes is extremely difficult. To make matters more complicated, while many authors agree that Alephis is an early Bovini and Parabos is not [15,20,21,24], others have proposed that Alephis evolved from Parabos [53,63]. If the latter statement is true, the separation of the two genera in different tribes would be difficult to justify.

The new CN material shed light on the morphology of Parabos tigneresi, a species that until now was classified within the genus Alephis [52] suggesting a close relationship between the two genera. Indeed, although the stratigraphic ranges of Parabos and Alephis remains uncertain, this does not rule out an ancestor-descendant relationship. As demonstrated in this work, key characters shared by these two taxa and other Eurasian genera commonly referred to early Bovini/stem-Bovini (e.g., Pachyportax, Selenoportax, Grevenobos) address to a relationship between them and support their inclusion in Pan-Bovini [24,44,69,74,77]. This would imply the presence of a common ancestor which evolved in Eastern Palaearctic and, somewhat during Miocene, reached Europe, giving rise to the Parabos-Alephis lineage which, however, preserved relatively primitive features compared with the roughly coeval African early Bovini (i.e., Ugandax). In contrast, the highly primitive morphology of P. soriae and P. savelisi suggests a close relationship with the Tragoportacini, indicating that the genus may have originated from a boselaphin-like lineage at the Mio-Pliocene boundary and evolved convergently toward the Bovini trough the Ruscinian. The presence of multiple Bovini-like offshoots in Europe during the Late Miocene to Early Pliocene should not be ruled out (see also Samokeros). The widely recognized recurrence of homoplasies in Bovidae complicates this issue [2].

On the light of the current knowledge, Parabos and Alephis: might be either a lineage diverging from the stem group of Pan-Bovini (sensu [69]) or a late surviving lineage of Tragoportacini morphologically convergent with early Bovini (Fig 25). A third hypothesis sees Parabos and Alephis not part of the same lineage and independently nested within the two tribes. A comprehensive and updated phylogenetic analysis of Mio-Pliocene Bovinae is of the utmost importance to shed light on the first evolutionary steps of Bovini and the potential presence of ghost stem groups.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 25. Suggested phylogeny of Bovinae showing the possible relationships and chronological ranges of Parabos, Alephis and other Miocene to Pliocene bovids.

Alternate scenarios: 1, Parabos and Alephis are part of a lineage diverging from the stem group of Pan-Bovini; 2, Parabos and Alephis represent a late surviving lineage of Tragoportacini.

https://doi.org/10.1371/journal.pone.0340256.g025

Palaeoenvironment and palaeoecology of Parabos and Alephis

By the end of the Miocene, the stable dry conditions that had facilitated the expansion of the so-called Pikermian chronofauna across Eurasia and Africa were replaced by a more humid climate marked by regional differentiations, which culminated in widespread reforestation, particularly in Mediterranean Europe [106,107 among others]. At the onset of the Early Pliocene (ca. 5.33–3.6 Ma), the increased seasonality and humidity fragmented the vast open landscapes of the Western Palaearctic. This transition triggered the final decline of the Pikermian chronofauna and severed long-lasting ecological connections between Africa and Eurasia [106,107]. In western Mediterranean Europe, the Late Miocene bovid community, primarily composed of Tragoportacini and Antilopini, was already poorer than that of southeastern Europe, which, in contrast, was characterized by a high diversity and strong affinities with assemblages ranging from Anatolia to Afghanistan [22,108,109]. Possibly from the latest Miocene (MN13), but certainly by the Early Pliocene, Parabos and Alephis became the sole large-sized bovid in Mediterranean Europe, especially in its western regions (Fig 2).

The revised age of Camp dels Ninots at ca. 4.4 Ma places it within the Early Pliocene warm period. Fossil pollen indicates that the vegetation in the region of Camp dels Ninots included many extinct thermophilus and hygrophilous species, as well as temperate taxa with Mediterranean affinities and high-altitude conifers [10,110]. The immediate surroundings of the lake at Camp dels Ninots were dominated by a riparian forest, characterized by plant species adapted to wet conditions, whereas the regional vegetation around the site was composed of a mixture of subtropical and temperate elements, reflecting a transitional environment [10,110,111]. The CN paleoenvironment was relatively warm and humid, and characterized by small-scale cyclical changes between hygrophilous and xerophilous plants, correlated to precipitation oscillation [10,110].

Kingdon [112] suggested that Bovini ancestors were forest-dwelling browsers with simple and mesodont/brachydont dentition. Bibi [69] proposed that the transition towards more open habitats marked by seasonal drying during the Middle to Late Miocene led to a dietary adaptation that produced the first stem bovins. Teeth showing higher crowns, increases pillars and occlusal complexity appear as early as 8.9 Ma, in the Siwaliks (Pakistan, India) and in the Irrawaddy beds of central Myanmar [44,69]. The westward spreading of C₄ grass during the later Miocene led to the expansion of these early Bovini in the Middle East and Africa [69]. In contrast, the first bovin-like dentitions in Western Europe only appear in the Pliocene, with Parabos and Alephis. Whether the more humid and forested habitats of Western Europe [106,113] were responsible for the delay in the appearance of these traits in the continent is still to be demonstrated. Crown Bovini with fully developed bovin dentition appears in Africa and South Asia by the earliest Pliocene (e.g., Ugandax coryndonae, Proamphibos spp.) while in Europe this is only with the arrival of the first Leptobos from the Eastern Palaearctic at the very end of Pliocene [22,28 and reference therein]. According to the hypsodonty index proposed by Fortelius et al. [29] for large mammals the values found for CN P. tigneresi corresponds to a mesodont dentition (i.e., 0.8–1.2). The index computed for the lower molars of the specimen IPHES.CN’12-B7-1 is below the values typical for extant hypsodont tribes such as Alcelaphini, Caprini and Bovini, and higher than the ones of more brachiodont bovids including Boselaphini, Cephalophini and Tragelaphini (Fig 21). The same index calculated for early Bovini from Asia provided values that fit within the variation found in Bovini, higher than in the specimen from CN. Tragoportax and Miotragocerus from the Miocene of Pikermi, show low crown dentition, especially in m1 and m3, similar to the one observed in Boselaphini and Tragelaphini (Fig 21). The intermediate position of the CN specimen between stem bovins and tragoportacins remarks the presence of an incipient hypsodonty in P. tigneresi which, however, still retain a certain degree of plesiomorphic characteristics.

According to Köhler [108], Parabos cordieri was well-adapted to wooded, humid habitats, exhibiting limb bone proportions and body mass like those of modern Tragelaphini and some Bovini. Its relatively short radius and tibia suggest a slow-paced mode of locomotion suited for forested environments, rather than cursorial adaptations for open landscapes which, on the contrary, are commonly associated with longer appendicular elements [e.g., 108, 114, 115, 116]. Köhler [108] also noted that the morphology of the metacarpal—characterized by a low distal articulation surface, weak trochlear ridges, and an inflated distal diaphysis—is typical of species inhabiting highly humid forests (metacarpal type A2 in [108]). The CN material shows that Parabos tigneresi does not significantly differ in general morphology and proportions from P. cordieri with the exception of the larger size. Phalanges are extremely informative elements for the type of habitat [108,117]. The proximal phalanx of the bovid from CN presents a rough surface on its axial side and its distal articulation is visible in anterior (=dorsal) view. The intermediate phalanx has a deep sagittal groove in its posterior (=palmar) side, there is an elongated post-articular plateau on the proximal section and the axial outline of the distal articulation is subtriangular. Lastly, the distal phalanx does not have a process above (i.e., anteriorly) the articulation and the latter is elevated from the ground level due to a conspicuous bone growth below (i.e., posteriorly) to it. All these characters fit with the morphology expected for bovids adapted to densely covered habitats [108,117]. Therefore, it is likely that both P. cordieri and P. tigneresi inhabited forested, relatively humid ecosystems. The analyses of Fig 22 show interesting results regarding other large Bovids of European Plio-Pleistocene. The limb lengths of Alephis are almost identical to Parabos suggesting that this taxon did not undergo significant specialization in its postcranial skeleton during the late Ruscinian (Fig 22). Notably, also Leptobos, despite being characterized by more derived dentognathic traits, exhibits limb bone proportions which are divergent from the variability observed in younger fossil and extant Bovini, especially in the metacarpal, positioning in between the groups formed by these Bovini and the Parabos-Alephis cluster (Fig 22B). This suggests that, at least during the earliest stages of the Quaternary, European Bovini did not yet display a significant shift toward stouter builds or gigantic body size; a transformation that occurred much later with the arrival of the first Bison.

Body mass estimates calculated on the basis of 11 dental and postcranial measures (S31 Table) for nine individuals from CN yields an average weight of 419 ± 31 kg for Parabos tigneresi, which makes it the largest bovid in Europe prior to the end Pliocene (Fig 26; S32 Table). The individuals with the smallest horncores are also the lightest with an estimated mass of 378 ± 60 kg (IPHES.CN’19-B15) and 397 ± 33 kg (IPHES.CN’04-B1) and therefore were interpreted as females. The rest of the sample, due to the size of the horns and postcranial, was referred to male individuals. In particular, the individuals bearing the largest horns, were estimated at 420 ± 37 kg (IPHES.CN’05-B2), 456 ± 55 (IPHES.CN’12-B10) and 408 ± 80 kg (IPHES.CN’17-B14). The difference between the minimum and maximum body mass estimations is approximately 100 kg (min = 378 ± 60 kg; max = 481 ± 51 kg) with the lightest specimen weighing slightly 20% less than the heaviest. In extant large bovins, females are typically smaller than males, often averaging 60–70% of male mass [118]. Therefore, it appears that sexual body mass dimorphism in P. tigneresi was not as pronounced as in extant bovins. Rather this is more similar to the condition in Alcelaphini and Hippotragini, in which males and females show larger mass overlap. This relative monomorphism is well-reflected in the low degree of horn size variation within the CN sample (see Figs 6–8; Table 1). An example of this can be found in IPHES.CN’11-B4 which, despite being the heaviest individual, and tentatively referred to a male, does not possess large horns. According to Packer [119], large-bodied animals rely less on rapid flight or crypsis as primary anti-predator strategies and instead tend to develop prominent cornual appendages as a defensive adaptation, especially in open environments where concealment within vegetation is less viable. Consequently, females of large-sized bovids living in such habitats are more likely to possess horns [120]. Despite its large size, the CN bovid was likely living in forested habitats and not in open landscapes, partially deviating from this predictive model. A comparable case is observed in another large Bovidae, the bongo (Tragelaphus eurycerus), where females possess impressive horns despite inhabiting dense forests [120]. This trait has been linked to strong intraspecific social competition; an explanation that may also apply to P. tigneresi, considering also its inferred ecological and morphological similarities to T. eurycerus. On the other hand, Estes [121] proposed that horns in female bovids may evolve not for defence or competition per se, but as a form of andromimicry (i.e., mimicking male traits) to shield male offspring from despotic aggression by dominant males, thus enhancing their survival. It must be remembered, however, that phylogenetic constraints play a significant role in horn development in females [120], thus the ancestry of Parabos may have influenced the retention of large horns in females more than natural selection. It is indeed true that also in P. cordieri the sexual dimorphism is quite reduced as stated by Duvernois [56] who found difficult distinguishing male and female metapodials. This condition, however, is not common in any other European Tragoportacini or Pliocene Bovini, as in Miotragocerus and Tragoportax the females are either hornless or showing clearly different morphology and size in the horns, and in Leptobos the females are noticeably smaller than the males and hornless [38,56,57,122].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 26. Violin plot of estimated body mass for selected large-bodied late Neogene and early Quaternary Bovidae from Europe.

Number of specimens shown above each sample. See S32 Table for specimens, localities and references.

https://doi.org/10.1371/journal.pone.0340256.g026

The body mass of A. lyrix was estimated to be similar to that of P. tigneresi, i.e., in the 300–500 kg range (422 ± 31 kg; Fig 26; S32 Table). Among extant bovids, only the bongo and small ecomorphs of the African buffalo (e.g., Syncerus caffer ‘nanus’ and ‘brachyceros’) are in this size range. While smaller than most extant Bovini (>500 kg), P. tigneresi is still larger than Parabos cordieri (281 ± 89 kg), Miotragocerus (85 ± 28 kg) and Tragoportax (156 ± 29 kg) (Fig 26; Table S32). Except for the caprine-like antelopes Protoryx and Palaeoryx, the largest bovid of Europe during this period was the obscure and relatively rare Samokeros minotaurus. The clear trend toward large body size (exceeding 400 kg) among multiple lineages of European Bovinae emerged only at the beginning of Pliocene with Parabos tigneresi, Alephis lyrix (421 ± 135 kg), and Grevenobos antiquus (398 ± 107 kg), and continued into the Pleistocene with Leptobos ex gr. SEM (404 ± 99 kg), Leptobos ex gr. EV (446 ± 98 kg) and Bison (Eobison) (489 ± 129 kg; Fig 26; S32 Table). The underlying causes of the shift toward larger sizes in European Bovinae remain unclear, though they might be linked, as it is possible for the increased hypsodonty, to the long-term cooling and aridification trend that affected the Eurasian continent culminating at the Quaternary and eventually leading to the replacement of forested, subtropical ecosystems by drier and more open habitats [123, 124, 125 and references therein] The relationship between habitat shifts and body mass fluctuations in herbivores has been widely studied showing that animals’ size is indirectly influenced by climate changes through vegetation [e.g., 126, 127]. The Jarman–Bell principle predicts larger body sizes in herbivores allow greater digestive efficiency when consuming coarse, low-nutrient vegetation such as grasses [128,129]. Global cooling and the spread of grasslands during the Neogene-Quaternary could have driven increases in body mass [130]. Such a hypothesis predicts increasing proportions of grasses in the diets of bovids from the Miocene onwards, as is also suggested by the higher hypsodonty index values. The diet of the CN Parabos tigneresi, however, is not known, and it remains to be seen if it consumed any appreciable quantity of grass.