First record of Macaca (Cercopithecidae, Primates) in the Middle Pleistocene of Greece.

In this article, we describe an almost complete macaque mandible from the Middle Pleistocene locality Marathousa 1 in the Megalopolis Basin of southern Greece. The mandible belonged to a male individual of advanced ontogenetic age and of estimated body mass ∼13 kg. Comparative metric analysis of its teeth permits its attribution to the Barbary macaque Macaca sylvanus, a species that was geographically widely distributed in Western Eurasia during the Plio-Pleistocene. The dental dimensions of the Marathousa 1 macaque fit better within the variation of the Early Pleistocene M. s. florentina and the Middle to Late Pleistocene M. s. pliocena rather than with the extant representative M. s. sylvanus. Moreover, principal component analysis reveals a better match with M. s. pliocena. However, because no clear-cut diagnostic criteria have been defined to differentiate these European fossil subspecies, we attribute the Marathousa 1 specimen to M. s. cf. pliocena, in agreement with the chronology of the locality. Previously known only from the Early Pleistocene of Greece by some isolated teeth, this is the first record of Macaca in the Middle Pleistocene of the country and one of very few in the eastern sector of the peri-Mediterranean region. We discuss the presence of macaques in the paleolake environment of Marathousa 1, as well as their predation risks from both carnivores and hominins present at the locality.

Introduction

Taxonomy of European fossil macaques

Cercopithecoidea, a morphologically diverse clade of small- to medium-sized primates, are today widely distributed across a vast region of the Old World, from West Africa to Japan, inhabiting a considerable variety of environments. Their earliest record is documented in the Late Oligocene of Africa, where they remained endemic until the late Miocene (Jablonski and Frost, 2010; Stevens et al., 2013), when members of both cercopithecid subfamilies, Colobinae and Cercopithecinae, dispersed toward Eurasia. The cercopithecine Macaca (Lacépède, 1799) is thought to have entered Europe toward the end of the Miocene, at ∼5.5 Ma (Köhler et al., 2000; Alba et al., 2014). Macaca, although evolutionarily conservative (Szalay and Delson, 1979: 355), is among the longest-lived (>5 Myr) cercopithecids in Europe, surviving under various paleoenvironmental conditions and thought to be continuously present on the continent until the Late Pleistocene (Elton and O'Regan, 2014).

All European fossil macaque forms are attributed to the Barbary macaque Macaca sylvanus (e.g., Delson, 1980; Shearer and Delson, 2012; Alba et al., 2019)—with the exception of Macaca majori Azzaroli, 1946 from Sardinia (Italy), considered an endemic insular species (Rook and O'Higgins, 2005; Zoboli et al., 2016). They are generally attributed to three (chrono)subspecies (Szalay and Delson, 1979; Delson, 1980; Alba et al., 2011): M. sylvanus prisca Gervais, 1859 from the Pliocene (type locality: Montpellier, France; Ruscinian; Palombo and Valli, 2003); M. sylvanus florentina (Cocchi, 1872) from the Early Pleistocene (type locality: Le Forre, Upper Valdarno, Italy; late Villafranchian; Rook, 2009; Rook et al., 2013); and M. sylvanus pliocena Owen, 1846 from the Middle until the early Late Pleistocene (type locality: Grays Thurrock, England; Marine Isotope Stage [MIS] 9; Schreve, 2001). The diagnosis of these subspecies relies mainly on differences in dental size and proportions. However, there is great overlap in almost all parameters among the fossil, as well as the extant, subspecies. Subspecific assignments are therefore often based on (bio)chronological grounds rather than on diagnostic morphological features. Indeed, several authors refrain from a subspecific attribution altogether (e.g., Bona et al., 2016; Reumer et al., 2018), whereas others use open nomenclature (e.g., Castaños et al., 2011; Alba et al., 2019).

Although the extant M. s. sylvanus is endangered and currently geographically restricted to Morocco, Algeria (native) and Gibraltar (introduced by humans; Fooden, 2007), its fossil relatives were widely distributed in Europe and Northern Africa (Delson, 1974; Szalay and Delson, 1979; Ardito and Mottura, 1987; Elton and O'Regan, 2014). Nevertheless, very few occurrences of the taxon are known from the eastern sector of the peri-Mediterranean region (Mecozzi et al., 2021: Fig. 4). In Greece, Macaca was only represented by isolated dental specimens from the fissure filling of Tourkovounia-2, Athens, biochronologically dated to the Early Pleistocene, and attributed to Macaca florentina (Symeonidis and Zapfe, 1976; van der Meulen and Doukas, 2001; Doukas and Papayianni, 2016). Here we report a new, almost complete, mandible, MAR-1-9B, recovered during the excavation of the Middle Pleistocene (Lower Paleolithic) locality Marathousa 1, Megalopolis Basin, southern Greece. Furthermore, we discuss the paleonenvironmental setting of Marathousa 1, as well as the potential prey-predator relationships among hominins, carnivores, and macaques, aiming to clarify the conditions under which this individual lived and to contribute to the better knowledge of the fossil macaques of Europe.

Geological, paleontological, archaeological, and chronological context

The Megalopolis area (Arcadia, Peloponnesus) is an intramontane basin, which was filled during the late Neogene and the Pleistocene by fluviolacustrine deposits. During the Middle Pleistocene, the basin hosted a large, shallow lake, which resulted in a stratigraphic sequence composed mainly by lacustrine sediments containing lignite seams, which have been mined for energy production since 1970. As a result of the mining operations, long fossiliferous sections are exposed, offering the unique opportunity to study the stratigraphy and the paleoenvironment of the Pleistocene paleolake. In 2013, a targeted survey of Pleistocene sediments in the basin was conducted by a joint team from the Ephorate of Palaeoanthropology-Speleology (EPS) of the Hellenic Ministry of Culture and Sports and the Palaeoanthropology working group of the University of Tübingen (Thompson et al., 2018). During this survey, the new locality Marathousa 1 was discovered within the Marathousa Member of the Choremi Formation (Fig. 1A; Panagopoulou et al., 2015, 2018; Harvati, 2016; Harvati et al., 2016, Harvati et al., 2018; Tourloukis and Harvati, 2018). Systematic excavations of the site were conducted in 2014–2019, yielding a large Lower Paleolithic lithic assemblage, found in stratigraphic association with faunal remains, some of which exhibit evidence of human processing (Konidaris et al., 2018; Tourloukis et al., 2018a). Post-infrared infrared stimulated luminescence, electron spin resonance, magnetostratigraphy and biochronology indicate an age between ∼0.5 and 0.4 Ma for the site (Blackwell et al., 2018; Doukas et al., 2018; Jacobs et al., 2018; Konidaris et al., 2018; Tourloukis et al., 2018b), and place it during the glacial MIS 12 (Panagopoulou et al., 2018). As such, Marathousa 1 represents the earliest currently known chronometrically dated evidence of human presence in Greece.

Figure 1

A) Geographic position of the Megalopolis Basin and of the Marathousa 1 locality within the Marathousa mine, and distant view of the locality showing the two excavation areas (map of Greece from Copernicus Land Monitoring Service, https://land.copernicus.eu/, satellite image of the mine from Google Earth, and Marathousa 1 photo by V. Tourloukis). B) Simplified stratigraphic column of Area B showing the stratigraphic units and the position of the Macaca mandible (modified from Panagopoulou et al., 2018); stratigraphic data in Karkanas et al. (2018). C) Distribution map of the southern corner of Αrea B trench (UB4c and upper part of UB5a) showing the position of the large mammal fossils and the lithic artifacts (excavation season 2019); red stars denote the Macaca mandibular fragments, which were found separated and refit (map made by D. Giusti, photographs of macaque hemimandibles by N. Thompson). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

The locality preserves a stratified and exceptionally well-preserved archaeological and paleontological assemblage between two lignite seams, in what was once the lake shore (Panagopoulou et al., 2015, 2018). Two excavation areas were defined, Area A and Area B (Fig. 1A), both yielding lithics, microfauna, and macrofauna (insects, ostracods, mollusks, fishes, amphibians, reptiles, birds, mammals), as well as microflora and macroflora (Doukas et al., 2018; Field et al., 2018; Konidaris et al., 2018; Michailidis et al., 2018; Panagopoulou et al., 2018; Tourloukis et al., 2018a; Bludau et al., 2021). The main find-bearing units, UA3c and UB4c, in the two excavation areas, respectively, are correlated to each other and represent the same depositional event (Giusti et al., 2018; Karkanas et al., 2018). They consist of dark gray, massive organic and intraclast-rich silty sand, interpreted to represent mudflows and hyperconcentrated flows that plunged into the shores of the lake and close to mudflats (Karkanas et al., 2018). Area A represents mainly a dense accumulation of elephant bones, belonging to a single individual of the European straight-tusked elephant, Palaeoloxodon antiquus, bearing evidence of human butchering activities (cut marks) and associated with lithic artifacts and other faunal remains, while Area B is characterized by a higher number of lithic artifacts, again found in association with faunal remains, including at least one additional elephant individual (Konidaris et al., 2018; Tourloukis et al., 2018a). In addition to P. antiquus, the locality's large mammal fauna includes the following taxa (Konidaris et al., 2018): Castor fiber (beaver), Mustela sp. (weasel), Lutra simplicidens (otter), Felis sp. (wildcat), Vulpes sp. (fox), Canis sp. (wolf-sized canid), Hippopotamus antiquus (hippo), Bison sp. (bison), Dama sp. (fallow deer), and Cervus elaphus (red deer).

During the 2019 excavation campaign, Marathousa 1 yielded a new faunal element: a cercopithecine primate represented by an almost complete mandible, MAR-1-9B, discovered in the southern corner of the Area B trench (Fig. 1C). It was found within the stratigraphic unit UB4c (Fig. 1B), and in close spatial and stratigraphical association with lithic artifacts and other faunal remains, including freshwater mollusks, fishes, amphibians, reptiles, birds, micromammals, and macromammals. Among the latter are dental and postcranial elements (e.g., complete tusk, vertebrae, ribs) of the elephant P. antiquus, possibly belonging to a single individual, the second one known from the locality.

Materials and methods

MAR-1-9B is housed at the EPS in Athens (Greece). The specimen originally consisted of two hemimandibles, deposited about 1.6 m apart (Fig. 1C): MAR-1-926/587-47, preserving the complete left ramus and corpus and the right symphyseal area around to the anterior part of the P3 alveolus, as well as the dental row from the lower canine to M2; and MAR-1-925/587-33, preserving the right ramus and corpus from its caudal end up to the level of the P3 alveolus, and the dental row P4–M3. Although damaged in the area to the right of the symphysis, the two hemimandibles fit together under the alveolus of the right P3 (Fig. 2). The refitted complete mandible was assigned the new accession number MAR-1-9B.

Figure 2

Macaca sylvanus cf. pliocena (EPS-MAR-1-9B) from Marathousa 1, in left lateral (A), right lateral (B), and dorsal (C) views. Scale = 5 cm.

Standard dental measurements (following Alba et al., 2011, 2019) of maximum mesiodistal crown length (MD) and maximum buccolingual crown breadth (BL) were taken using a digital caliper to the nearest 0.1 mm. For the molars, BL was measured both at the mesial (BLm) and distal (BLd) lophids. A breadth/length index (BLI) was calculated for each tooth (BLI = BL/MD × 100). All dental measurements are reported in Table 1 and mandibular measurements in Table 2. Comparative dental measurements (Table 3, Table 4; following in some cases the taxonomic attributions of Alba et al., 2019, 2021) were obtained from the literature and from publicly available sources (the New York Consortium in Evolutionary Primatology PRImate Morphometrics Online [PRIMO] database; http://primo.nycep.org).

Table 1

Dental measurements (in mm) of the macaque mandible (EPS-MAR-1-9B) from Marathousa 1.

Tooth	Side	MD	BL	BLm	BLd	BLI
C1	Left	8.1	6.3
P3	Left	6.8	5.2			76.5
P4	Right	5.6	6.1			108.9
	Left	5.8	6.1			105.2
M1	Right	7.5		6.4	6.2	85.3
	Left	7.7		6.5	6.3	84.4
M2	Right	8.9		7.3	6.7	82.0
	Left	9.4		7.1	–	75.5
M3	Right	12.7		7.8	7.2	61.4
M1–M3		28.5

Abbreviations: MD = mesiodistal crown length; BL = buccolingual crown breadth for the canine and the premolars; BLd = distal buccolingual crown breadth for the molars; BLm = mesial buccolingual crown breadth for the molars; BLI = breadth/length × 100.

Table 2

Mandibular measurements (in mm) of the macaque mandible (EPS-MAR-1-9B) from Marathousa 1.

Measurement	Right	Left
Total length	88.8	83.5
Total breadth	(62.0)
Height of ramus at articulation	45.5	47.0
Height of ramus at coronoid process	53.0	52.5
Length of ramus	28.7	29.5
Height of corpus below M1	24.3	24.6
Height of corpus below M3	23.5	19.8

The total length was measured from the rostralmost point of the mandible to the caudal margin of the ramus (parallel to the occlusal surface, and parallel to the buccal face of the corpus). Parentheses indicate an inaccurate measurement, due to specimen distortion. Height measurements were taken buccally.

Table 3

Comparative dental sample (fossil and extant) used in the analyses.

Taxon	Extant/fossiliferous site	Source
Macaca sylvanus sylvanus	Extant	Alba et al. (2019, 2021), PRIMO (2021))
Macaca majori	Capo Figari (Italy)	PRIMO (2021)
Macaca sylvanus prisca	Montpellier (France)	PRIMO (2021)
Macaca sylvanus florentina or M. s. cf. florentina	Estació de Vallparadís (Spain)	Alba et al. (2008)
	Monte Peglia (Italy)	Petronio et al. (2020), PRIMO (2021)
	Mugello (Italy)	Alba et al. (2021)
	Quibas (Spain)	Alba et al. (2011)
	Tegelen (The Netherlands)	van Kolfschoten (2020), PRIMO (2021)
	Tourkovounia-2 (Greece)	Symeonidis and Zapfe (1976)
	'Ubeidiya (Israel)	PRIMO (2021)
	Untermassfeld (Germany)	Zapfe (2001)
	Upper Valdarno (Italy)	PRIMO (2021)
	Vallonnet (France)	de Lumley et al. (1988), PRIMO (2021))
Macaca sylvanus pliocena or M. s. cf. pliocena	Cova Negra (Spain)	Pérez Ripoll (1977)
	Gajtan (Albania)	PRIMO (2021)
	Gombasek (Slovakia)	PRIMO (2021)
	Grotta degli Orsi Volanti (Italy)	Mazza et al. (2005)
	Gruta da Aroeira (Spain)	Alba et al. (2019)
	Lezetxiki II cave (Spain)	Castaños et al. (2011)
	Monte Sacro (Italy)	PRIMO (2021)
	Montsaunès (France)	PRIMO (2021)
	North Sea (The Netherlands)	Reumer et al. (2018)
	Orgnac 3 (France)	PRIMO (2021)
	Quecchia Quarry (Italy)	Bona et al. (2016)
	Saint-Estève-Janson (France)	PRIMO (2021)
	Torre in Pietra (Italy)	PRIMO (2021)
	Valdemino (Italy)	PRIMO (2021)
	Zlaty Kun C718 (Czechia)	PRIMO (2021)

Table 4

Descriptive statistics and calculated z-scores for the dental measurements of EPS-MAR-1-9B compared with Macaca sylvanus sylvanus, Macaca sylvanus pliocena, and Macaca sylvanus florentina (see Materials and methods).

		n	Mean	SD	Minimum	Maximum	z-score
P3 MD	MAR-1-9B	1	6.80	–	–	–
	M. s. sylvanus	17	8.06	1.87	5.10	11.50	−0.67
	M. s. pliocena	7	8.37	1.93	6.10	11.10	−0.81
	M. s. florentina	11	9.61	1.70	6.90	12.50	−1.65
P3 BL	MAR-1-9B	1	5.20	–	–	–
	M. s. sylvanus	17	4.60	0.53	3.80	5.60	1.12
	M. s. pliocena	8	4.44	0.55	3.40	5.30	1.39
	M. s. florentina	12	5.23	0.79	3.80	6.40	−0.04
P4 MD	MAR-1-9B	2	5.70	0.14	5.60	5.80
	M. s. sylvanus	19	5.98	0.55	5.10	7.00	−0.51
	M. s. pliocena	10	6.17	0.36	5.50	6.70	−1.31
	M. s. florentina	20	6.14	0.59	4.90	7.30	−0.74
P4 BL	MAR-1-9B	2	6.10	0.00	6.10	6.10
	M. s. sylvanus	19	5.16	0.47	4.20	6.50	2.00
	M. s. pliocena	11	5.34	0.35	4.70	5.70	2.20
	M. s. florentina	20	5.36	0.62	4.70	7.20	1.20
M1 MD	MAR-1-9B	2	7.60	0.14	7.50	7.70
	M. s. sylvanus	26	7.53	0.68	6.00	8.70	0.10
	M. s. pliocena	12	8.14	0.38	7.50	8.90	−1.42
	M. s. florentina	24	8.09	0.39	7.40	9.00	−1.25
M1 BLm	MAR-1-9B	2	6.50	0.07	6.40	6.50
	M. s. sylvanus	24	5.98	0.33	5.20	6.60	1.58
	M. s. pliocena	11	6.47	0.38	6.00	7.00	0.07
	M. s. florentina	25	6.20	0.42	5.70	7.40	0.72
M2 MD	MAR-1-9B	2	9.20	0.35	8.90	9.40
	M. s. sylvanus	25	9.22	0.95	7.10	11.10	−0.02
	M. s. pliocena	15	9.73	0.64	8.80	10.80	−0.84
	M. s. florentina	23	9.66	0.46	8.90	11.00	−0.98
M2 BLm	MAR-1-9B	2	7.20	0.14	7.10	7.30
	M. s. sylvanus	24	7.55	0.72	6.60	8.90	−0.50
	M. s. pliocena	15	7.73	0.54	6.50	8.50	−0.99
	M. s. florentina	24	7.61	0.53	6.70	8.50	−0.77
M3 MD	MAR-1-9B	1	12.70	–	–	–
	M. s. sylvanus	21	11.18	0.99	9.30	13.10	1.53
	M. s. pliocena	13	12.78	0.81	11.70	14.40	−0.09
	M. s. florentina	17	12.40	0.99	10.20	14.50	0.30
M3 BLm	MAR-1-9B	1	7.80	–	–	–
	M. s. sylvanus	21	7.90	0.83	6.40	9.30	−0.12
	M. s. pliocena	13	8.20	0.62	7.30	9.40	−0.64
	M. s. florentina	16	7.76	0.55	6.80	8.80	0.08

Abbreviations: MD = mesiodistal crown length; BL = buccolingual crown breadth for the premolars; BLm = mesial buccolingual crown breadth for the molars; SD = standard deviation.

When z > |1.96|, the null hypothesis that EPS-MAR-1-9B fits within the variation of the comparative sample can be rejected at p < 0.05; in such cases, the z-score is typed in bold.

To estimate body mass, we used the allometric regressions of Delson et al. (2000: Table 7) for the M1 mesiodistal crown length. This measurement is considered the most reliable dental variable for the estimation of body mass of fossil male cercopithecines. However, owing to the heavily worn state of attrition of the M1, we additionally used the M2 and M3 mesiodistal crown length of male cercopithecines and calculated the mean value for all molar positions. Box and whisker plots, statistical computations, and principal component analysis (PCA) were performed with PAST, version 4.05 (Hammer et al., 2001; https://www.nhm.uio.no/english/research/infrastructure/past/). For the PCA, we used only measurements from complete M1–M3 tooth rows of M. s. sylvanus, M. s. florentina, and M. s. pliocena (Supplementary Online Material [SOM] Table S1). To account for the effects of size, dental measurements used in the PCA were transformed to Mosimann's log-shape ratios by dividing each value by the geometric mean of all the variables for each observation and logarithmizing (with the natural algorithm ln) the results (Mosimann, 1970; Jungers et al., 1995). For each cheek tooth measurement and M. sylvanus subspecies used in the analysis, the z-score was computed as z = (x − m)/SD, where x is the dental measurement of MAR-1-9B (mean value in case both right and left premolars/molars are preserved) and m and SD, the mean and standard deviation of the comparative sample, respectively (Table 4).

Systematic paleontology

Order Primates Linnaeus, 1758

Family Cercopithecidae Gray, 1821

Subfamily Cercopithecinae Gray, 1821

Tribe Papionini Burnett, 1828

Subtribe Macacina Owen, 1843

Genus Macaca Lacépède, 1799

Macaca sylvanus (Linnaeus, 1758)

Macaca sylvanus cf. pliocena Owen, 1846

(Fig. 2)

Studied material Left mandibular fragment with canine and P3–M2, EPS-MAR-1-926/587-47; right mandibular fragment with P4–M3, EPS-MAR-1-925/587-33 (the fragments refit forming a complete mandible, designated EPS-MAR-1-9B). The material originates from Marathousa 1 (Megalopolis basin, Arcadia, Greece), dated between 500 and 400 ka (Middle Pleistocene).

Description The specimen is almost complete, preserving both corpora and rami, and bearing the right P4–M3, and the left C1–M2 (Fig. 2). The breakage apparently caused the loss of the right C1 and P3, as well as some damage at their alveoli. A certain degree of deformation is also observed, as a weak skew to the right side. In lateral view, the symphysis is rounded (Fig. 2A, B). The ventral margin of the right corpus is straight, and its height remains the same along the molar row. The left corpus is shallower at the level of the missing M3, resulting in a markedly concave ventral margin in this area; below the M1–M2, it is more robust, exhibiting a convex ventral profile. This unusual morphology of the left corpus may have resulted from a pathological condition. The M3 alveolus is open and shows no macroscopic signs of bone resorption, indicating that the M3 was present during life and only lost perimortem or postmortem. Two mental foramina are located low at the corpus at the level of P3–P4, but their size and relative position differ on each side. On the left, the rostroventral one is larger, with a long, beveled opening, while the caudodorsal one is very small and probably directly connected to the former. On the right side, the rostroventral one is small, while the other one is much larger, round, and situated higher on the corpus. An additional large foramen occurs only on the left side in a high position at the level of the mesial border of the left M2. On the lingual side of the symphysis, traces of two vertically positioned lingual mental foramina are observable. A mandibular corpus fossa is present on the buccal side, which is more evident in the better preserved left hemimandible. The rostral and caudal margins of the rami are slightly inclined caudally relative to the corpora. The coronoid process is positioned somewhat higher than the condyle. In dorsal view (Fig. 2C), the dental arcade is almost V-shaped, with the postcanine tooth series forming a slight curve, and the canines apparently positioned quite close to each other, leaving a narrow space for the incisors, which must have been very small.

The dentition shows an advanced degree of dental wear, with dentine exposure in all postcanine teeth (less in the right M3), indicating an individual of older age. The single preserved left canine is projecting labially, when observed dorsally (Fig. 2C). In lateral view, it protrudes well above the occlusal plane of the cheek teeth (Fig. 2A, B). Its crown is curved distolabially, with an originally oval cross section, whose long axis is directed linguolabially. A large honing facet on the distal face of the tooth, produced by the occlusion with the upper canine, extends from the tip to almost the base of the crown and is flat linguolabially, but concave dorsoventrally, particularly close to the crown base. Mesially, the canine exhibits a second, subtle facet, owing to the contact with I2. The cheek teeth are too worn to permit detailed morphological observations. The left P3 has a subtriangular shape in lateral aspect, with a broader distal and a higher mesial part; a weak cingulid is preserved on the lingual side (Fig. 2C). At its mesial part, the P3 shows a long, rostrally sloping, sharp median ridge, which forms the lingual margin of a steeply inclined, elongated honing facet, created by occlusion with the upper canine. This facet extends on the entire mesiolabial side of P3, from the crown tip to the deepest part of the mesial root. Both P4s are substantially worn; however, the left P4 is slightly less worn and shows the metaconid that is higher and stronger than the protoconid. The M1 and the longer and broader M2 are deeply worn, and the dentine is confluent on the occlusal surfaces; only their rectangular shape and the presence of two lophids (bilophodont pattern) separated by a buccal cleft is visible. The M2 is rotated ∼15° relative to the M1, and the succeeding M3 8° relative to the M2. The M3 is the longest molar, bearing a third lophid, formed by the hypoconulid, clearly separated from the second one by a distal buccal cleft (shorter than the median one); in occlusal aspect, the tooth is characterized by distal tapering and is less worn than the more anterior molars. The second and the third lophids are connected on the buccal side of the occlusal surface. As indicated by the shape of the left M3 alveolus, the two M3s would have been almost parallel to each other.

Ontogenetic age, sex, and body mass estimation

MAR-1-9B belongs to a senile individual within the dental age stage (IDAS) 5 (loss of inner profile in both M1 and M2, much worn M3) of Anders et al. (2011). This stage corresponds to the ‘old’ age classes of Paul et al. (1993) and Berghänel et al. (2011) and thus to an approximate age of >19–20 years.

MAR-1-9B clearly plots with males of both extant and extinct subspecies in both the Box and Whisker plot of lower canine MD and BL values, and the MD vs. BL biplot (Fig. 3), where a clear separation between males and females is evident. The dimensions of the premolars also distinguish between males and females in extant M. s. sylvanus, although some overlap exists (Fig. 5). The BL of the MAR-1 P3 exceeds that of M. s. sylvanus males; however, its MD is rather small, possibly due to the advanced stage of wear. Similarly, the MAR-1-9B P4 is wide and short, but overall it plots with the larger specimens of all subspecies (Fig. 5). We therefore consider that MAR-1-9B represents a male individual.

Figure 3

Box and Whisker plot (A) and biplot (B) comparing the MD (length) and BL (breadth) of lower canine of modern and fossil Macaca sylvanus subspecies and M. majori from various localities (see Materials and methods).

Figure 5

Bivariate dental plots of buccolingual (BL; BLm, mesial lophid for the molars) breadth vs. mesiodistal (MD) length (in mm) for the lower premolars and molars of the macaque mandible from Marathousa 1, compared with those of modern and fossil Macaca sylvanus subspecies and Macaca majori from various localities (see Materials and methods). A) P3; B) P4; C) M1; D) M2; E) M3.

To estimate body mass, we applied the allometric regression of Delson et al. (2000) for male cercopithecines to the MAR-1-9B M1 MD (7.7 mm). The resulting body mass estimate was 10.5 kg. This is at the lower range of body size values for males reported by Delson et al. (2000: Appendix Table 1) and Fooden (2007: Table 2); however, it may be an underestimate, as the M1 dimensions are likely affected by the advanced degree of wear. If the M2 or the M3 MD dimensions are used instead (9.4 and 12.7 mm, respectively), the resulting body mass estimate is 12.6 and 15.3 kg, respectively. Calculating the mean value from all molar positions produces a body mass estimate of 12.8 kg for MAR-1-9B.

Comparisons

Two genera of fossil Macacina are known from Europe, Paradolichopithecus and Macaca. A taxonomic attribution of MAR-1-9B to Paradolichopithecus (recorded also in Greece during the Early Pleistocene; de Vos et al., 2002; Kostopoulos et al., 2018) can be excluded based on the more rostrocaudally elongated corpus and the more caudally inclined ramus (e.g., Szalay and Delson, 1979: Fig. 185), and the larger dental size of the latter (Paradolichopithecus arvernensis including both males and females: M1 MD = 9.9–11.8 mm; M2 MD = 12.4–15.5 mm; PRIMO fide Alba et al., 2019). On the other hand, mandibular and dental morphology, and dental size and proportions of the MAR-1-9B fit well within the range of extinct and extant M. sylvanus (Figure 3, Figure 4, Figure 5, Figure 6, Figure 7; Table 4).

Figure 4

The Macaca sylvanus cf. pliocena (EPS-MAR-1-9B) specimen from Marathousa 1 compared with specimens of Macaca sylvanus florentina and M. s. pliocena from various localities. A, B) EPS-MAR-1-9B, in left lateral (A), and dorsal (B) views. C, D) Holotype mandible of M. s. florentina from Upper Valdarno, Italy (IGF10034; Museo di Storia Naturale, Florence, Italy), in left lateral (C), and dorsal (D) views. E) Right hemimandible fragment (reversed) of M. s. pliocena from Zlaty Kun C718, Czech Republic (NMP-uncatalogued; Národní Muzeum, Prague, Czechia), in right lateral view. F–I) Occlusal views of the right tooth rows EPS-MAR-1-9B from Marathousa 1 (F), IGF10034 from Upper Valdarno (G), NMP-uncatalogued from Zlaty Kun C718 (H), and NMP-uncatalogued of M. s. pliocena from Gombasek, Slovakia (I). Scale = 5 cm. Photographs other than EPS-MAR-1-9B courtesy of and copyright Eric Delson.

Figure 6

Box and Whisker plots of buccolingual (BL; BLm, mesial lophid for the molars) breadth vs. mesiodistal (MD) length (in mm) for the lower premolars and molars of extant and fossil macaques compared with those of the macaque mandible from Marathousa 1: A) P3; B) P4; C) M1; D) M2; E) M3. Horizontal lines represent the median, boxes the 25 and 75 percentiles (interquartile range); whiskers the maximum-minimum values. Comparative sample: Macaca sylvanus sylvanus (n = 17–27), Macaca prisca (n = 1–2), Macaca sylvanus florentina (n = 11–25), Macaca sylvanus pliocena (n = 7–15), and Macaca majori (n = 5–15) from various localities (see Materials and methods).

Figure 7

Logarithmic ratio diagram comparing the buccolingual (BL; BLm, mesial lophid for the molars) breadth and mesiodistal (MD) length of the lower premolars and molars of the Marathousa 1 macaque mandible with Macaca sylvanus sylvanus, Macaca sylvanus florentina, Macaca sylvanus pliocena, and Macaca majori from various localities (see Materials and methods). Standard of comparison: mean values of M. s. sylvanus.

A comparison of the MAR-1-9B mandible with those of the European M. sylvanus subspecies is limited because of the fragmentary nature of the known specimens. Similar to MAR-1-9B, a mandibular corpus fossa on the buccal side (for this character, refer to the study by Delson, 1980) seems to be present in the mandible of unidentifiable sex from Quecchia quarry (Italy; Bona et al., 2016: Fig. 3a), attributed to M. s. cf. pliocena by Alba et al. (2019), as well as a deeper one in the female mandible of the same taxon from Lezetxiki II (Spain; Castaños et al., 2011: Fig. 2b). Likewise, a fossa is markedly present in the holotype of M. s. florentina (IGF 10034) of a male individual from Upper Valdarno (Italy; Fig. 4C); however, a fossa is absent in the male mandible NMB VA 352 also from Upper Valdarno (Delson, 1980: Fig. 2-2C). Compared with the mandible of M. s. pliocena from Zlaty Kun C718 (Czechia), of unidentifiable sex, MAR-1-9B shows a less inclined rostral margin of the ramus (Fig. 4E).

Owing to the advanced stage of dental wear, a morphological comparison of the MAR-1-9B dentition is not possible; nonetheless, the teeth (in particular the less worn M3) do not show any substantial morphological difference from those of M. s. florentina and M. s. pliocena (e.g., those from Upper Valdarno, Zlaty Kun C718 and Gombasek; Fig. 4F–I). Therefore, we focus on dental crown dimensions. Because of the advanced wear, dental measurements are likely slight underestimates of the original, unworn, crown dimensions. We consider that the M3 measurements are probably not affected to the same degree because it exhibits a lower degree of attrition than the preceding dentition.

With respect to the lower canine, other than a clear separation of M. majori from the larger canines of M. sylvanus, the canine dimensions overlap among the Macaca subspecies; only the male canines of M. s. florentina tend to be larger, in particular in MD (Fig. 3). However, the sample of M. s. prisca and M. s. pliocena is too small to allow safe conclusions. Except for M. majori, the MAR-1-9B P3 and P4 MD values fall below the mean and median values of all M. sylvanus subspecies. Nonetheless, apart from the roughly similar P3 BL mean and median values with M. s. florentina, the MAR-1-9B P3 and P4 BL mean and median values surpass those of M. majori, M. s. sylvanus, M. s. prisca, M. s. florentina, and M. s. pliocena, from the latter exceeding its known maximum range for the P4 BL (Figure 6, Figure 7).

In contrast to the lower canine and the P3–P4, sexual dimorphism in the molars is less pronounced and there exists substantial overlap between males and females (Fig. 5). For all molar positions, the MAR-1-9B crown dimensions are greater than the MD and BL mean and median values of M. majori and M. s. prisca—although the latter is represented by very few known specimens (Figure 6, Figure 7). As noted by Alba et al. (2011, 2019), the M1 of M. s. florentina and M. s. pliocena overlap in size but tend to be longer and wider than those of M. s. sylvanus (Figure 6, Figure 7). The MD dimensions of the two MAR-1-9B M1s fall at the average of M. s. sylvanus and close to the lowest values of M. s. florentina and M. s. pliocena (Figure 6, Figure 7; Table 4). However, their BL values exceed the upper quartile of M. s. sylvanus, stand at the upper limit of the interquartile range of M. s. florentina, and fit best with M. s. pliocena (Fig. 6; Table 4). It must be noted here again that the MAR-1-9B M1s are affected by severe attrition, and therefore, the reported crown dimensions are underestimates of their original, unworn values. The MAR-1-9B M2s plot in the lower range of MD and BL values of M. s. florentina and M. s. pliocena. The M2s of M. s. pliocena tend to be wider (greater median BL value) but still overlap with M. s. florentina and M. s. sylvanus (Fig. 6). In the M3, M. s. pliocena tends to be larger than M. s. sylvanus (although with some overlap), in particular as far as the MD is concerned (Figure 6, Figure 7). In agreement with Alba et al. (2011), Macaca s. florentina tends to be also longer than M. s. sylvanus, although their median BLs are similar. The MAR-1-9B M3, the tooth position least affected by the specimen's heavy wear, falls near the maximum MD values of M. s. sylvanus and between the median MD values of M. s. florentina and M. s. pliocena. In terms of BL, it stands near the mean and median values of M. s. sylvanus and M. s. florentina (Figure 6, Figure 7).

Similar results were obtained by the statistical comparison using z-scores of MAR-1-9B with the subspecies of M. sylvanus (excluding M. s. prisca whose sample size is insufficient). The analysis does not detect significant differences of MAR-1-9B from the comparative sample except of the P4 BL of M. s. sylvanus and M. s. pliocena (Table 4). MAR-1-9B is most similar with the M2 MD of M. s. sylvanus, P3 BL and M3 BL of M. s. florentina, and M1 BL and M3 MD of M. s. pliocena.

We carried out a PCA using only complete M1–M3 tooth rows of M. s. sylvanus, M. s. florentina and M. s. pliocena (34 specimens in total), combining M1 MD, M1 BL, M2 MD, M2 BL, M3 MD, and M3 BL (Fig. 8; SOM Tables S1–S4; M. s. prisca is not included in the analysis owing to the absence of complete tooth rows, while the sample of complete tooth rows of M. majori is too limited to be included). The first two principal components (PC1 and PC2) account for 64.0% of the total variance. PC1 is defined by positive loadings of M2 MD, M2 BL, and M3 BL and negative loadings of M1 MD, M1 BL, and M3 MD; PC2 by positive loadings of M1 BL and M2 BL and negative loadings of M1 MD, M2 MD, M3 MD, and M3 BL. Results show that the convex hulls of the two Pleistocene subspecies, despite some overlap, are relatively well separated from that of M. s. sylvanus, influenced by the long and broad M1 (M1 MD, M1 BL) and the long M3 (M3 MD) of the fossil subspecies and the short and narrow M2 (M2 MD, M2 BL) and wide M3 (M3 BL) of the extant subspecies. The convex hulls of the two fossil taxa overlap greatly, highlighting the strong similarities in the dental dimensions of M. s. florentina and M. s. pliocena. MAR-1-9B (the right M1–M3 tooth row) is distinct from the extant Barbary macaque and plots comfortably within the convex hull of M. s. pliocena. In the PCA plot, it falls closest to the as yet unpublished M. s. pliocena tooth row from the late Middle Pleistocene of Gajtan (Albania; PRIMO; Shearer and Delson, 2012).

Figure 8

Principal component analysis (PCA) of dental variables among Macaca sylvanus sylvanus (n = 17), Macaca sylvanus florentina (n = 9) and Macaca sylvanus pliocena (n = 8) using Mosimann's Log-shape ratio transformation (see Materials and methods and SOM Tables S1–S4 for details). Only complete M1–M3 tooth rows are included. The principal component biplot depicting the projection of the original axes (variables) and the principal component loadings for the first two principal components (PC1 and PC2) are also shown.

Discussion

Taxonomic attribution

Our dental metric comparisons show that MAR-1-9B is distinct from the smaller-sized M. majori and that it can be attributed to M. sylvanus, the sole Plio-Pleistocene representative of macaques in continental Europe (Elton and O'Regan, 2014). With the exception of the smaller-toothed Pliocene M. s. prisca (whose sample, however, is too limited to draw secure conclusions), the MAR-1-9B dental dimensions fall in the ranges of all M. sylvanus comparative samples included here. However, concerning the molars, its values are closer to the mean and median values of M1 BL and M3 MD of M. s. florentina and M. s. pliocena rather than those of M. s. sylvanus. The PCA provides additional evidence that these two fossil subspecies differ from the modern representative in their longer and broader M1, and longer M3. Although M. s. florentina and M. s. pliocena overlap greatly, the MAR-1-9B macaque fits better with M. s. pliocena and does not fall within either the M. s. florentina or the extant M. s. sylvanus range. Nevertheless, taking into account the absence of clear-cut diagnostic criteria for the European fossil macaque subspecies, their great size overlap, as well as the single specimen status and heavy dental attrition of MAR-1-9B, we believe that this result is strongly suggestive but not conclusive. We therefore distinguish MAR-1-9B from the extant M. s. sylvanus, and we provisionally attribute it to M. sylvanus cf. pliocena (for the use of open nomenclature, refer to the study by Bengtson, 1988), an attribution that is also consistent with the specimen's chronology.

Paleoenvironmental setting—Prey-predator relationships among hominins, carnivores, and macaques

Our previous work at Marathousa 1 permits the detailed reconstruction of the paleoenvironmental conditions under which the MAR-1-9B macaque lived. The sedimentological study indicated that the (spatially and stratigraphically associated) fossil and cultural material was deposited on an extensive mudflat surrounding a lake shore, at the end of the glacial MIS 12 and the gradual transition to the interglacial MIS 11 (Karkanas et al., 2018), with a temperate prevailing climate, as reconstructed by paleobotanical remains. Plants were diverse and dominated by aquatic, as well as waterside and damp ground, taxa (Field et al., 2018). The recovered wood assemblage was dominated by willow (Salix) and alder (Alnus), but also included elm (Ulmus), deciduous oak (Quercus), and maple (Acer; Field et al., 2018). Temperate conditions are also inferred from the diverse bird (particularly rich in anatids and rallids) and large mammal (e.g., beavers, otters, deer, hippos, elephants) assemblages, which point to a landscape with substantial woodland components as well as more open areas, close to permanent and large freshwater bodies (Konidaris et al., 2018; Michailidis et al., 2018). Furthermore, a multiproxy analysis using ostracods, sponge spicules, diatoms, grain sizes, total organic carbon, total inorganic carbon, and conventional X-ray fluorescence, indicated a mostly dry and rather cold climate, with summer water temperatures between 10 and 15 °C (Bludau et al., 2021). These transitional temperate conditions and temperature range are suitable for M. sylvanus according to Elton and O'Regan (2014). Additionally, the Marathousa 1 landscape would have offered macaques the protection and the plant resources of the nearby woodlands, as well as access to freshwater and a variety of feeding opportunities throughout the year. Indeed, modern M. sylvanus is omnivorous, with diets encompassing mainly leaves, seeds, fruit, fungi, and invertebrates (e.g., snails, scorpions, spiders, beetles, millipedes, ants), but it is also known to consume bird eggs and hunt vertebrate prey such as birds, squirrels, and rabbits (Martin, 2003; Fooden, 2007; Young et al., 2012).

Middle Pleistocene hominins, as well as large carnivores, would also have benefited from the lake shore environment, which would allow the employment of ambush strategies and other hunting techniques (Konidaris and Tourloukis, 2021). Overall, the Marathousa 1 paleoecosystem would have been characterized by a multicomplex ecological network with intense predator-prey relationships among hominins, carnivores, and herbivores, including macaques. Both hominins and carnivores would have been potential predators for macaques (Meloro and Elton, 2012), especially because M. sylvanus is a diurnal and predominantly terrestrial species when moving and feeding (Martin, 2003; Fooden, 2007). The advanced age of the MAR-1-9B individual, combined with the possible pathologies, suggest that it was vulnerable and an easy target because intrinsic power, that is, physical strength and fighting ability of the individual, reduces with age (Berghänel et al., 2011). Although it does not show any evidence of anthropogenic or carnivore modifications, it was found in close spatial association with lithic artifacts as well as other faunal material bearing cut marks or carnivore gnawing (Fig. 1C). Indeed, prey-predator relationships and interactions among hominins, carnivores, and macaques at Marathousa 1 are likely: lithic artifacts and bones with anthropogenic modifications (cut marks, percussion damage, bone tools) recovered at the site are spatially and stratigraphically associated with carnivore remains (e.g., the wildcat Felis sp., the fox Vulpes sp., and the wolf-sized Canis sp.) and carnivore-modified bones of their prey (Konidaris et al., 2018; Tourloukis et al., 2018a). Spatial taphonomic analysis of the site has shown that these assemblages have undergone minimal transport, as also attested by several refits of broken faunal remains, including the MAR-1-9B mandible (Giusti et al., 2018; Konidaris et al., 2018). The integrity of the assemblage and the close association of bones showing anthropogenic and carnivore modifications indicate that both hominins and carnivores exploited a broad spectrum of prey, from medium-sized mammals (the fallow deer Dama) to megafauna (the straight-tusked elephant Palaeoloxodon), and thus contributed to the formation of the site. Among the carnivore species that have been found at Marathousa 1 (Konidaris et al., 2018; Tourloukis et al., 2018a), Canis is known to interact with and occasionally hunt M. sylvanus even today (Riley et al., 2015; Waters et al., 2017), and this was likely also the case with the pursuit- and pack-hunting Pleistocene large canids (Konidaris and Tourloukis, 2021). Although Felis and Vulpes may have preyed on juvenile macaques, they would have been unlikely predators of adults (Meloro and Elton, 2012). Additional large carnivores have been reported from the Pleistocene of the Megalopolis Basin and would have been members of the greater ecosystem at Marathousa 1, including hyenas (Hyaenidae indet.), lions (Panthera leo), and perhaps leopards (Panthera pardus; Sickenberg, 1976; Athanassiou, 2018), all of which are documented to prey on primates (Isbell, 1994; Hart, 2007).

Interactions and conflicts between macaques and humans are also widely reported today (Priston and McLennan, 2013). Although scarce in the fossil record, direct evidence of Macaca exploitation by hominins exists from South-Southeast Asia (e.g., Rabett and Piper, 2012; Wedage et al., 2019), while Blasco and Fernández Peris (2012: Table 5) noted anthropogenic breakage on a M. sylvanus specimen from layer XII of Bolomor Cave (Spain). In Europe, several Middle Pleistocene localities document the coexistence of macaques and hominins (Szalay and Delson, 1979; Elton and O'Regan, 2014; Alba et al., 2019), including a few which, such as Marathousa 1, have also yielded evidence of elephant butchering, for example, Fontana Ranuccio and La Polledrara in Italy, Bilzingsleben in Germany, Ebbsfleet in the United Kingdom, and Ambrona and Bolomor Cave in Spain (Blasco and Fernández Peris, 2021; Konidaris and Tourloukis, 2021). The MAR-1-9B specimen marks the first documented such record in southeast Europe.

Conclusions

Dental morphology and dimensions of the Marathousa 1 specimen permit its attribution to the Barbary macaque M. sylvanus. Although clear-cut diagnostic criteria for the fossil subspecies are currently lacking, the multivariate (PCA) analysis of molar crown dimensions supports an attribution of MAR-1-9B to M. s. cf. pliocena, in agreement with its chronology but pending future discoveries and the clarification of the differential diagnoses among the fossil macaque subspecies.

MAR-1-9B represents the most complete fossil European macaque mandible. It adds to the meager fossil record of the genus in the eastern sector of the peri-Mediterranean region (Mecozzi et al., 2021: Fig. 4) and is the first recorded occurrence of Macaca in the Middle Pleistocene of Greece. It therefore extends the chronological range of cercopithecids and documents, for the first time, the coexistence of macaques and hominins in the country. Finally, the exquisitely preserved, rich and well-documented assemblages of Marathousa 1, allow not only a secure chronological placement of this specimen but also a detailed reconstruction of the environmental conditions and predator-prey relationships under which it lived.

Conflict of interest

There is no conflict of interest.