The site, Cauca (11°18'51"N, 70°17'41"W), was so named because of its proximity to the homonymous fishing village, in the coastal area of the Gulf of Venezuela. It is located approximately 14.6 km northwest of the Town of Urumaco, Urumaco Municipality, following the dirt road that connects Urumaco with the hamlets of Cauca and Río Seco (Fig. 1). The site is on the western coastal plain about 2.3 km south of the coastal zone and approximately 12 metres above sea level. Currently, this plain is influenced by the trade winds (“vientos alisios”), whose incidence generates a semi-arid or arid, seasonal climate with prolonged periods of water deficit, with rainfall that ranges between 200 and 600 mm annually, average temperatures of 30 °C and a predominantly xerophytic vegetation (Matteucci et al. 1999).

The area is affected by an intermittent runoff system, generating laminar erosion of the sediments. The geology is represented by a not formally defined sedimentary unit, characterised mainly by facies of unconsolidated fine to coarse sands of light brown to ochre colour, with underlying coastal palaeodunes. The level of oxidation of the sand and clay layers carrying the bone assemblages suggests that Cauca was probably deposited in a low-energy more humid environment, contrasting with the environmental conditions currently present in the area. For now, there is no precise evidence of whether this deposition environment was a permanent or intermittent body of water. The coastal plains of Falcón State were subject to a negative water balance during the Late Pleistocene (Ochsenius 1980) and the Cauca site could have offered attractive water resources for animals with swamp patch areas. An example of archaeological sites, such as Muaco, Taima-Taima and Cucuruchú, which were deposited in sedimentary environments influenced by the action of resurgent springs, offer an oasis that may have attracted animals during dry periods (Cruxent 1970; Ochsenius and Gruhn 1979; Ochsenius 1980).

The fossiliferous site covers an area of approximately 1600 m², where eight groupings of bone remains have been identified emerging on the surface (Figs 1B, 3). The first reports from this locality were made by one of the authors (R.S.) from the Palaeontology Department of the Alcaldía Bolivariana del Municipio Urumaco, in January 2004. During a survey of the site on 16 November 2019, several of the authors (J.D.C.B., A.J. and R.S.) collected from surface partial lithic projectiles and other preformatted artefacts in the vicinity of the Cauca site (see Jaimes et al. (2024a)). Thanks to the support and legal authorisation from the Alcaldía Bolivariana del Municipio Urumaco and the Instituto del Patrimonio Cultural de Venezuela (permits N° 00019-01/31/2020, 00110-04/09/2021, 00178-07/14/2021 and 00522-08/18/2022) the first systematic excavations were planned at the Cauca site.

Figure 3. 

Remains of fossil vertebrates emerging on the surface at the Cauca site. Skull (A), pelvis (B) and appendicular elements (C) of megatheriid Eremotherium. A, C. correspond to unexcavated S2 (see Fig. 1B). Upper molars (D) of gomphoteriid Notiomastodon platensis, with the root completely eroded, associated with the mandible AMU-CURS-1269 from Cauca “A” (Fig. 1B).

To date, three excavation field seasons have been carried out between 2021 and 2022, totalling five work areas. These excavations were under the direction of A.J. and R.S., excavating only the outcropping bone groupings defined as 1, 3–5 and 8 (see Fig. 1B). The five excavations include: 1) Cauca “A”, carried out between 3 and 4 August 2021, with a 4 × 4 m grid (Fig. 4B); 2) Cauca “B”, carried out between 4 and 26 October 2021, with a 4 × 4 m grid (Fig. 4B), and 3) Cauca “C”, “D” and “E” carried out between 3 and 20 November 2022, and all with 2 × 2 m grids (Figs 1B, 4B).

Figure 4. 

Excavations at the Cauca site. Cauca “A” (A), Cauca “B” (B), and Cauca “E” (C), showing the defence of Notiomastodon platensis (AMU-CURS-1359) in situ. Mandible of N. platensis in plaster jacket in Cauca “A” (D). Cauca “D” with remains (tibia and ribs) of cf. Eremotherium laurillardi (E).

The depth of all excavations did not exceed 40 cm because that is the thickness of the bearing layer (e.g. Fig. 4B, C). There is no evidence of fossil under/below this layer. The stratigraphy of the site, based on the five excavations and from the top or surface to the excavated base, is characterised by at least three well-defined layers. The first is the surface layer, about 5 cm thick and composed of unconsolidated sediments that include pellets and other small clasts that are transported by rain runoff and wind action that occurs from east to west. The second layer varies between 20 and 35 cm thick, composed of fine silt-clayey sands, heterogeneous, compact, and well-defined with colours between light brown and ochre, with evidence of oxidation. In Cauca “D”, specifically grid “C”, the sediments showed a greater clay composition, which could be related to its accumulation in the lowest deposition area in the northern section of the site. The deposit of this second layer may have occurred in a water-saturated, or intermittent, very low-energy environment, favoured by the low inclination of the terrain (~ 10°). This second layer is the carrier of the bone remains and it could have had a greater thickness, which was likely eroded by laminar erosion, explaining the large amount of fossil bones exposed on the surface. The third layer, or the basal part, is characterised by an ancient relief of slightly more consolidated dark and oxidised sands, in which no remains of vertebrates or other micro/macrofossil organisms have been found.

Most of the bone and dental remains identified on the surface in different surveys (pre-excavation) were practically disintegrated or in a state of fragmentation that did not allow any type of surface rescue (Fig. 3A–C). Possibly, this was because of its exposure to environmental factors for decades before the first survey at the site. In pre-excavation activities, only the relatively complete crown of an M2 molar from a gomphothere and fragments of equid teeth were recovered on the surface by the authors (J.D.C.B., R.S.) in 2013.

Two dental fossil samples were selected for radiocarbon dating (Table 1). The dated material comes from a Notiomastodon platensis mandible (AMU-CURS-1269) found in excavation Cauca “A” and an isolated tooth of Equus sp. (AMU-CURS-1365), found in the vicinity of the latter (Fig. 1B). Both specimens correspond stratigraphically to the most superficial or exposed part of what we recognise here as the second layer.

In the absence of bone collagen, tooth enamel was selected from these two fossils. The enamel surface was cleaned and the dentine was removed with a dremel to isolate the enamel from the rest of the dental tissue. The enamel (approx. 2 g) was then ground using a steel mortar and pestle, followed by grinding in an agate mortar to a particle size of < 100 microns. The powder was then further ground using a McCrone Microniser Retsch following the methodology described in Wood et al. (2016). Briefly, this approach enables better separation of crystallite clusters and increases the efficiency of the acetic acid pre-treatment aimed at removing diagenetic carbonates. To achieve this, the powder was ground three times for 10 min, with a 5 min pause between each stage to avoid heating. For some samples, part of the hand-milled fraction (< 100 microns) was retained for comparison with the conventional pre-treatment approach. The resulting powder was pre-treated under light vacuum for 20 h with a solution of 1 N acetic acid (1 ml acetic acid for approx. 50 mg powder). The extraction yields were comprised between 70–80%. The pre-treated powder was then rinsed with milliQ water and dried at 50 °C in an oven. About 250 mg of powder was then reacted under vacuum with orthophosphoric acid at 70 °C for around 20 min. The CO₂ released was then separated cryogenically from the water produced and chemically purified, thanks to several passages through a trap filled with a copper-silver wool mixture. The CO₂ gas was then reduced in the presence of hydrogen and iron to produce graphite. Samples were then pressed into targets and ¹⁴C ages were measured on the compact AMS ECHoMicadas (Archéozoologie, Archéobotanique, Sociétés, Pratiques et Environnements, Equipe SAPOA, France).

Five mammal taxa are reported here, including three xenarthrans, one proboscidean, an equid and a felid. Other fragmentary and no diagnostic elements are referred to here as indeterminate mammals.

Xenarthra Cope, 1889

Phyllophaga Owen, 1842

†Megatheriidae Gray, 1821

†Eremotherium Spillmann, 1948

†cf. Eremotherium laurillardi (Lund, 1842)

Fig. 5

Referred material

Remains of Eremotherium cf. E. laurillardi have been recovered at the Cauca site in bone groupings 2–6, with only groupings 3–5 being excavated (Fig. 1B). Fourteen cranial and postcranial remains that include a right hemi-mandible (with mf 1 to 4), skull (with right Mf 1 to 5 and left Mf 2 to 4), right clavicle, right radius, presumably left navicular, cuneiform, IV left metatarsal, a worn element that we presume corresponds to a scaphoid, fragment of pelvis with acetabulum and part of the pubic ramus and five incomplete vertebrae, from excavation Cauca “B” (Fig. 4B) and assigned to catalogue number AMU-CURS-1268. A probable right tibia (AMU-CURS-sn) and seven rib fragments were found in Cauca “D”. The other cranial and postcranial elements identified in Cauca “C” and unexcavated groupings 2 and 6 were not collected due to their poor state of preservation, being in some cases disintegrated (Fig. 3A–C).

Descriptions

The 14 disarticulated cranial and postcranial elements recovered in Cauca “B” most probably belong to the same individual (AMU-CURS-1268). The cranial material is composed of only the skull (AMU-CURS-1268a) (Fig. 5A1–A4) and the right hemi-mandible (AMU-CURS-1268b) (Fig. 5B1, B2). The skull has a Tl of 580 mm and preserves only the palatal section, since part of the cranial vault, nasal bones and part of the maxilla bones were destroyed by erosive processes. The palatal section is relatively well preserved, with both the right Mf1–Mf5 and the left series Mf2–4 and the alveolus of left Mf1 and Mf5. The hemi-mandible has a Tl of 450 mm, preserving the four molariforms mf1–mf4 (Fig. 5B2). This hemi-mandible has a convex ventral margin and projects downwards at the level of the first molars, forming a moderate mandibular protuberance. The molariforms preserved in the skull and hemi-mandible are of quadrangular section with crowns characterised by two transverse and parallel ridges, separated by a “V”-shaped valley. Measurements of the molariforms are presented in Table 2. Associated with these cranial remains were recovered the right clavicle with a Tl of 240 mm (AMU-CURS-1268c; Fig. 5D1, D2), complete right radius with a Tl of 700 mm (AMU-CURS-1268d; Fig. 5C1–C3), presumably left navicular (AMU-CURS-1268e), cuneiform (AMU-CURS-1268f), IV left metatarsal (AMU-CURS-1268g; Fig. 5G), a worn element that we presume corresponds to a scaphoid (AMU-CURS-1268h), fragment of pelvis with the acetabulum and part of the pubic ramus (AMU-CURS-1268i) and five incomplete vertebrae (AMU-CURS-1268j–n; Fig. 5E1–F). Other small, fragmented and indeterminate bone elements were observed during the excavation and presumably could be associated with the same individual.

Table 2.

Download as

CSV

XLSX

Dental measurements of cf. Eremotherium laurillardi (AMU-CURS-1269a, b) collected at the Cauca site. Abbreviations: upper molariform (Mf), lower molariform (mf), right (d), left (i). (*): incomplete or missing pieces. Measures in millimeters.

	Mfr1	Mfr2	Mfr3	Mfr4	Mfr5*	Mfl1	Mfl2	Mfl3	Mfl4	Mfl5*	mfr1	mfr2	mfr3	mfr4
Long	25	35	35	35	20	25	35	35	37	20	40	40	40	30
Width	36	50	47	40	25	35	45	47	40	15	44	45	42	40
Height	40	32	22	10	5	10	10	27	20	–	40	30	10	10

Figure 5. 

Megatheres (A1–G) and mylodontids (H1, H2) from the Cauca site. A1–A4. Skull of cf. Eremotherium laurillardi (AMU-CURS-1268a) in right lateral, ventral, posterior and anteroventral view; B1, B2. Right hemi-mandible (AMU-CURS-1268b) in right lateral view; C1–C3. Right radius (AMU-CURS-1268d) in proximal, dorsal and distal view; D1, D2. Right clavicle (AMU-CURS-1268c); E1, E2. Thoracic vertebra (AMU-CURS-1268j) in anterior and left lateral view; F. Thoracic or lumbar vertebral centrum (AMU-CURS-1268k) in anterior view; G. Left IV metatarsal (AMU-CURS-1268g); H1, H2. cf. Mylodontidae, proximal and dorsal right radius (AMU-CURS-1363). Abbreviations: (alv.) alveolus, (fm) foramen magnum, (mf) lower molariform, (Mf) upper molariform, (oc) occipital condyle.

In Cauca “C”, only a few remains of postcranial elements of an Eremotherium, were found emerging on the surface. These were identified as vertebrae, ribs, pelvis fragments and other small indeterminate fragments. However, the poor state of preservation and disintegration in some cases, did not allow their recovery. Due to the proximity of these materials to the individual collected in Cauca “B”, a possible association amongst them is not ruled out here. In Cauca “D”, an association of seven fragments of ribs and a right tibia were recovered, also in a poor state of preservation (Fig. 4E). This bone accumulation is located about 8 m north of the Cauca “B” individual, so a possible association with the latter would not be unlikely.

Bone grouping 2 (Fig. 1B) was not excavated due to the partial or total disintegration in very little scraps of bones of the outcropping materials. However, a skull in a state of fragmentation (Fig. 3A) still offers evidence that allows us to identify this association as a second specimen of Eremotherium. In bone grouping 6 (not excavated) to the northwest of Cauca “D”, a humerus of Eremotherium was also identified, but its poor state of preservation prevented collection or conservation in situ.

Remarks

The ground sloth Eremotherium laurillardi distributed from south-eastern USA to Brazil, is the only species of the genus known from the Late Pleistocene in the Americas (Cartelle and De Iuliis 1995). The taxonomic validity of Eremotherium rusconi Schaub, 1935, another species from the Late Pleistocene of South America, is questionable (Cartelle and De Iuliis 1995; Faure et al. 2014; Cartelle et al. 2015).

Remains referable to Eremotherium are common and widely referred to in the Pleistocene fossil record of Venezuela (e.g. Schaub (1935); Simpson (1939); Nectario Maria (1941); Aguilera (2006); Carrillo-Briceño (2015); Carrillo-Briceño et al. (2016); Chávez-Aponte (2022) and references therein); specific taxonomic assignments of many of these specimens are still uncertain and should be clarified.

Remains assigned to cf. Eremotherium laurillardi have been reported in Venezuelan territory in palaeontological and archaeological sites ranging from sea level (e.g. Muaco, Taima-Taima and Cucuruchú; see Aguilera (2006); Carrillo-Briceño (2015); J.D.C.B pers. obs. (2022)), to mountain areas above 1200 m (e.g. El Vano; Jaimes et al. (2024b)).

The cranial and postcranial materials of Eremotherium collected from Cauca likely belong to an adult individual. The assignment of these specimens to the genus Eremotherium is based on the Hypsodontic index (HI) known for ground sloths. According to Bargo et al. (2006), HI is an element that is quantified by using the depth of the mandible (DM) (below the third molariform tooth) divided by the length of the molariform tooth row (LTR). In the right hemi-mandible collected in Cauca “B” (AMU-CURS-1268b), the DM is 140 mm and the LTR is 180 mm, with a resulting IH equal to 0.77, a value that coincides with Eremotherium whose range oscillates between 0.73 and 0.83 (De Iuliis 1996; Bargo et al. 2006). Based on the taxonomic validity of the Late Pleistocene Eremotherium species suggested by Cartelle and De Iuliis (1995) and Cartelle et al. (2015), we tentatively suggest assigning the specimens collected at the Cauca site as cf. Eremotherium laurillardi.

†Mylodontidae Gill, 1872

† Glyptodontidae indet.

Fig. 6

Referred material

Two isolated osteoderms (AMU-CURS-1047 and -1360) were collected on the surface at the Cauca site (Fig. 1B).

Descriptions

Both isolated osteoderms are from the shell. AMU-CURS-1047 has a Tl of 47 mm and is eroded on both the internal and external faces, which does not allow the identification of the ornamentation pattern (Fig. 6A1, A2). AMU-CURS-1360 possibly corresponds to an osteoderm from the anterior lateral region of the shell and this has a Tl of 31 mm; it is also eroded on the inner face, and partially preserving the ornamentation of the outer face with a central figure of irregular polygonal shape, preserving some foramina that could belong to hair follicles and surrounded by a groove (Fig. 6B1, B2).

Figure 6. 

Indeterminate Glyptodontidae osteoderms from Cauca site. A1, A2. AMU-CURS-1047; B1, B2. AMU-CURS-1360; both in external and transversal view.

Remarks

The state of preservation of these two osteoderms does not allow a more precise taxonomic determination than Glyptodontidae, Glyptodontinae. However, these osteoderms could belong to the genus Glyptotherium, widely known in the Late Pleistocene of the Falcón State, especially at the sites of Muaco and Taima-Taima (Carlini et al. 2008, 2022). A recent report suggests that these Glyptotherium-related armadillos had a broader distribution in what is now Venezuela during the Late Pleistocene, with reports including mountain areas above 1200 m altitude (see Jaimes et al. (2024b)).

Proboscidea Illiger, 1811

†Gomphotheriidae Hay, 1922

† Notiomastodon Cabrera, 1929

† Notiomastodon platensis (Ameghino, 1888)

Fig. 7

Referred material

A mandible (Fig. 7A1), the proximal part of a humerus and the distal end of the radius with what appears to be parts of very deteriorated carpals and metacarpals, all assigned as AMU-CURS-1269 and collected from Cauca “A” (Fig. 4A). An isolated M2 (AMU-CURS-1045) collected about 5 m northwest of Cauca “C” (Fig. 1B) and a complete and isolated tusk (AMU-CURS-1359) (Fig. 7C) from Cauca “E” (Fig. 4C). Other dental and postcranial fragments (see below) in a very poor state of preservation could not be collected.

Figure 7. 

Notiomastodon platensis remains from Cauca site. A1, A5. Dorsal and anterodorsal views of the mandible (AMU-CURS-1269); A2, A3. Occlusal and labial views of preserved right m2 and (A4), occlusal view of m3; B1, B2. Occlusal and lateral view of M2 (AMU-CURS-1045). C. lateral view of tusk (AMU-CURS-1359).

Descriptions

The mandible AMU-CURS-1269 (Fig. 7A1) preserves the complete right m2 (Fig. 7A1–A3, A5) and erupted m3 (Fig. 7A1, A4) and erupted left m2 and m3 in a fragmentary state (Fig. 7A1, A5). Part of the mandibular ramus and condylar processes are not preserved. The right m2 is a bunodont and trilophodont molar with a well-defined heel and cingulum (Fig. 7A2, A3); it is 150 mm in Tl. This molar presents a wear state of “3” (following Mothé et al. (2010)), which suggests that it could correspond to a subadult/adult individual. Abundant fragments referring to upper molars M2–M3 of the same individual were recovered on the mandibular remains (Fig. 3D). A small portion of a tusk was also found a few centimetres from the specimen, which was not collected due to the high degree of disintegration. Of the postcranial remains associated with the lower jaw, the proximal part of a right humerus is only preserved in the posterior section and what appears to be the distal end of the radius with carpal and metacarpal fragments are in a very poor state of preservation.

The isolated tusk AMU-CURS-1359 (Fig. 7C), whose position on the skull is undetermined, measures about 1700 mm long and could belong to an adult. It is curved, although it shows no apparent natural twist and lacks an enamel band. The tusk lies about 14 m north of the Cauca “A” site and its association with the jaw and the other elements is uncertain. Near the collection site of this tusk and on the surface, some fragmented and indeterminate postcranial bones were observed.

A few metres northeast of the Cauca “D” and “E” excavations, a tusk and bone fragments were found emerging on the surface. However, these were not collected due to the poor state of preservation (grouping 7; Fig. 1B). In addition, small fragments of gomphotherid molars were scattered on the surface of the site, possibly because of transport by runoff. A possible upper left M2 (AMU-CURS-1045) not worn and with ~ 96 mm in Tl (Fig. 7B1, B2), was collected a few metres northeast of excavation “C”, suggesting the probable presence of a juvenile individual at the site. On the other hand, we do not have any evidence that allows us to associate the cranial and dental remains found in Cauca “A”, with the other gomphothere elements reported herein, resulting in an uncertain number of adult individuals present at the site.

Remarks

In recent times, consensus has emerged suggesting the validity of only two genera of gomphothere in the southern continent, Cuvieronius hyodon Fischer, 1814 and Notiomastodon platensis (= Stegomastodon platensis) (e.g. Mothé et al. (2012, 2013, 2017a); Alberdi and Prado (2022)). The stratigraphic and geographic distribution, as well as the diagnostic features that have been used to differentiate both genera, which are based on the skull and mandibular symphysis morphology and upper tusk and molars, are discussed in detail by Mothé et al. (2016, 2017a, 2017b), Alberdi and Prado (2022) and Carrillo-Briceño et al. (2023).

Based on the taxonomic differences mentioned above and used to differentiate both genera of gomphotheres in South America, we assigned the better-preserved specimens referred to in this section to N. platensis. This assignment is supported by: 1) the double to single clover wear pattern present in the right m2 of the AMU-CURS-1269 (Fig. 7A2), which resembles the pattern present in molars with advanced wear state in N. platensis (e.g. Mothé et al. (2012, 2013, 2017a) and references therein); 2) the robustness, curved shape and lack of twist and enamel bands in the tusk AMU-CURS-1359 (Fig. 7C), which contrast with the shape and morphology of the C. hyodon tusk (Mothé et al. 2012, 2013; Mothé and Avilla 2015); and 3) a jaw with no trace of lower tusks; meanwhile, C. hyodon presents a pair of lower incisors or its corresponding vestigial alveoli (Mothé and Avilla 2015). By analogy, we believe the other remains of gomphotheres found at the Cauca site (e.g. tooth fragments and probably postcranial elements), and with a poor state of preservation (for example these from grouping 7) could also belong to N. platensis.

Abundant specimens of gomphotheres have been reported for several locations in the Falcón State and other sites in Venezuela (Carrillo-Briceño et al. 2008a, 2008b; Chávez-Aponte et al. 2008; Carrillo-Briceño 2012, 2015); however, their taxonomic assignment should be re-evaluated. The gomphothere remains collected from the palaeontological/archaeological sites of Muaco, Taima-Taima and Cucuruchú in Falcón State, should be assigned to N. platensis (e.g. J.D.C.B., pers. obs. (2022)).

Perissodactyla Owen, 1848

Equidae Gray, 1821

Equus Linnaeus, 1758

 Equus sp. and Equidae indet.

Fig. 8

Referred material. The equid material corresponds to isolated dental elements collected on the surface of the second layer (Fig. 1B) and is represented by a complete m3 (AMU-CURS-1365) and another five molar fragments (AMU-CURS-523: n = 2, and AMU-CURS-1046: n = 3).

Descriptions. The left m3 AMU-CURS-1365 is almost complete, only missing part of the mesial surface (Fig. 8A1). The occlusal surface of the m3 is relatively well preserved with a length of 28 mm and it is characterised by a subtriangular protoconid, an oval hypoconid, a metaconid and metastylid forming a more rounded double knot, an oval postflexid and the presence of a well-defined pli caballinid fold. The rest of the five fragments could belong to both upper and lower molars; however, the poor state of preservation of these prevents a more detailed taxonomic identification.

Figure 8. 

Equids, felids, and indeterminate mammals from Cauca site. A1, A2. Left m3 of Equus sp. (AMU-CURS-1365) in occlusal and labial view; B1–B3. Right astragalus of cf. Leopardus pardalis (AMU-CURS-1361) in dorsal, plantar and lateral view; C. A bone of the foot of an indeterminate mammal (AMU-CURS-1362). Abbreviations: (hypc) hypoconid, (mtcd) metaconid, (mty) metastylid, (plc) pli caballini fold, (prtc) protoconid.

Remarks. The occlusal pattern, present in the m3 with a subtriangular protoconid, an oval hypoconid and the presence of a pli caballinid fold, can likely be associated more with Equus than Hippidion Owen, 1869 (see Prado and Alberdi (2017); Carrillo-Briceño et al. (2023) and references therein). The poor state of preservation of the other dental elements does not allow them to be assigned beyond Equidae for now.

For the Pleistocene of Falcón State, equid remains assigned to Equus neogeous Lund, 1840 and Equus santaeelenae Spillmann, 1938, have been reported for the sites of Muaco, Taima-Taima, Cucuruchú and Quebrada Ocando (Royo and Gómez 1960; Aguilera 2006; Rincón et al. 2006). Other reports of fossil equids from Venezuela have also been referred to by Rincón et al. (2006) and Carrillo-Briceño (2015).

Prado and Alberdi (2017), based on an extensive review of the morphological and morphometric characters of cranial and postcranial elements of fossil horses from South America, recognise only three species as valid (e.g. E. andium, E. insulatus and E. neogeous). Machado and Avilla (2019) questioned the diagnoses and taxonomic validity of the three valid species proposed by Alberdi and Prado (2017), suggesting the possibility of a single Equus species of South America which should be recognised as E. neogeous. In the case of our specimens, only the m3 AMU-CURS-1365 is likely associated here with Equus; despite this, until now, as there is no consensus on the taxonomy of the valid Equus species of South America and the lack of more diagnostic characters in our specimen, it is appropriate to keep this assignment tentatively as Equus sp

Carnivora Bowdich, 1821

Felidae Fischer, 1817

Felinae Fischer, 1817

Radiocarbon age of 33750 ± 350 and 37050 ± 500 BP were obtained from teeth of Notiomastodon platensis and Equus sp., respectively. Results were then calibrated using OxCal 4.4 (Bronk Ramsey 2009) and indicate that the Notiomastodon is 37.5–39.5 ka cal. BP and that the Equus is 41.3–42.3 ka cal. BP. These dates are amongst the oldest dates ever produced on bioapatite (see Cherkinsky (2009)) and are close to the limit of the radiocarbon dating technique. It is important to note here that these are minimum ages because diagenesis makes bioapatite samples look younger due to carbon isotope exchange between soil dissolved carbonates and bone carbonate (Zazzo and Saliège 2011; Zazzo 2014). It is, thus, very likely that these samples are older than 40 ka.

Preformatted lithic artefacts, on the surface and without evidence of discrete accumulations, were in the adjacent areas of the Cauca “B–D” excavations (Fig. 1B). None of these artefacts was recovered in direct association with the bone materials in the fossil-bearing stratum or layer. These artefacts are represented by a chopper (AMU-12-1; Fig. 9A–B2), a planoconvex scraper (AMU-12; Fig. 9C–D3) and three flakes or lithic debris (AMU-12-2; Fig. 9E).

Figure 9. 

Lithic artefacts found in the vicinity of the excavations at the Cauca site (see Fig. 1B). A, C. in situ artefacts. Chopper (AMU-12-1; B1, B2), planoconvex scraper (AMU-12; D1–D3) and flake (AMU-12-2; E).

Between 100 and 150 m northeast of the Cauca excavation site, we have found a relatively flat area with the presence of abundant preformatted lithic artefacts on the surface. The carrier layer corresponds to unconsolidated fine to coarse sand facies of light brown and ochre colour that underlie palaeodunes. This area and its surroundings have been prospected by members of our team since 2019 and, in it, we have collected at least three different typologies of lithic projectiles that include El Jobo, Clovis and Fish Tail technologies, amongst other lithic artefacts (Jaimes et al. 2024a). However, so far, we do not have precise evidence that allows us to put this diversity of lithic technologies around the Cauca site in a chronological and stratigraphic context or in any relationship with the megafauna reported for the place.

According to Jaimes et al. (2024a), the contemporaneity of these three projectile typologies in the surroundings of Cauca in such a small area remains an uncertainty and their presence there could probably be due to erosion/laminar washing of the most overlying layers which may have generated decanting and mixing of artefacts corresponding to different antiquities. There is no doubt that the site of Cauca and its surroundings have been subjected to these processes of laminar erosion and this could be inferred by the flat characteristics of the terrain and the almost null evidence of transportation shown by both the fossils and the lithic artefacts (Jaimes et al. 2024a).

The bioapatite dating of two fossil remains from Cauca provided an age of at least 40,000 cal. BP. In contrast, a biochronological approach of the site using identified fossil taxa does not offer a more precise age due to a wide chronological record for these taxa. For example: 1) Late Pleistocene for Eremotherium laurillardi (Cartelle and De Iuliis 1995), 2) Pliocene (~ 3.8 Ma, only for North America)–Late Pleistocene (for South America) for Glyptotherium (Carlini and Zurita 2010; Gillette et al. 2016; Zurita et al. 2018), 3) Early Pleistocene (1.2–0.4 Ma)–Late Pleistocene (~ 11,770 years BP) for N. platensis (Alberdi and Prado 2022) and 4) Middle Pleistocene (~ 2 Ma)–Early Holocene (~ 8,000 years BP) for Equus (MacFadden 2013; Prado and Alberdi 2017; Villavicencio et al. 2019).

Fossil bone elements emerging on the surface at the Cauca site were categorised within conservation stage “5” on the scale of Behrensmeyer (1978) (i.e. as the bone is fragile and breaks easily, it may lose the original shape of the bone). These materials could not be collected or identified in most cases due to their degree of fragmentation or total or partial disintegration (e.g. Fig. 3A–C). In contrast, the elements of the excavated groupings (not exposed) present a conservation stage of “3 to 4”, on the scale of Behrensmeyer (1978). These specimens have a thick fibrous surface, rough texture, large and small splinters that tend to break off when the bone is moved. Weathering has penetrated the internal cavities. The cracks are open and chipped with rounded edges and, in some cases, there is an incursion of sediment and biological activity (plant roots). These taphonomic preservation conditions limit the identification of micro modifications of animal or anthropogenic origin on the bone surface. Some specimens recovered from excavations are fractured or incomplete, where the missing part was exposed on the surface and eroded.

The factors that influenced the arrangement, alteration and disarticulation of the bones could be related to natural processes (meteoric and animal action) or human action (Fernández-Jalvo and Andrews 2016). At the Cauca site, indirect evidence of disarticulation was observed. An example is the postcranial remains of a probable adult individual assigned as cf. E. laurillardi and identified in excavation groupings 3 (Cauca “B”), 4 (Cauca “C”) and 5 (Cauca “D”) (Fig. 1B). In Cauca “B”, only the skull, right hemi-mandible, and a few other postcranial elements of the individual (see the faunal assemblage section) were identified. Other remains that probably belong to the same individual from Cauca “B” are found scattered between Cauca “C” and “D”. In Cauca “C”, a badly deteriorated fragment of the pelvis and other fragments of ribs and vertebrae were identified less than six metres west of Cauca “B”. In Cauca “D”, about 12 m north of Cauca “B”, a grouping of fragments of disarticulated ribs and a right tibia were recovered (Fig. 4E). These ribs show high fragmentation, with an arrangement pattern where some are on top of others and very close to the tibia.

At the Cauca site, large, compact and dense postcranial elements such as humeri and femurs of large megaherbivores (e.g. Eremotherium or mylodontids, Notiomastodon), both on the surface and in excavated groupings 1–5 (Cauca “A–D”), are scarce. Some examples are the tibia reported in Cauca “C”, the humerus in very poor condition observed in grouping 6 and what appear to be fragments of large bones destroyed amongst the remains of the Eremotherium identified in unexcavated grouping 2 (Fig. 3C). Gravitational dispersion models of large mammal carcasses in deposition environments with inclinations greater than 20° and subjected to hydraulic forces of 152 cm/s, considered a high rate (Voorhies 1969), are easier to disperse, starting with phalanges, ribs, teeth and jaw, while the last to be transferred from the place of death would be the skull. Although gravitational dispersion can organise and select bones for transport, others tend to anchor especially in clay environments with a lower inclination and subject to lower rates of hydraulic currents (Fernández-Jalvo and Andrews 2016). However, the spatial arrangement of the bone elements of the probable same individual from the cf. Eremotherium laurillardi in Cauca “B–D” breaks with the natural anatomical pattern that could be expected in high-energy deposition environments. The same could be inferred from the remains of Notiomastodon platensis from Cauca “A” and the tusk found in Cauca “E” if there is a possible relationship between them. Despite this, we do not rule out the possibility that most of the large bones and other cranial and postcranial elements have been eroded and destroyed before or after the biostratinomic and fossil-diagenetic processes. As mentioned above, many fossils that remain on the surface have suffered fragmentation and disintegration due to the direct action of meteoric processes (e.g. very dry environment with concentrated rain) once they were exposed due to the erosion of the carrier layer and a clear example is the unexcavated grouping 2 (Fig. 3A, C).

The greatest disadvantage present when identifying potential evidence of direct action by animals (predators and scavengers) or humans in the skeletal remains of the Cauca site, as well as in their distribution pattern, includes: 1) the poor state of preservation of the remains, which limits the identification of micro modifications of anthropogenic origin on the surface of the bones, 2) absence of direct association of lithic artefacts and the remains of fossil fauna and 3) the high degree of exposure of the site to the external agents (e.g. laminar erosion) that, for years or decades, has eroded and degraded its sedimentary context. This last limitation prevents us from putting into stratigraphic context the few preformatted lithic instruments found in the vicinity of excavations B–D (Fig. 9). The origin of the pruning layer of these artefacts is uncertain and its association with the site’s megafauna is speculative and cannot be ratified, especially if the radiocarbon ages obtained for the site (40,000 cal. BP.) are taken into consideration. The latter would contrast with new cultural evidence for the American continent with ranges within the Last Glacial Maximum (LMG) (e.g. Pansani et al. (2023)).

The presence of lithic artefacts in adjacent areas of the excavations and surroundings of the Cauca site, with different lithic typologies, such as El Jobo, Clovis and Fish Tail (Jaimes et al. 2024a), undoubtedly sheds new light on technologies and distribution of the artefacts used by the megafauna hunters who roamed the region at the end of the Pleistocene. However, the lack of a chronological and stratigraphic context of the layers carrying the lithic also prevents a secure association with the megafauna of the Cauca site. The only known sites with an association of megafauna and humans from the Late Pleistocene of Venezuela (with dating), are Muaco, Cucuruchú, Taima-Taima and El Vano (see Cruxent (1970); Bryan (1973); Carrillo-Briceño (2015); Carlini et al. (2022)). In Taima-Taima (Bryan et al. 1978; Haynes 2023) and El Vano (Jaimes 1998, 1999, 2003, 2005; Jaimes et al. 2024b), there is evidence of hunting/butchering of gomphotherids (Notiomastodon platensis) and megatheres (cf. Eremotherium laurillardi), respectively, are evident. At both sites, lithic points associated with the Jobo typology were found in association with the bone remains (Cruxent 1967, 1970, 1979; Bryan et al. 1978; Jaimes 1998, 2003, 2005). Meneses and Gordones (2021) recently suggested the existence of early human presence alongside megafauna (Notiomastodon sp.) at the El Llano del Anís site in Mérida State (Fig. 2). However, the absence of precise dating, comprehensive palaeontological examinations of the site’s fossil material and substantial evidence of human involvement in the bone remains pose limitations in interpreting the site.