Pre-Clovis Archaeological sites of the Americas 8: Chiquihuite Cave, Mexico

Last year Ciprian Ardelean et al. (2020), published a paper in Nature: “Evidence of human occupation in Mexico around the Last Glacial Maximum.” This is a seminal paper, as every possible, modern archaeological technique, was thrown at the excavation by the team from the Autonomous University of Zacatecas and other institutions.

During the 8 years of excavations at Chiquihuite, starting in 2012, Ardelean and his team found ca. 2000 stone tools of various types. These included scrapers, hand axes, spear points, and projectile points.

The location of Chiquihuite Cave and other pre-Clovis sites in the Americas. It is interesting that a section of the cave is marked for future excavation. Original caption reads: The state of Piauí includes five sites: Toca do Boqueirão da Pedra Furada, Vale da Pedra Furada, Toca do Sitio do Meio, Toca da Tira Peia and Toca da Janela da Barra do Antonião-North. Sources: the journal Nature (Chiquihuite cave, early human settlements); Geological Survey of Canada (ice sheets); Paleoindian Database of the Americas (late Pleistocene coastline) Image credit: Hotz (2020).

The team on the approach to Chiquihuite Cave Sierra El Astillero mountains. No external view of the cave seems to be available on the internet, beyond the long shot in the paper itself. This may be due to the region being controlled by drug cartels. Consequently the scientists needed a police escort when trekking to and from the cave. Image credit: Brackley (2020).

Location of the cave below Chiquihuite Peak, from Extended Data Fig. 1c Ardelean et al. (2020), shown by the arrow.

Archaeologists entering Chiquihuite Cave, from Middle East Online (2020).

Ardelean explained the circumstances of the initial exploration of the cave, in an interview to Hotz (2020): “For a decade, Dr. Ardelean searched the state of Zacatecas for evidence of early human settlement. He explored 35 sites with no luck. In 2010, local villagers told him of the remote Chiquihuite Cave. When Dr. Ardelean and his colleagues first ventured into its inky interior, he expected little more than a coyote den.

He was surprised to discover two vast vaulted interconnected chambers steeply sloping down into the heart of the mountain. “We were shocked by the size of everything,” he said.

In 2012, they dug a test pit in the loosely compacted gravel, sand and fragmented rock about 150 feet from the cave entrance. The deeper they dug, the harder it was to keep the walls of the pit from collapsing. “When I left the cave, I was convinced I had nothing,” he said. “I had several bags of rocks with me that somehow looked suspicious.”

In the laboratory, they identified three hand-flaked stone chips created during toolmaking, the charred remains of palm plants and the bone from a bear’s penis. Radiocarbon dating of the bone and charcoal suggested they might be around 27,000 years old, he said.

“That rang the alarm,” he said.

In three years of fieldwork, they recovered more than 1,900 tools made from small distinctive chunks of greenish limestone not normally found inside the cave. The sharpened points, blades and scrapers had been crafted with soft hammer blows, likely from a wooden or bone striker, Dr. Ardelean said. They “looked like nothing else he had seen in the Americas.”

In the body of the paper, Ardelean et al.  (2020) explain how the chronology of the cave was elucidated using radiocarbon dating and optically stimulated luminescence (OSL) methods:

“We obtained 46 radiocarbon determinations from a total of 59 samples (bone, charcoal and sediment), as well as 6 optically stimulated luminescence (OSL) dates from a total of 8 samples. The chronology at Chiquihuite Cave is principally constructed using radiocarbon dating, with OSL dates providing an independent control. Sample selection was aimed at maintaining a close spatial relationship with lithic artefacts found along the sequence.

All bone samples with reported elemental and isotopic data (n = 6) had collagen yields, C: N atomic ratios and per cent carbon values that fell within accepted ranges (greater than 1% (weight), 2.9–3.5 and 30–50% (weight), respectively. This indicated acceptable collagen preservation and likely low levels of contamination.

All the chronometric data we obtained were incorporated into a Bayesian age model..

Modelled output estimates the start of SC-C as being before the LGM, at 33,150–31,405 calibrated years before present (ad 1950) (cal. bp) (all ranges are given at 95.4% probability); SC-B dates to between 16,605–15,615 cal. bp and 13,705–12,200 cal. bp (close in date to the Younger Dryas, at about 12,900–11,700 years ago). The entire sequence spans 20,090–17,830 years. Within the sequence, we identified a considerable gap between strata 1218–1219 and stratum 1217 that fits within the LGM, and probably indicates decreased human presence and reduced sedimentation. Overall, the chronological sequence for Chiquihuite Cave is in excellent agreement with the stratigraphic evidence.”

Ardelean et al.  (2020), Fig. 2 showing the chronology of the sediment layers. The key point is that layer SC-C with the lithics that “looked like nothing else he had seen in the Americas.” dates to 33,150–31,405 cal. bp (pre-LGM).

A larger view of the excavation. Soil sample collection in progress, from Ars Technica (2020).  Additionally, a good digital, 3D model of Chiquihuite Cave has been produced by Gandy (2021).

Let us look at the lithics in question.

Lithics from layer SC-C. All from Extended Data Fig. 5: Additional Chiquihuite Cave Lithics Ardelean et al.  (2020). Scale bar = 1cm. i and n flakes; o blade; b’, c’ and f’ points.

Top: k’ point from layer SC-C. All from Extended Data Fig. 5: Additional Chiquihuite Cave Lithics Ardelean et al.  (2020), scale bar 1cm. Middle: flake from SC-C; Bottom: Point from SC-C both from Fig 3. Ardelean et al.  (2020), middle and bottom, scale bar = 3cm. With respect to m, the Fig 3. caption gives the following, additional information: “One Pseudotsuga sp. (Douglas fir) charcoal fragment closely associated with the bifacial preform shown in m in stratum 1223 was dated to 27,929 ± 82 uncalibrated radiocarbon years bp.” This lithic point or ‘bifacial preform’ must have a calibrated date, close to the 33,000 bp maximum date, for the chronological sequence, stated above.

Point from Chiquihuite Cave, from Hotz (2020). Original caption reads: Scientists said they unearthed hundreds of unusual green limestone spear points, blades and other implements from the Chiquihuite Cave. Photo: C. Ardelean.

Scraper (?) excavated from below the LGM boundary at Chiquihuite Cave, from Curry (2020a). Original caption reads: Stone tools like this one, from deep in Chiquihuite Cave in Mexico, suggest people lived there at least 26,000 years ago. C. Ardelean.

Whilst Ardelean, stated in interview [Hotz (2020)], that the lithics “looked like nothing else he had seen in the Americas.”, others disagree. Boëda et al. (2020), provide the following observations on the lithics from layer SC-C: “Regarding the lithic material, the authors emphasize that the artifacts made on flakes do not reflect any known technical tradition. Only 13 lithic artifacts of the SC-C component are illustrated in the article, so it is not possible to comment further on the accuracy of this claim. In the paper’s Methods section, the authors note that “Inductive references to known Paleoamerican typologies were avoided”. From our perspective, this is the best way to approach the lithic industries that we do not know, given that early Paleo-American typologies are currently being (re)constructed, and none of them presents sufficient extraregional heuristic power to be used in a generalized manner of comparison.

The preliminary morpho-technical typological classification adopted by the authors reports the presence throughout the Chiquihuite sequence of cores, flakes, blades, bladelets, medial fragments of blades, modified or used flakes, scrapers (circulars forms and end scrapers), burins, backed knives, points, adzes, and point-like objects. All of these types of artifacts, individually or collectively, are present in the South American Pleistocene sites dating to before, during, and after the LGM, made on limestone and other raw materials, notably basalt and andesite in the western sector of the southern continent, and mainly quartz and quartzite in the eastern sector. Perhaps the only novelty is the presence of bifacial pieces and laminar elements in the older components. The so-called “bifacial preform” from stratum 1223 (27,929 ± 82 14C yr BP) is the most ancient bifacial piece reported to date in the Americas. Given the excessive attention given to bifacial pieces in American prehistory, perhaps this evidence will promote a recognition of the anthropogenic character of associated non-bifacial artifacts, largely forgotten from any detailed technological

study. Nevertheless, the few lithic pieces illustrated in Ardelean et al. (2020) present an indisputable anthropogenic character, in addition to the obvious techno-functional coherence of each piece individually.

The fact that some researchers are already ruling against the anthropogenic nature of these artifacts (e.g., Curry 2020b) reflects a usual ad hoc assumption – proof of a lack of knowledge of the technical specificities of lithic materials, and in particular of South American industries. The authors emphasize that “transverse points” are a recurring feature throughout the Chiquihuite stratigraphic sequence, also present in the SC-C component.

Another recurring element is marginal retouching (referred to as “edge trimming”). Whether it is unifacial or bifacial retouching, this technique was used to modify flakes into tools. These two technical characteristics are also present in South American Pleistocene sites, in particular those located in northeastern Brazil, where we have recognized “convergence” of worked margins, avoiding the ambiguous term “point.” We use the term “convergence” to refer to a particular techno-type of tool in northeastern Brazil, made up of two edges that form a convergence due to a retouching or shaping operation. The term “point” is functional in nature, while the term “convergence” is rather technofunctional in nature.”

Boeda et al., (2020), then provide figures drawn from other, recent papers, that show, that similar lithics have been excavated in South America:

Figure 1 Locations of sites referenced in the text, dates of those sites, and artifacts representative of their lithic industries (the figure elaborated from and based on Boëda et al. 2014, 2016; Lahaye et al. 2019; Lahaye et al. 2013; Vialou et al. 2017).

Boeda et al. (2020), go on: “In the northeast of the Brazilian state of Piauí (Brazil), our team and others (Guidon 1989; Parenti 2001) have recorded several different geological contexts (cuesta, terrace, valley, limestone massif, cave, rock shelter) with long archaeological sequences. In the limestone karst zone, we reported three sites with Pleistocene archaeological sequences, containing remains of mesofauna and megafauna: Toca da Pena, Tira Peia, and Toca da Janela da Barra do Antonião-Norte (TJBA-N) (Figure 1). These sites share several characteristics with Chiquihuite Cave. For example, in the rock shelter of Tira Peia, a site located on a limestone massif, we have recovered a large number of lithic artifacts made on quartz and quartzite pebbles, and limestones rich in silica in the C6 and C7 layers, dated between 20,000 and 15,100 cal yr BP (Lahaye et al. 2013). Additionally, a dozen similar artifacts were found in layers C8 and C9, exceeding 22,000 cal yr BP (Boëda et al. 2013). One km northeast of Tira

Peia, the C5a layer of TJBA-N has been dated between 19,000 and 20,400 cal yr BP, containing lithic artifacts on quartz and quartzite (Lahaye et al. 2019). Likewise, in the cuesta area, the first occupation of Sítio do Meio, dated between 35,000 and 28,000 cal yr BP and with an occupation gap between 28,000 and 24,000 cal yr BP, presents a phenomenon of microlithism evident in the quartz-pebble industry (Boëda et al. 2016, figure 6). All of these sites yielded lithic artifacts that cannot yet be traced to any known Paleo-American technocultural tradition, but that does not mean they do not exist.”

Faunal and human DNA from Chiquihuite Cave

The most interesting aspect of the excavation was the collection of sediment samples for environmental DNA testing. The main reason was to ascertain whether any ancient DNA, comparable to the date of the oldest layers with human-made lithics. As this ancient DNA would be from pre-Palaeoindian times, it would be extremely interesting to find out who the makers of the tools were most closely related to, and thus where these early migrants to the Americas, originated. Unfortunately, according to the authors, they found none. Except that isn’t strictly true. They actually, did detect human DNA at various points throughout the sequence. I’ll come back to that later.

Scientist collecting soil samples for eDNA testing, from NBC News (2021) 

In terms of the faunal DNA, a great deal was found, as Ardelean et al. (2020) say: “Among the fauna, bat DNA is present in all layers. This is dominated by Eptesicus sp., Myotis sp. and Vespertilionidae, until stratum 1204, in which Phyllostomidae, Desmodus sp. and new Microchiroptera species replace the previous assemblage. By contrast, although bear DNA (Ursus sp.) appears during the LGM, the highest abundance occurs during the termini, which is in concordance with the archaeological evidence2,19 (Supplementary Information 1.5, Supplementary Fig. 7). Rodents (Marmota sp., Microtus sp., Ictidomys sp., Urocitellus sp. and Peromyscus sp.) are present throughout the sequence, with a higher incidence in a few of the strata. Deer mouse (Peromyscus sp.), vole (Microtus sp.) and marmot (Marmota sp.) appear to be more abundant than other rodents, with kangaroo rat (Dipodomys sp.) appearing in the youngest stratum (stratum 1201). DNA findings also include goat (Caprinae, probably Oreamnos sp.) and sheep (Ovis sp.), as well as a low proportion of DNA from birds, such as sparrow (Zonotrichia sp.), falcon (Falco sp.) and tanagers (Thraupidae).”

I found this vague list, with its assignment to the level of genera only, quite frustrating. I therefore looked at the relevant Supplementary Information sections, to see if more detail was given. I was surprised to find which reference genomes the authors chose to use. Here are a few examples.

Marmota sp. From their Supplementary Data Ardelean et al. (2020), the reference genome used was Marmota marmota. This is the Alpine Marmot, of European only distribution. This is particularly odd as a recent paper resolving the phylogeny of all 15 Marmota sp. was published by Steppan et al. (2011). The paper notes areas for fossil finds of Marmota, with Marmota flaviventris, the yellow-bellied marmot, previously found in New Mexico. It was therefore, probably this species that is represented by the DNA and fossils from Chiquihuite Cave.

Another important group of fossils and eDNA, samples, from the cave are those assigned to Equus or horse species. Some of these are from early in the sequence, meaning they were native Equus species making them doubly interesting. Consulting “Extended Data Fig. 4 Taxonomic profiles of animals (Amniota) and plants (Viridiplantae) identified by ancient environmental DNA.”, one finds that Equus sp. DNA was found in strata UE1217, UE1218 and UE1223. These strata, range in modelled age from ca. 25,000 bp, (UE1217) or around the LGM to ca. 31,550 bp UE1223 – see Ardelean et al. (2020), Fig 2. Not only that but two species of Equus were found. This is very exciting! Looking at the reference (Supplementary Data) genomes used were Equus przewalskii and Equus asinus. Once again these are both odd choices for the reference genomes.

Taking Equus przewalskii first, this is the Eurasian Wild Horse, which, in the prehistoric era ranged over central Asia, China, and western Europe. This is, and never was an American species. Again, reference genomes for north American wild horses from the same region of Chiquihuite Cave, were available from a paper by Barron-Ortiz et al. (2017). Their conclusion reads: “Two equid species, Equus ferus and E. conversidens, are identified for the late Pleistocene of the Western Interior of North America, based on molecular and morphological analyses of the cheek teeth. A third species, E. cedralensis, is provisionally recognized based exclusively on the morphological analyses of the cheek teeth.” Despite the rather confusing assignment of one clade from their DNA and morphological studies to Equus ferus, the fact remains that they identified three, distinct species. All were found in the region of Chiquihuite Cave. Lastly Machado and Avilla (2019) used morphometric on a large number of specimens of to suggest that all species of Equus from south America should be synonymised under a single species: Equus neogeus. While only a partial sequence for this species have been published (see Orlando et al. 2008), I would have hoped that, Ardelean et al. (2020), would have compared their sequences to this available data and the other, Equus DNA sequences available.

A second species of Equidae was also discovered via eDNA sampling by Ardelean et al. (2020). Once again looking at the Supplementary Data spreadsheet we find an even more extraordinary choice of reference genome, namely Equus asinus. By ‘Equus asinus’ I can only assume the authors mean the Asinus subgenera of Equus, commonly known as the Eurasiatic wild ass. These include the African wild asses, (ancestor of the domesticated Donkey), Onagers from the Asian steppes and Kiang, the Tibetan wild ass. While it is possibly found in the topmost layer [UE1201] as a feral species introduced at the time of the conquistadores, the other DNA samples recovered from throughout the sequence going right down to UE1223, dated ca. 31,500 bp simply cannot be members of this species group.

I cannot fathom the assignment of these DNA samples to this reference genome, nor the lack of comment on this glaring anomaly by other reviewers of the paper.

 

The last example I wish to look at is Urocitellus parryii, the Artic Ground Squirrel. Once again a quick check on the phylogeny of the genus, brings up adequate references to DNA studies on Urocitellus. Herron et al. (2004) and Helgen, et al. (2009) revise the taxonomy of this genus and give DNA sequences. To cut a long story short, the ground squirrel DNA found by Ardelean et al. (2020) in stratum UE1207C (modelled at 25,000 to 17,000 bp) cannot be Urocitellus parryii as this is a circum-polar species found in Canada, Alaska and Siberia. The most likely species seem to be, Urocitellus beldingi; Urocitellus mollis; or Urocitellus townsendii; considering their current distribution. Once again, I must ask why the authors chose the least likely reference genome form the 12 other, known species which might have provide a better fit in terms of likelihood of occurrence in the Chiquihuite Cave region.

I must therefore ask why these reference genomes were chosen. Was it the fact that they were the ones most widely available? Was it that, the DNA sequences recovered by Ardelean et al. (2020), were truly closest to the species that they were assigned to – although this seems unlikely from my discussion above. Another possibility is that the range of these Marmota, Equus, and Urocitellus species were formerly much greater, and once extended to northern Mexico. I guess we’ll have to wait for a more in depth paper? It would make most interesting reading, if indeed it is forthcoming.

Lastly, I come to the human DNA. Ardelean et al. (2020), unequivocally state: “we found no evidence of ancient human DNA within the samples”. This however is not the case. Abundant human DNA was found throughout the sequence of strata. It is the source that it originated from that is in question. In their XXXXX Ardelean et al. (2020) state: “We investigated the presence of ancient human DNA out by mapping sequencing reads of each sample against two different reference indices (see Methods). We first determined the presence of mitochondrial (MT) sequences by mapping against a reference index containing all mitochondrial genomes contained in the RefSeq database (release 92). Reads mapping uniquely and with high quality (MQ25) to a single MT reference contig were extracted and assessed for genomic coverage and ancient DNA damage. We find that only the sample from UE1210 contains sufficient human MT reads for analysis, with a total of 189 reads mapping at MQ25 and covering ~58% of the MT genome (contig NC_012920.1, Supplementary Metadata file).

However, rates of characteristic ancient DNA damage substitutions (5’ C>T or 3’ G>A) were indistinguishable from other substitution types, indicating that the reads originated likely from contaminating modern human DNA. This contrasts with reads mapped to the American black bear MT (contig NC_003426.1) from the same sample, which showed similar genomic coverage but elevated rates of 5’ C>T and 3’ G>A substitutions, consistent with authentic ancient DNA (Supplementary Metadata file, Fig. S62).

When using the full human genome as a reference index, we find reads mapping from all samples, with coverage ranging from 595 up to 32,727 reads at MQ25 (Supplementary Metadata file). For the majority of samples rates of ancient DNA damage substitutions were ≤ 0.01, again suggesting their modern origin. However, three of the strata (UE1210, UE1212, UE1215) exhibit elevated rates, ranging from 0.03 up to 0.07 (Supplementary Metadata file). As remnant, human background contamination present in the reagents used in the laboratory preparations will dilute or decrease the DNA damage signal of ancient human reads if present in only low quantities, we considered those samples as putative candidates for follow up.

A further complication for the ancient DNA authentication stems from the possibility of spurious mapping of reads that originate from DNA sequences conserved between humans and closely related species. If ancient DNA sequences of a closely related species are present in sufficient numbers in the sample, their spurious alignment to the human genome can create a false-positive signal of ancient DNA damage. To investigate whether this was the case for our samples, we re-calculated the substitution rates restricting to reads assigned to Old world monkeys (Hominidae) using the ‘Holi’ pipeline, parsing reads with mismatches ≤ 5 for all samples and controls. Substitution rates of 5’ C>T and 3’ G>A changes were indistinguishable from the other types after this filtering step (Fig S63-S65), suggesting spurious mapping as the main culprit for the elevated rates observed before.” 

Up to the part “However, three of the strata (UE1210, UE1212, UE1215) exhibit elevated rates, ranging from 0.03 up to 0.07”, I am in full agreement with Ardelean et al. (2020). Elevated levels of DNA damage is a key indicator that ancient DNA is present. The suggested reason for these samples being discarded as truly ancient human DNA, is ‘spurious mapping’ of the sequences onto ‘closely related species’. The modelled dates for strata UE1210 to UE1215 are ca. 17,000 to 19,000 bp, consequently if these samples DO represent ancient human DNA, they would be between 4,400 and 6,600 years older than that of the oldest DNA from the Americas: the Anzick Child at ca. 12,600 bp. Human DNA, from the Americas not only definitively break the ‘Clovis first’ hypothesis, it would almost certainly reveal a great deal about the origins of the earliest people (so far discovered) to migrate to the Americas.

So let us investigate what closely what ‘closely related species’ could be in northeastern Mexico at this time. Looking at the temporal window between 17,000 and 19,000 bp we are still in the slow thawing phase of the last glacial maximum. According to Vázquez-Selem and Lachniet (2017): “The LLGM (last local glacial maximum) in the mountains of Mexico (20-14 ka) and Central America (~21-18 ka) overlaps with the final part of the global Last Glacial Maximum (26.5-19 ka).” This is the period from which the ‘human’ DNA from stratum UE1215 originates. One has, therefore, to ask which ‘closely related species’ could be extant in northeastern Mexico at this time? The obvious candidates are the New World Monkeys (Infraorder Platyrrhini). A check on the phylogeny of New World Monkeys and Old World anthropoid revealed a paper by Schrago and Russo (2003), estimates the split of Platyrrhines split Catarrhines (the branch leading to (Homo sp.) at around 35 MYA.

Phylogeny of New World Monkeys and Old World Anthropoids, from Schrago and Russo (2003).

Roughly speaking, the conservation of stretches of DNA through time and the points at which these sequences diverge is used to put an age estimate on the split times for the different groups above. While some hominoid proteins evolved significantly slower than those of other primates, the mitochondrial clock rate was found to be more or less unaffected. In consequence, DNA sequences from before the split of NWM and OWA are conserved in NWM.

Therefore, the statement that ‘spurious mapping’ of the sequences onto ‘closely related species’ by Ardelean et al. (2020) is plausible.

Having said that, the question we must ask is whether there were any NWM in the Chiquihuite Cave region during the slow alleviation of the inhospitable climate in the region, in the wake of the LGM, between 19,000 and 17,000 bp? Consulting the literature very widely, I was unable to find any reference to fossil NWM from the relevant time-period. This is not to say that such fossils do not exist, but that they are not mentioned in any of the texts I consulted.

If we step back for a moment and consider the environmental adaptations of monkey species worldwide, one species capable of ling in cold, icy conditions at altitude does spring to mind. I am of course referring to the Japanese Snow monkey, Macaca fuscata. This Japanese macaque, inhabits subtropical forests in the southern part of its range and subarctic forests in mountainous areas in the northern part of its range. The macaque can cope with temperatures as low as −20°C and live up to and beyond 3,000m in altitude.

Japanese macaques seen in deep snow, from JRPASS (2021).

While it is possible that an analogous NYM, similar to the Japanese Snow Monkey did exist in northeastern Mexico at the relevant time, no fossils have ever been found, even by Ardelean et al. (2020).

One can therefore, only conclude on the balance of the evidence that, no New World Monkeys were extant in the region of Chiquihuite Cave between 19,000 and 17,000 years ago. This leads us to the inevitable position of having to accept that the samples of ‘human’ DNA from strata UE1210 to UE1215, with their characteristically damaged DNA and elevated rates of 5’ C>T and 3’ G>A substitutions, were consistent with authentic ancient DNA and were actually human and not from a ‘closely related species’.

Conclusions

1. The evidence of the lithics, C13 dating and OSL dating leads unassailable conclusion that humans were in north America ca. 31,500 bp.

2. Extremely valuable ancient Equus DNA samples were recovered from Chiquihuite Cave sediments which could help reveal the true phylogenic relationships, geographic distribution and dispersals of the genus in the Americas.

3. A robust chronology for the strata within Chiquihuite Cave was built using complimentary scientific methods. Human DNA was found in the sequence of strata and could be dated to 19,000 – 17,000 bp using this chronology.

4. The reference genomes used to identify various species/genera was somewhat inexplicable, while exclusion of human DNA and its attribution to ‘closely related species’, on flimsy evidence seems deliberately disingenuous. On this point, perhaps the authors didn’t want to give critics ammunition to knock the paper down. Put simply, I believe they were just playing it safe.

Acknowledgement: Hat tip to Marnie Dunsmore (2021) at the Linear Population Model blog who put many of us onto this site in 2017/2018(?). Unfortunately, I can’t go back and read her posts as the blog is now private and only open to invited guests.

References

Ardelean, C.F., et al. (2020). Evidence of human occupation in Mexico around the Last Glacial Maximum. Nature 584, 87–92.

Ars Technica (2020). People may have lived in North America by 30,000 years ago. At: https://arstechnica.com/science/2020/07/people-may-have-lived-in-north-america-by-30000-years-ago/ accessed 06/08/2021

Barron-Ortiz, C. I. et al. (2017). Cheek tooth morphology and ancient mitochondrial DNA of late Pleistocene horses from the western interior of North America: Implications for the taxonomy of North American Late Pleistocene Equus. PLoS ONE 12(8): e0183045.

Boëda, E., I. Clemente-Conte, M. Fontugne, C. Lahaye, M. Pino, G. Felice, N. Guidon, et al. (2014). “A New Late Pleistocene Archaeological Sequence in South America: The Vale da Pedra Furada (Piauí, Brazil).” Antiquity 88: 927–941.

Boëda, E., R. Rocca, A. Da Costa, M. Fontugne, C. Hatté, I. Clemente-Conte, J. Santos, L. Lucas, G. Felice, et al. 2016. “New Data on a Pleistocene Archaeological Sequence in South America: Toca do Sítio do Meio, Piauí, Brazil.” PaleoAmerica 2 (4): 286–302.

Eric Boëda, Ruth Gruhn, Agueda Vilhena Vialou, Carlos Aschero, Denis Vialou, Mario Pino, Maria Gluchy, Antonio Pérez & Marcos Paulo Ramos (2020): The Chiquihuite Cave, a Real Novelty? Observations about the Still-ignored South American Prehistory, PaleoAmerica. At: https://www.researchgate.net/publication/346718010_PaleoAmerica_The_Chiquihuite_Cave_a_Real_Novelty_Observations_about_the_Still-ignored_South_American_Prehistory accessed 09/08/2021

Brackley, P. (2020). How dangerous mission to Mexican cave uncovered evidence of people in Americas 15,000 years earlier than thought. At: https://www.cambridgeindependent.co.uk/news/how-dangerous-mission-to-mexican-cave-uncovered-evidence-of-people-in-americas-15-000-years-earlier-than-thought-9118951/ accessed 08/08/2021

Curry, A (2020a). Were humans living in a Mexican cave during the last ice age? At: https://www.sciencemag.org/news/2020/07/were-humans-living-mexican-cave-during-last-ice-age accessed 07/08/2021

Curry, A. (2020b). “Tools Suggest People Reached Americas Early.” Science 369: 1416–1417.

Dunsmore, M. (2021). At: http://linearpopulationmodel.blogspot.com/

Gandy, D. (2021). Chiquihuite Cave 3D model. At: https://sketchfab.com/3d-models/chiquihuite-cave-excavation-x-12-2caa3caeb04d4170a02a42ae685d9d54 accessed 06/08/2021

Guidon, N. (1989). “On Stratigraphy and Chronology at Pedra Furada.” Current Anthropology 30 (5): 641–642.

Herron, Matthew D.; Castoe, Todd A.; Parkinson, Christopher L. (2004). "Sciurid phylogeny and the paraphyly of Holarctic ground squirrels (Spermophilus)". Molecular Phylogenetics and Evolution. 31 (3): 1015–30.

Helgen, Kristofer M.; Cole, F. Russel; Helgen, Lauren E. & Wilson, Don E (2009). "Generic Revision in the Holarctic Ground Squirrel Genus Spermophilus". Journal of Mammalogy. 90 (2): 270–305.

Hotz, R. L (2020). Mexican Cave Find Hints That People Lived in North America 30,000 Years Ago. Wall Street Journal. At: https://www.wsj.com/articles/mexican-cave-find-hints-that-people-lived-in-north-america-30-000-years-ago-11595430002 accessed 07/08/2021

JRPASS (2021). Monkey Around in Japan: The Snow Monkeys of Jigokudani Monkey Park. At: https://www.jrpass.com/blog/monkey-around-in-japan-the-snow-monkeys-of-jigokudani-monkey-park accessed 13/08/2021

Lahaye, C., M. Hernandez, E. Boëda, G. D. Felice, N. Guidon, S. Hoeltz, A. Lourdeau, et al. (2013). “Human Occupation in South America by 20,000 BC: The Toca da Tira Peia site, Piauí, Brazil.” Journal of Archaeological Science 40 (6): 2840–2847.

Lahaye, C., G. Guérin, M. Gluchy, C. Hatté, M. Fontugne, I. Clemente-Conte, I. J. C. Santos, et al. (2019). “Another

Site, Same Old Song: The Pleistocene-Holocene Archaeological Sequence of Toca da Janela da Barra do Antonião-North, Piauí, Brazil.” Quaternary Geochronology 49: 223–229.

Machado, H. and Avilla, L., 2019. The diversity of south American Equus: did size really matter? Frontiers in Ecology and Evolution, 7, p.235. 

Middle East Online (2020).  New evidence of humans in America is twice as old as thought. At:  https://middle-east-online.com/ accessed 05/08/2021

NBC News. (2021). Ancient stone tools suggest first people arrived in America earlier than thought. At: https://www.nbcnews.com/science/science-news/ancient-stone-tools-suggest-first-people-arrived-america-earlier-thought-n1234578 accessed 06/08/2021

 Orlando, L., Male, D., Alberdi, M. T., Prado, J. L., Prieto, A., Cooper, A., et al. (2008). Ancient DNA clarifies the evolutionary history of american late pleistocene equids. J. Mol. Evol. 66, 533–538.

Parenti, F. 2001. Le gisement quaternaire de Pedra Furada (Piaui, Brésil): Stratigraphie, chronologie, évolution culturelle. Ed. Paris: Recherche sur les civilisations.

Schrago, C.G. and Russo, C.A., (2003). Timing the origin of New World monkeys. Molecular biology and evolution, 20(10), pp.1620-1625.

Steppan S. J. et al. (2011). Molecular data resolve placement of the Olympic marmot and estimate dates of trans-Beringian interchange. Journal of Mammalogy, 92(5):1028–1037. 

Vázquez-Selem, L. and Lachniet, M.S., 2017. The deglaciation of the mountains of Mexico and Central America. Cuadernos de Investigación Geográfica, 43(2), pp.553-570.

Vialou, D., M. Benabdelhadi, J. Feathers, M. Fontugne, and A. V. Vialou. (2017). “Peopling South America’s Centre: The Late Pleistocene Site of Santa Elina.” Antiquity 91: 865–884.