Paisley Caves

What happened

1. 2008Gilbert et al. report human mtDNA (haplogroups A2, B2) from coprolites directly AMS-dated to 12,300 14C yr BP (~14,300 cal BP), predating Clovis.
2. 2009Goldberg et al. and Poinar et al. publish comments questioning coprolite identification and dating reliability; Gilbert/Rasmussen teams rebut with new data.
3. 2012Jenkins et al. report 100+ AMS dates, Western Stemmed points at Clovis-age levels, and blind independent-lab DNA confirmation.
4. 2013Hockett & Jenkins report two cut-marked bones directly dated before Clovis (>13,100 cal BP), independent of coprolite DNA.
5. 2014Sistiaga et al. re-test the disputed coprolite with steroid biomarkers and confirm a herbivore, not human, origin.
6. 2014Fiedel publishes continued skepticism about the pre-Clovis claim.
7. 2018Shillito et al. flag that ancient DNA may be mobile in cave sediment, questioning reliance on mtDNA for coprolite identification.
8. 2020Shillito et al. confirm 3 pre-Clovis coprolites as human via lipid biomarkers, date a basketry fragment, and flag mtDNA errors in 6 others.
9. 2023Saban et al. publish a pollen study describing Paisley Caves' dated sequence from 13,000-6,000 cal BP.

Objections 1 of 3 proved wrong

• ✓
  Micromorphology and FTIR analysis of coprolite 1374-5/5D-31-2 indicated a herbivore, not human, origin; a 2014 steroid-biomarker re-test agreed, conflicting with its mtDNA identification as human.

  Natural, not human · Paul Goldberg, Francesco Berna, Richard I. Macphail; Sistiaga et al. (2014) · 2009

  Held up — Confirmed for this specimen by three methods (2009-2014); Shillito et al. (2020) showed mtDNA can misclassify non-human coprolites as human, and separately confirmed three other coprolites as human by biomarkers.
• ✓
  DNA authenticity was questioned: without demonstrated intact stratigraphy, ancient DNA in coprolites might be mobile in cave sediment (leaching) rather than intrinsic to the tested specimen.

  Dating or samples · Hendrik Poinar, Stuart Fiedel, et al. · 2009

  Held up — Shillito et al. (2020) found 6 of 21 mtDNA-typed 'human' coprolites gave carnivore biomarkers, and one Camelidae coprolite dated ~810 14C years apart between fractions, consistent with mobile solutes.
• ✕
  In the absence of demonstrated intact stratigraphy and diagnostic artifacts, anomalous carbon-isotope values were said to make the oldest coprolites' radiocarbon dates unreliable.

  Layers or association · Hendrik Poinar, Stuart Fiedel, et al. · 2009

  Proved wrong — Jenkins et al. (2012) added 100+ AMS dates and pre-Clovis cut-marked bone; Shillito et al. (2020) found basketry (12,273±56 14C BP) stratified above coprolite 185, confirming intact stratigraphy and pre-Clovis age.

Evidence at the time within ~5 years of the claim

• Unambiguous evidencePartly — mtDNA indicated human origin, but Goldberg et al. (2009) found micromorphology/FTIR analysis of one dated coprolite hard to reconcile with a human source.
• Documented contextPartly — Excavated in situ by a university field school with recorded stratigraphy, but Poinar et al. (2009) argued the deposits lacked demonstrated stratigraphic integrity at claim time.
• Independent dating methodsNo — Only AMS radiocarbon dating was used to directly date the coprolites at claim; no second independent dating method was applied yet.
• More than one labYes — By 2012 (four years after the claim), an independent laboratory blind-tested coprolite samples and confirmed human DNA in pre-Clovis specimens.
• Professional recoveryYes — University of Oregon archaeological field school directed by Dennis Jenkins recovered and recorded the material.
• Peer-reviewed reportYes — Published as a Report in Science (2008), a leading peer-reviewed journal.
• Open to criticsYes — Critics Goldberg, Berna and Macphail obtained and independently analyzed one of the actual dated coprolites.
• Independent teamsPartly — Critics (Goldberg, Berna, Macphail) were independent, but the confirming lab was commissioned by proponents and already counted under multiple_labs; no independent corroboration by 2013.
• Outside stakeNo — Ordinary university field-school research; no commercial, religious or nationalist motive documented.

What would settle it for a field partner

1. Independent-lab lipid-biomarker re-analysis of split samples from the 3 human-confirmed coprolites, using material archived from the original extraction.Medium cost
   Would show: Whether an outside lab replicates the human biomarker signal on the same specimens.
2. Second AMS date, at a lab independent of the original program, on the basketry fragment and the two cut-marked bones.Low cost
   Would show: Whether the pre-Clovis age holds independent of the Oregon program's own dating.
3. New excavation or coring by a team independent of the Jenkins program, targeting undisturbed pre-Clovis strata.High cost
   Would show: Whether an independent team can recover further pre-Clovis material at the site.

What settled it Replication, Forensic analysis

What would change this record

Failure of an independent lab to replicate the human lipid-biomarker signal on split samples of the 3 confirmed coprolites, or evidence that the basketry fragment or cut-marked bones are intrusive into pre-Clovis strata, would overturn this status.

Notes

• Status. Core claim vindicated: cut-marked bone, dated basketry and 3 biomarker-confirmed coprolites put people here ~14,300 cal BP. The mtDNA method misfired (6 of 21 'human' coprolites were carnivore); all evidence is from one program.
• Accepted view then. Clovis/Nenana (~13,500-13,000 cal BP) was North America's oldest accepted culture; Monte Verde, Chile (~14.5 ka) was already accepted by 2008.
• Now understood. ~14,000-14,300 cal BP, reaffirmed since 2008
• Dating basis. Direct AMS 14C on a coprolite yielding human mtDNA: 12,300 14C yr BP (Gilbert et al. 2008), commonly calibrated to ~14,300 cal BP. Further coprolites and a basketry fragment later dated to a similar range.
• Refined since. mtDNA proved unreliable for typing coprolites: lipid biomarkers showed 6 of 21 mtDNA-'human' coprolites were carnivore, probably because DNA moves through cave sediment (Shillito et al. 2018, 2020).
• Refined since. Coprolite 1374-5/5D-31-2, typed human by mtDNA in 2008, is herbivore by micromorphology and FTIR (Goldberg et al. 2009) and by steroid biomarkers (Sistiaga et al. 2014).
• Refined since. The pre-Clovis case now rests on lines independent of mtDNA: biomarkers on 3 coprolites, basketry dated 12,273±56 14C BP above coprolite 185, and two cut-marked bones below Clovis levels (Hockett & Jenkins 2013).
• Correction. Some summaries describe the 2012 Western Stemmed points as pre-Clovis evidence; Jenkins et al. (2012) date them to 11,070-11,340 14C BP and title them 'Clovis Age', not pre-Clovis.
• Correction. Some accounts place Shillito et al.'s 2020 lipid-biomarker study in Science, by association with Gilbert et al. 2008; it appeared in Science Advances 6(29).
• Pre-Clovis evidence now rests on three lines independent of the original mtDNA calls: lipid biomarkers on three coprolites, a basketry fragment above dated coprolite 185, and two cut-marked bones below Clovis levels (Hockett & Jenkins 2013). The 2012 Western Stemmed points are Clovis-age, not pre-Clovis. All this evidence comes from one research program; no independent excavation has occurred. Of 21 re-tested coprolites, 6 that mtDNA called human proved carnivore by biomarkers; one mixed case may reflect coprophagy rather than contamination alone. Fiedel's 2014 final verdict could not be confirmed. No precise coordinates recorded, per heritage-ethics guidance.

People and evidence

• Proponents. Dennis L. Jenkins, M. Thomas P. Gilbert, Loren G. Davis, Thomas W. Stafford Jr.
• Critics. Hendrik N. Poinar, Stuart J. Fiedel, Paul Goldberg, Francesco Berna, Ana Sistiaga
• Evidence types. radiocarbon, dna, coprolites, lithics, artifacts, stratigraphy, chemistry, plant remains, fauna · 4 independent lines
• Record confidence. medium — Core chronology (S1, S6, S9 full text) is well verified. S7's abstract is elided (TLDR used); S8 only partly readable. Medium confidence given DNA-reliability complexity.