These notes distinguish reporting discrepancies from objections to the archaeological interpretation. None is silently repaired, and none alone decides whether people made the objects.
Conflicting model summaries
Table S2 prints Unit III end as 27,225–25,225 cal BP (95.4%) and 26,230–26,230 (68.3%); supplementary narrative instead gives 27,574–25,225 and 27,253–26,230. Unit II start ends at 16,880 in Table S2 and the main text, versus 16,850 in the supplement narrative. Preserve both versions; do not plot these disputed boundaries as settled results.
Input list and outline need reconciliation
Table S1 assigns the wood to III(2), whereas the supplied OxCal outline names a III(3) phase. The outline contains placeholders rather than the full numerical model. A complete executable input and output are needed for reproduction. UW462 appears in the 2017 report but is absent from the 35-row Table S1.
Published uncertainty and calibration history
UW465 is 25,100 ± 2,500 years in 2017 and Table S1, but the 2023 main narrative abbreviates its error to ±2 ka. The 2017 text repeats 27,818–26,887 cal BP for Gif-9365 and GIFA-99177 despite different raw measurements. This ledger retains the raw ages; no replacement calibration has been calculated.
What would strengthen the association?
Check original field registers for object-ID-to-position mapping and dating-sample coordinates. Osteoderms with imprecise positions were excluded from the precise tool–bone spatial analysis. Spatial proximity does not identify the tool that made a particular mark.
Human manufacture versus natural damage
Costopoulos (2023, scholarly blog commentary) asks whether rare natural lesions could account for three selected objects in a large assemblage, and calls for stronger perforation-wall comparisons. S77 already supplies FR3A/FR6 images and an 80-osteoderm survey; the open question is their diagnostic adequacy. This critique does not demonstrate a natural origin.
Do not conflate the experimental controls
S77 Text S6, PDF p.7: modern dry armadillo UFCAT-MAM 1/6 was polished without drilling; fossil sloth SEI8004 was drilled and polished. The commentary describes drilling modern osteoderms. Preserve the primary protocol. Experimental similarity establishes plausibility, not exclusive human causation.
Tooth bioapatite and converted estimates are different results
S186 publisher abstract: Unit III4 tooth bioapatite gives a reported minimum of 22,339–22,534 cal a BP; the authors’ conversion yields 22,042 ± 40 conventional 14C BP, calibrated to 25,994–26,396 cal a BP. Unit II2 gives 14,944–15,239 cal a BP minimum and a converted 14,547 ± 40 conventional 14C BP / 17,450–17,906 cal a BP. These are not four independent assays or extracted-collagen measurements. Full methods, IDs and confidence conventions remain unverified; neither tooth dates ornament manufacture.
Possible burning: later Unit II, preliminary report
The October 2026 PB-86 conference abstract reports heat indicators in a Unit II osteoderm and rib, concluding possible burning that could be human-caused at approximately 13 kya BP. No precise calibrated range or specimen-to-hearth ledger is provided there. Do not assign this to the Unit III ornaments or infer cooking, hunting or deliberate fire from dark colour alone.