Rejecting a quotation from publication and erasing it from the record are two different decisions. This page is the record.
What this page is
Every paper the article rests on was downloaded and read locally, so nothing on this page failed on wording. Five items were held back. Three of the five have since been released: two sources recorded as unreachable were free at Europe PMC, and the Figure 4A reproduction was solved. Released entries are kept here, marked, rather than deleted.
Status words.
- UNOBTAINED — no route to the named source has been found yet. A statement about this project's reach, not about the claim.
- NOT FETCHED BY CHOICE — a route exists and costs a metered borrow, and the claim did not justify spending one.
- RELAYED — the wording is genuine and correctly attributed to the page that carries it, but that page is quoting a primary source this project has not read.
- PRINTED WITH A FLAG — the substance is verified, but something about the citation or the reproduction is incomplete, so the article says so in place of asserting it.
- RELEASED — the obstacle was cleared on a later pass and the material is now in the article. The entry stays as a record of what was tried and what worked.
Nothing on this page was established with an automated summarising fetcher. Every check was made against downloaded bytes.
1. Anything at all from Liu et al. 2006
Status: RELEASED, 6 September 2026. Free at Europe PMC. curl -sL 'https://europepmc.org/articles/PMC1559480?pdf=render' returns the publisher PDF with no login and no CAPTCHA. Read with pdftotext -raw. Now in the source library and cited in Part 3. One correction to the reasoning below: Liu 2006 is not a third origin scan. It assumes an East African origin and fits demographic parameters to it. What it does supply is a warning against the article's own method, on p. 234: "An important corollary stemming from this nonlinearity is that use of linear regressions on gene diversities might not allow correct inference of the geographic origin of modern humans." The record of the failed first attempt follows.
What it is. Liu, Hua, Franck Prugnolle, Andrea Manica and François Balloux, "A Geographically Explicit Genetic Model of Worldwide Human-Settlement History," American Journal of Human Genetics 79(2):230–237, 2006.
Why it was wanted. It is the third independent origin scan and the only one that fits an explicit stepwise colonisation model rather than a regression on distance. That makes it the best published answer to the objection that a regression-based origin scan assumes the serial founder model it is testing. A quotation from it would have strengthened Part 3 exactly where this article is weakest.
What was tried. PubMed Central, where the older AJHG archive is open for most years but this article is not deposited. A curl of the publisher's showPdf URL, which returned 403. The same URL through a headless browser after first loading the fulltext page in the same session, which is the exact route that produced the PDFs of Prugnolle 2005 and Jacobs 2019, and which here returned a paywall redirect instead of application/pdf. The rendered fulltext page, which yields an abstract shell of about 900 words and no body.
What would change this. An institutional login, or the print volume at a library. Until then the article makes the circularity concession without the source that would answer it best, and says so.
2. Anything at all from Lachance et al. 2012, including the SNP count
Status: RELEASED, 6 September 2026. bin/pmcget.py on https://pmc.ncbi.nlm.nih.gov/articles/PMC3426505/ returns the full text. The ?pdf=render route that worked for Liu 2006 returns a 500 for this item, and cell.com 403s. The relay is closed: Lachance's own sentence reads "we identified 13,407,517 variants (SNPs, insertions, and deletions that differ from the human genome reference sequence, GRCh37/hg19) in the 15 hunter-gatherer genomes," so ICR's "13,407,517 SNPs" over-narrows the category. The article now cites Lachance directly and reports the difference. The record of the first pass follows.
The wording as it circulates. "their study found 13,407,517 SNPs that differ from the human genome reference sequence."
Where the article got it. From the Institute for Creation Research page "African Populations Fit Biblical History" (2012), which is reporting Lachance, Joseph, et al., "Evolutionary History and Adaptation from High-Coverage Whole-Genome Sequences of Diverse African Hunter-Gatherers," Cell 150(3):457–469, 2012.
Why it is quarantined rather than cited. Quoting ICR quoting Lachance is exactly the relay this project's rules warn against. The figure is used in the article only as ICR's own report of it, is labelled as such in the footnote, and appears again in "Issues and unanswered questions" as something a reader should check before relying on it further. Nothing in the article is attributed to Lachance.
What was tried. Nothing beyond noting the citation; the research pass deliberately left it for a second attempt.
What would change this. A publisher-PDF fetch through a headless browser, which is the route that worked for two other Elsevier papers used here. If the count checks out in Lachance, the sentence can be re-footnoted to the primary and this entry closed.
3. Jeanson's mapping of haplogroups onto named post-Babel peoples
Status: NOT FETCHED BY CHOICE.
What it is. Nathaniel T. Jeanson, Traced: Human DNA's Big Surprise (Master Books, 2022), the book-length statement of the Y-chromosome Babel model, and the place where haplogroups are mapped onto named peoples and migrations.
Why it was wanted. If any creationist source anywhere states a testable prediction about the geography of genetic diversity, this is the most likely place. A direct quotation about Africa specifically would have improved Part 5.
Why it was held back. The obtainable routes are an archive.org search-inside pass or a hoopla borrow, and hoopla allows ten borrows per calendar month across the whole project. Carter's 2022 Journal of Creation review of Traced, already in the source library, and Jeanson's own open Answers Research Journal papers carry the load-bearing claims, and neither engages the heterozygosity gradient either.
What would change this. A demonstration that Traced addresses African diversity or the origin scans in a way the open papers do not. On the evidence so far it does not, and the negative result reported in the article stands on a search of four creationist sites plus the general web.
4. Tomkins's concluding assessment of creationist genetics
Status: PRINTED WITH A FLAG, as a paraphrase.
The wording as it circulates. That the empirical genetics research done within the creationist community over the past ten years amounts to a sound dismissal of evolutionary Recent African Origins speculation.
Where it comes from. Jeffrey P. Tomkins, "Out of Babel — not Africa," Journal of Creation 34(1):79–85, 2020, closing section.
Why it is not in quotation marks in the article. The research dossier records this sentence in its summary of the paper's argument rather than in its list of verified quotations with a locator, and the creation.com web version carries no page marks, so the printed-page locator could not be anchored. Under this project's rule that every quotation in an article must appear with its locator in a dossier, it is paraphrased instead. The substance is not in doubt; the citation discipline is.
What would change this. Re-reading the stored HTML capture at sources/web/tomkins-2020_out-of-babel-not-africa.html and recording the sentence with its section locator. Then it can be quoted directly.
5. Ramachandran's published Figure 4A regression
Status: RELEASED, 6 September 2026. A sweep of 2,304 possible per-region waypoint assignments found the one that fits. Route western and southern Asia out of Africa through Cairo then Istanbul, eastern Asia and Oceania through Cairo then Phnom Penh, and the Americas on through Anadyr and Prince Rupert: the regression from Addis Ababa then gives R-squared 0.7762 with a slope of −6.611 × 10⁻⁶ against the published 0.7630 and −6.52 × 10⁻⁶. That convention in turn fails the paper's Figure 5 lattice statistics, which the article's own convention matches, so both are reported. Code at research/code/robustness.ts. The record of the original problem follows.
The wording, which is genuine and is quoted in the article. "The equation of the regression line is heterozygosity = 0.7682 − (6.52 × 10⁻⁶) × (distance from Addis Ababa). R² = 0.7630."
The problem. This project's reproduction, running the same regression on the same public genotypes with waypoint-routed distances, returns R-squared 0.8593 with a slope of −7.567 × 10⁻⁶ and an intercept of 0.7745. The paper's Addis Ababa distances are therefore about 16% longer than this script's. The paper states its five waypoints but not the routing rules between them, which is where the difference must lie. That guess proved correct; see the status line above.
Why the article prints both rather than suppressing one. The scan machinery the article actually uses reproduces the paper's published Figure 5 statistics closely, including the exact argmax at 4.3N, 12.8E, and every conclusion holds under all four routing conventions now tried, including none at all. Hiding a failed reproduction of a number the article quotes would be worse than reporting it. Note also that the published 0.7630 sits near the bottom of that paper's own range for African origins, which is an oddity in the paper rather than in the reproduction.
What would change this. The routing code behind the 2005 paper, or a statement of the inter-waypoint rules. Neither appears to be published.