---
description: Research dossier for Eric's request "Try to find more evidence to support young ages" for mitochondrial Eve.  Tests new leads not already in the three mtDNA articles, each checked against its primary source, priced numerically and graded for how much it helps a young age.
created: 2026-10-27
agents: Claude Code (research subagent)
models: claude-opus-5-5
assigned: [Eric]
---

# Young-age evidence for mitochondrial Eve:  new leads, tested

This file is a research dossier only.  No article was edited.  It is written for the agents merging `mitochondrial-eve`, `mitochondrial-eve-2` and `mitochondrial-eve-astra` tonight (plan:  `research/merge-plan-2026-10-27.md`, section 8).

## Baseline used for every price below

These are the numbers already established in `agents/memory/mitochondrial-eve.md`;  they are not re-derived here.

1. Requirement:  Behar 2012 mean root-to-tip **57.1** substitutions;  a 6,000-year Eve at 25-year generations is 240 generations, so **0.238 retained changes per genome per generation**.
2. Measured fixing (retained-change) rates:  Connell 2022 fixed-only 1.30e-8/site/yr (about 0.0054 per genome per 25-yr generation);  Soares 2009 1.665e-8 (about 0.0069);  aDNA tip rates 2.1-2.7e-8 (0.0087-0.0114).  Shortfall against these:  21x to 44x.
3. Measured arising rates:  Wei 2019 0.26/gen;  Arbeithuber 2025 0.743/gen;  Jeanson 2015 0.158/gen.
4. Already in the articles, so not redone here:  the arising-versus-fixing joint;  Norfolk Island;  deCODE;  Arbeithuber;  maternal age;  Henn 2009's 5x/2x;  Subramanian & Lambert 2011;  Molak & Ho 2015 (cited, slope not priced);  Cabrera 2021;  Endicott & Ho 2008 (162 to 108 ka);  Rieux S13 consistency;  Ötzi;  our 229-genome regression;  saturation (Soares S3, deCODE);  clade-rate ceiling about 3x;  created heteroplasmy and burst scenarios (priced in EVE2 lines 389-420);  Trifunovic 3-5x;  Stewart 2008 purifying selection;  the Neanderthal joint.

## Lead list

1. Jeanson's own stated predictions (ARJ 2013, 2015, 2016) and whether anything published since has tested them.
2. Mutator-polymerase magnitude:  is Trifunovic's 3-5x the right figure for what a single nuclear polymerase defect can do to the mtDNA mutation rate?  (Vermulst 2007, Kennedy 2013, Ross 2013, Stewart 2008 germline counts.)  Price a mutator burst against the tree.
3. How far time-dependency reaches:  the fitted slope in Molak & Ho 2015 and the curves in Ho 2005, 2011, 2015;  price what a 6,000-year calibration window would buy if the metazoan curve applied to humans.
4. Henn 2009's youngest calibration points:  what rate do calibrations that fall inside the young-earth window give?
5. Historically documented matrilines (Richard III, the Romanovs):  multi-century pedigrees dated by records, not radiocarbon, inside the young-earth window.
6. Germline studies 2020-2026 not yet in the articles.
7. Nuclear genetic control of mtDNA mutation and heteroplasmy (Gupta 2023 and similar):  heritable rate variation as a basis for past rate change.
8. Ancient-DNA tip-dating assumptions not already covered (tree prior, oldest-sample leverage).
9. Creationist primary work not yet engaged:  Carter 2008 Eve consensus predictions, Wood, Criswell, Sanford 2018.

Grades:  **strong** = a reader who checks the source will agree it supports a younger Eve;  **moderate** = real and supportive but bounded well short of what a 6,000-year Eve needs;  **weak** = true but does little;  **fails** = tested and goes the other way.

## Lead 1.  Jeanson's own predictions, and what has tested them since

**Already done elsewhere.**  `research/jeanson-case.md` reproduces every Jeanson 2013/2015/2016 number and grades his four-genus test (Homo inside, Caenorhabditis 9% outside, Drosophila inside a 22-fold range, Daphnia fails 1.5-2x) and the yeast failure (Jeanson & Lisle 2016).  It does not say which of his numbers were later re-measured, and it misses one methodological point in his favour.  Both are added here.

### 1a.  A point in Jeanson's favour that the articles do not make

**What he measured.**  Jeanson 2013 (*ARJ* 6:467-501) compared within-genus mtDNA diversity in *Caenorhabditis*, *Drosophila* and *Daphnia* against mutation rates from laboratory mutation-accumulation (MA) lines:  Denver 2000, Haag-Liautard 2008, Xu 2012.  His Table 6 (p. 495, anchored on the printed running head of PDF page 29; "Unrealistic mutation rate requirements for the evolutionary timescale"), years per mutation, evolution / creation / actual:  *Caenorhabditis* 21,752 / 12 / 6-11;  *Drosophila* 34,247 / 17 / 9-202;  *Daphnia* 23,259 / 31 / 48-62;  *Homo* 36,364 / 2,020 / 1,244-2,915.

**Why this matters.**  The article's whole case against the human clock is that pedigree studies count mutations *arising* (heteroplasmies at a few percent), while phylogenies count mutations *retained*.  MA lines are different in kind.  Each line is passed through one individual per generation for hundreds of generations (Haag-Liautard 2008:  "an average of 262 and 187 generations of MA";  Denver 2000:  "approximately 215 consecutive generations";  Konrad 2017:  409), so most mutations that are counted have had time to drift to fixation or loss inside the line.  Under Kimura's identity, a neutral mutation fixes inside a single-individual line at the rate it arises.  **So Jeanson's invertebrate comparison largely escapes the arising-versus-fixing objection that sinks his human clock.**  That is a fair thing to say for him, and none of the three articles says it.

**Limits.**  (1) The escape is partial:  Konrad 2017 counted "nine base substitutions in frequencies ranging from 3% to near fixation" (Results, "Spontaneous Mitochondrial Mutation Rates and Spectrum in C. elegans"), so MA counts still include some heteroplasmies.  (2) MA lines suppress purifying selection, which is the point of the design;  Haag-Liautard's own Discussion says their high rate "largely comes from mutations at nonsynonymous major-strand G sites;  these are subject to strong purifying selection in nature, and this contribute little to between-species divergence" (quoted from `research/clock-tests-in-other-species.md`, verified there against the Europe PMC XML).  So the MA rate overstates the natural neutral substitution rate by whatever fraction of mutations selection removes.  (3) The invertebrate test compares *species* (six *Drosophila* species;  *C. elegans* with *C. briggsae*), which requires those species to share one created sequence 6,000-10,000 years ago.  That is a separate claim with its own nuclear evidence, and Jeanson & Lisle 2016's nuclear control already fails it by "nearly an order of magnitude or more" (verbatim in `jeanson-case.md`).

### 1b.  The one Jeanson rate that has been re-measured since went against him

**Source.**  Konrad A, Thompson O, Waterston RH, Moerman DG, Keightley PD, Bergthorsson U, Katju V.  "Mitochondrial Mutation Rate, Spectrum and Heteroplasmy in *Caenorhabditis elegans* Spontaneous Mutation Accumulation Lines of Differing Population Size."  *Molecular Biology and Evolution* 34(6):1319-1334 (2017).  doi:10.1093/molbev/msx051.  https://pmc.ncbi.nlm.nih.gov/articles/PMC5850408/ .  Library `sources/papers/konrad-2017_mtdna-mutation-rate-elegans-ma-lines.{html,txt,yaml}` (copyright), read by curl of the PMC HTML on 2026-10-27.  (The quotations in `clock-tests-in-other-species.md` were WebFetch-extracted;  the three below are now verified against the bytes.)

**What it measured.**  20 MA lines, 409 generations, about 97% of the mtDNA.

**Quotes (1 from Results, "Spontaneous Mitochondrial Mutation Rates and Spectrum in C. elegans";  2 and 3 from Discussion):**

1. "The N = 1 MA lines contained nine base substitutions in frequencies ranging from 3% to near fixation (table 1), yielding a spontaneous rate of base substitution, μbs, of 4.32 × 10⁻⁸ per nucleotide site per generation (95% CI: 0.87 to 9.44 × 10⁻⁸)."
2. "We note the existence of a greater than two-fold difference in the mtDNA μbs estimates from our analysis of C. elegans MA lines relative to that of Denver et al. (2000)."
3. "Because we were able to detect mutations present at lower frequencies than Denver et al. (2000) due to greater depth of sequencing, the discrepancy between our mtDNA mutation rate estimates is even larger than it appears."

**Price.**  Back-calculating from Jeanson's Table 6 (6-11 years per mutation, a genome of about 14,100 bp averaged over *C. elegans* and *C. briggsae*, and his 3.5-4-day generations, i.e. 91-104 generations a year) gives an input rate of about 7-11 × 10⁻⁸ per site per generation, which is Denver's value (Konrad's range of published rates tops out at Denver's 9.7 × 10⁻⁸).  Replacing it with Konrad's 4.32 × 10⁻⁸:  4.32e-8 × 14,100 × 91-104 = 0.055-0.063 mutations per genome per year, i.e. **16-18 years per mutation against the 12 his model requires**.  The *Caenorhabditis* "match" turns from 9% outside to 1.3-1.5x short, at laboratory generation times;  wild *C. elegans* breeds far less often than every four days, which makes it worse.

**Skeptic's reply.**  The better measurement moved against the prediction, and Konrad says the true gap is "even larger than it appears."  Two of Jeanson's four genera now fail (Daphnia, Caenorhabditis), one passes inside a 22-fold window (Drosophila), and the human pass rests on the arising rate.

**Grade.**  1a:  **moderate** as a point of method (it should be stated for him);  1b:  **fails** as evidence for a young age.  No other Jeanson 2013-2016 prediction has a later independent measurement that I could find:  the human one is the Ding 2015 rate, superseded by Connell 2022 and deCODE 2024 (already in the articles), and nobody has re-measured *Drosophila* or *Daphnia* mtDNA MA rates since Haag-Liautard 2008 and Xu 2012 on the evidence of Konrad's own 2017 review of the field.

## Lead 2.  A polymerase mutator is about a hundred times stronger in the germline than the articles say

**This is the most useful new finding in this dossier for the young-earth side, and it corrects a number in the original article.**

**What the articles say now.**  The original article prices a proofreading-defective polymerase at Trifunovic 2004's "threefold to fivefold increase in the levels of point mutations" (footnote `trifunovic-2004:a`) and then cites Stewart 2008 only for purifying selection.  `mitochondrial-eve-archaic.md` repeats "Trifunovic 2004 abolishing PolgA proofreading gives 3-5x."  EVE2 says a burst needs 2.86 retained changes per generation for 500 years and calls these "requirements of the stated scenarios, not measured ancestral rates."  Nobody has put Stewart's own germline counts next to that requirement.

### What Stewart 2008 actually measured in the germline

**Source.**  Stewart JB, Freyer C, Elson JL, Wredenberg A, Cansu Z, Trifunovic A, Larsson N-G.  "Strong Purifying Selection in Transmission of Mammalian Mitochondrial DNA."  *PLoS Biology* 6(1):e10 (2008).  doi:10.1371/journal.pbio.0060010.  https://pmc.ncbi.nlm.nih.gov/articles/PMC2214808/ .  Library `sources/papers/stewart-2008_purifying-selection-mtdna.html` (already held);  re-read by curl of the PMC HTML on 2026-10-27.

**Design (Results, first paragraph).**  "We used eight mtDNA mutator founder females to establish independent maternal lines through 13 F1 females."  The founders were homozygous for proofreading-dead PolgA;  their daughters (N1) were heterozygous;  from N2 on, only wild-type PolgA females were bred.  So every mutation scored was made in the germlines of at most three mutator-carrying females in a row:  the founder's own (heterozygous) mother, the homozygous founder, and her heterozygous N1 daughter.  (The paper's Figure 1 note:  "The PolgA mut/PolgA mut mtDNA mutator females were derived from mothers with wild-type C57Bl/6 mtDNA.")  Heterozygotes mutate far less than homozygotes (somatically about one-fifth, 500x against 2,500x;  see below), so this is roughly 1.4 homozygous-equivalent generations.  Detection (Materials and Methods):  "The software identified heteroplasmy or substitutions at ≥25% signal intensity on both strands of DNA sequence."

**The numbers (Results, first paragraph), verbatim:**

> "We sequenced the entire mtDNA of 190 animals from generations N2 to N6 and identified 1,069 unique mutations (Dataset S1).  The typical animal carried approximately 30 mtDNA mutations (mean = 29.8 mutations, standard deviation = 9.2).  In each line, a large proportion of the identified mtDNA mutations (38.48%) were transmitted to the descendents of that particular N1 female, similar to the propagation of mtDNA haplogroups in human pedigrees."

**And the selection result that comes with it:**  dN/dS 0.6035 over all mutations, 0.4617 once single-animal and single-litter mutations are removed;  and (Results, the Figure 2 paragraph) the codon-position pattern matches humans' "2.0-fold reduction" vs "2.6-fold reduction", "surprising because the mtDNA mutator strains have undergone selection for at most six generations."  The abstract:  "There are striking similarities between the mutational patterns in our experimental mouse system and human mtDNA polymorphisms."  (That sentence refers to the codon-position selection pattern, not to the substitution spectrum;  do not quote it as a spectrum match.)

### The price

A neutral variant at frequency f eventually fixes along a followed lineage with probability f (the article's own Σf rule).  Every one of Stewart's 29.8 mutations per animal sits at f ≥ 0.25.

| assumption | eventual fixations per mutator generation | generations to supply all 57.1 root-to-tip changes | years at 25 yr/gen |
|---|---|---|---|
| all at f = 0.25, spread evenly over 3 mutator generations (most conservative) | 2.5 | 23 | 575 |
| all at f = 0.25, spread over 2 mutator generations | 3.7 | 15 | 380 |
| all at f = 0.25, made in 1 generation | 7.5 | 8 | 190 |
| mean f = 0.45, 2 generations | 6.7 | 9 | 210 |
| Trifunovic 3-5x on a fixing rate of 0.0054-0.0114 (the articles' figure) | 0.016-0.057 | 1,000-3,500 | 25,000-88,000 |

So a single known nuclear defect, in mice, makes in one to three germline generations about half of everything a human lineage carries from root to tip.  The burst that EVE2 priced at 2.86 retained changes per generation for 500 years is **of the same order** as what this mechanism delivers:  2.5-7.5 per generation before purifying selection is discounted, roughly 2-5 after a 25-50% discount taken from the dN/dS figures.  Measured another way:  roughly 15 mutations per mutator generation at ≥25% frequency against deCODE's 0.0703 per human transmission at a far looser 5% threshold is more than 200-fold.

The somatic figures point the same way.  Vermulst 2007 (*Nature Genetics* 39:540, abstract via PubMed, library `vermulst-2007`):  "a mitochondrial mutator mouse was able to sustain a 500-fold higher mutation burden than normal mice, without any obvious features of rapidly accelerated aging."  The genotype split, from the contemporary news report (Phillips, *The Scientist*, March 2007, library `phillips-2007`):  "Mice homozygous for the mutant gene have 2,500 times the mitochondrial mutations as wild-type mice and also show reduced lifespan and several features of premature aging.  However, mice with only one copy of the mutated gene showed no signs of premature aging, even though they have 500 times as many mitochondrial mutations as wild-type mice."  The difference from Trifunovic's 3-5x is mostly the assay:  Vermulst's random-mutation-capture baseline is "more than ten times lower than previously reported," so the fold-change is larger.  Use the germline count, not either somatic fold-change, in an article.

### Why it still does not deliver a 6,000-year Eve

1. **Cost.**  Ross 2013 (*Nature* 501:412, abstract via PubMed, library `ross-2013`):  "maternally transmitted mtDNA mutations can induce mild ageing phenotypes in mice with a wild-type nuclear genome.  Furthermore, maternally transmitted mtDNA mutations lead to anticipation of reduced fertility in mice that are heterozygous for the mtDNA mutator allele … Unexpectedly, a combination of maternally transmitted and somatic mtDNA mutations also leads to stochastic brain malformations."  Eight to fifteen generations of a homozygous mutator in every woman is a population carrying a growing, inherited load with falling fertility.
2. **Every lineage at once.**  The 57.1 is a mean over all lineages, so the mutator must run in every surviving maternal line in parallel, and only after those lines have split (EVE2's point:  a change fixed before the split is shared, not a difference).  It must also run on the Neanderthal and Denisovan branches to the same degree (the measured branch ratios are 1.01 and 0.96), then vanish from every population.  No exonuclease-dead POLG allele is common in any living population.
3. **Anchors.**  A burst common to all lineages is invisible in the tree's shape (it only rescales time), so only dated samples can see it.  Rieux's 21 single-sample calibrations agree within CV 3.2% across 2,535-39,464 radiocarbon years, so after the oldest sample the mtDNA rate would have to fall in lockstep with whatever inflates radiocarbon ages:  two unrelated processes, one curve.  Ötzi (dendro-anchored) caps the post-5.3-ka rate.
4. **No evidence it happened.**  This establishes that the mechanism is big enough.  It is not evidence that it operated.  The direct Neanderthal POLG find (Leu580Pro, outside the exonuclease domain, branch 1.01x) is the only candidate on file, and it fails.

**Skeptic's best reply.**  "You have found a way to break the clock, not evidence that it was broken.  The cost is a population-wide, multi-generation mitochondrial disease that leaves no allele behind and must stop exactly when the radiocarbon record says it should."

**Grade.**  **Moderate.**  It moves the burst rescue from "no known mechanism of that size" to "a known mechanism of that size exists;  its occurrence is unevidenced and costly."  **Article action:**  the merged article should replace the bare "3-5x" with Stewart's germline count, or say plainly that 3-5x is a somatic fold-change and not the germline output.  Leaving it as is understates the young-earth case by about two orders of magnitude, which a checking reader will find.

**Footnote-ready.**
- Stewart, James B.;  Freyer, Christoph;  Elson, Joanna L.;  Wredenberg, Anna;  Cansu, Zekiye;  Trifunovic, Aleksandra;  Larsson, Nils-Göran.  "Strong Purifying Selection in Transmission of Mammalian Mitochondrial DNA."  *PLoS Biology* 6(1):e10, 2008.  doi:10.1371/journal.pbio.0060010.  https://pmc.ncbi.nlm.nih.gov/articles/PMC2214808/ .  Quote locator:  Results, first paragraph ("We sequenced the entire mtDNA of 190 animals …").
- Ross, Jaime M.;  et al.  "Germline mitochondrial DNA mutations aggravate ageing and can impair brain development."  *Nature* 501(7467):412-415, 2013.  doi:10.1038/nature12474.  Abstract.
- Vermulst, Marc;  et al.  "Mitochondrial point mutations do not limit the natural lifespan of mice."  *Nature Genetics* 39(4):540-543, 2007.  doi:10.1038/ng1988.  Abstract.

## Lead 3.  How far time-dependency reaches, and what it can buy

**Already in the articles.**  Molak & Ho 2015 is cited for "a significant negative slope" across 239 estimates;  Henn 2009's 5x/2x;  Subramanian & Lambert 2011;  Cabrera 2021's 2.27x;  Peterson & Masel's ~2x over ~8,400 yr.  **Not in the articles:**  the fitted slopes themselves, the human-only slope, and what each slope buys when priced against the requirement.  This matters because a young-earth reader can take Molak & Ho's own curve and extrapolate it, and on one reading it delivers the whole requirement.

**Source.**  Molak M, Ho SYW.  "Prolonged decay of molecular rate estimates for metazoan mitochondrial DNA."  *PeerJ* 3:e821 (2015).  doi:10.7717/peerj.821.  PMC4358697.  CC-BY.  Library `sources/data/molak-2015_prolonged-decay-metazoan-mtdna/` (main text `molak-2015-fulltext.txt`;  supplements added today from the Europe PMC `supplementaryFiles` zip, since the PMC `bin/` links and peerj.com both return a browser challenge to curl).

**What it measured.**  239 published, re-estimated and new mtDNA rate estimates, each regressed (log-log) against the age of the calibration used to make it.

**Quotes.**
1. Results:  "We found significant correlations between these two variables in coding markers (slope = −0.28, p = 1.58 × 10⁻¹², R² = 0.24) as well as non-coding markers (slope = −0.43, p = 1.38 × 10⁻¹², R² = 0.60)."
2. Methods:  "For all data subsets tested, the randomized data produced an average regression slope of −1, which is the value expected when there is spurious self-correlation (Kenney, 1982)."
3. Discussion:  "We observed this pattern across the entire timescale that was analysed, beyond at least 10⁶ years in non-coding markers and 10⁷ years in coding markers … This is in contrast with previous estimates of the temporal depth of time-dependence, which has been variously estimated at a few hundred years (Richards M in Gibbons, 1998), around 50 kyr (Henn et al., 2009), 200 kyr (Burridge et al., 2008), or 1–2 Myr (Ho et al., 2005 …)."
4. Supplementary article (`supp-s001-human-case-study.txt`, "Intraspecific time-dependent rates – human case study"):  "This case study included 26 estimates of rates in the human mitochondrial control region … Calibrations … ranged from 25 years (one generation) to 12.5 Myr … Regression analysis produced strong evidence of a negative relationship between rate estimates and calibration times (slope -0.24, p = 1.51×10-7, R2 = 0.69; Figure S1, Table S4)."  Table S3 (`supp-table-s3-regressions.docx`) gives the human slope's standard error as 0.03 and rejects slope = −1 at p = 5.42 × 10⁻¹⁸.
5. Table S4 (time slices, non-coding):  slopes −0.34, −0.36, −0.48, −0.49, −0.58, −0.53 in the successive windows from 10¹-10⁵ to 10⁶-10⁹ years.

**The price.**  If the rate estimate scales as (calibration age)^b, moving the calibration from age T to 6,000 years multiplies the rate by (T/6,000)^−b.  The requirement is 21-44x the fixing/tip rate (34x against Soares' chimp-calibrated whole-genome rate).

| slope b (source) | chimp calibration 6 Myr → 6 ka | Eve-depth calibration 150 ka → 6 ka | aDNA calibration 30 ka → 6 ka |
|---|---|---|---|
| −0.24 ± 0.03 (humans, control region, n = 26) | 5.2x (3.5-7.9) | 2.2x | 1.5x |
| −0.28 (all metazoa, coding) | 6.9x | 2.5x | 1.6x |
| −0.43 (all metazoa, non-coding) | 19.5x | 4.0x | 2.0x |
| −0.48 to −0.58 (non-coding, deep time slices) | 27-55x | 4.7-6.5x | 2.2-2.5x |
| **needed for a 6,000-yr Eve** | **b = −0.44 to −0.55** | **b = −0.95 to −1.18** | b < −1.8 |

**What this shows, for the young-earth side.**  Take the all-animal control-region slope and extrapolate from the chimpanzee calibration straight down to a 6,000-year window, and the mainstream meta-analysis's own curve supplies 20-55x:  the whole requirement.  That is a real, citable, peer-reviewed curve, and a young-earth writer will eventually find it.  The articles should state it before he does.

**Why it fails anyway.**
1. **Humans have their own slope, and it is shallow.**  Molak & Ho fitted human data separately:  −0.24 ± 0.03, which buys 5x from the chimp calibration, not 34x.  The −0.43 to −0.58 slopes come from other animals' d-loops (birds, bears, bison), which saturate far more than a human tree only 57 substitutions deep.
2. **Self-consistency.**  If Eve is 6,000 years old, the right rate is one calibrated inside 6,000 years, and those exist for humans:  Holocene ancient DNA (Brotherton 2013, 2.4 × 10⁻⁸), Cabrera 2021's youngest bin (4.33 × 10⁻⁸ over ~1,100 years), and Henn 2009's coding-region rate from Polynesia (Lead 4).  None is within 10x of the requirement.  The extrapolated curve cannot overrule direct measurements made at the timescale it is being extrapolated to.
3. **A slope of −1 means no accumulation at all.**  From Eve-depth or aDNA calibrations, the requirement is b ≤ −0.95, i.e. a sequence 150,000 years old would carry no more differences than one 6,000 years old.  Molak & Ho use −1 as the signature of a meaningless artefact ("spurious self-correlation") and reject it for the human data at p = 5 × 10⁻¹⁸;  and the radiocarbon-dated branch-shortening data (Rieux S13, our 229-genome regression) show differences accumulating roughly linearly with sample age.

**Skeptic's reply in one line.**  "You borrowed a penguin's slope for a human, and you extrapolated it over a window where human measurements already exist."

**Grade.**  **Weak** as evidence;  **worth stating** because the extrapolation is the strongest-looking form of the time-dependency argument and nobody has priced it in print.

## Lead 4.  Calibrations that fall inside the young-earth window

**Already in the research, not the article.**  `research/phylogenetic-rates-and-calibration.md` (finding 4 and the Rieux Table 2 section) records that Henn 2009's fourteen HVR1 rates fall monotonically with calibration age, and that Rieux 2014's tip-calibrated clock overshoots every historically young colonisation by 3-6x (New Zealand 6.6x, Remote Oceania 4.9x, Canary Islands 3.9x, Madagascar 3.3x), concluding "the only calibration checks with genuinely known dates all show the clock r[unning fast]."  This is the strongest internal-calibration material on file for the young-earth side, and the merged article should carry it.  Two numbers it does not carry are added here.

**Source.**  Henn BM, Gignoux CR, Feldman MW, Mountain JL.  "Characterizing the Time Dependency of Human Mitochondrial DNA Mutation Rate Estimates."  *Molecular Biology and Evolution* 26(1):217-230 (2009).  doi:10.1093/molbev/msn244.  Library `henn-2009b_time-dependency-human-mtdna-rate.pdf` (Wayback copy of the OUP PDF).  Page numbers anchored on the printed running heads, checked on PDF pages 4, 6 and 7 (printed 220, 222, 223).

**Quote 1, HVR1 (p. 222, Results, "Trends in Genealogy-Based HVRI Mutation Rate Estimates"):**  "Genealogy-based rate estimates between 2,500 and 5,000 years ago are indistinguishable from pedigree-based mutation rate estimate (tables 1 and 2).  This remains true across different haplogroups and populations from vastly different geographic locations, for example, Canary Islands versus Samoa."

**The numbers (Table 2, p. 220,** ρ-based divergence rates per bp per Myr, calibrated on archaeological first arrival):  Canary Islands U6b1 0.81 (2.5 ka);  Samoa B4a1a1 0.74 (2.8 ka);  Nauru 1.03 (4 ka);  Palau 0.89 (4.2 ka);  Marianas 1.09 (4.3 ka);  Paleo-Eskimo A2 0.81 (5 ka).  Mean 0.90, against Howell's pedigree 0.95.  **For the control region, every calibration inside the young-earth window gives the pedigree rate.**

**Quote 2, coding region (p. 223, "Coding Region Results"):**  "The Polynesian-specific B4a1a1 haplogroup was the youngest haplogroup in our data set and gave an estimate of 0.046/bp/My.  Estimates from Native American haplogroups A2, B2, and D1 averaged 0.049/bp/My (n = 167) … There appears to be virtually no change in the coding region rate over the 10,000 year span between the Polynesian and American estimates."

**The price.**  Henn states that "All mutation rates discussed here will be in the form of divergence rates" (p. 218), so 0.046 per bp per Myr of divergence is a mutation rate of 2.3 × 10⁻⁸ per site per year, the same as the aDNA tip rates (2.1-2.7 × 10⁻⁸), about 0.0095 per genome per generation, and an Eve of roughly 150,000 years.  The control-region side, at the pedigree rate these young calibrations reproduce, gives the 15,200-year (mean pairwise) to 31,700-year (deepest split) ages already in the article.

**Skeptic's reply.**  (1) The fast control-region rates are founder-analysis artefacts:  a founder population imports diversity, which ρ counts as post-arrival mutation, so the youngest calibrations are biased upward, which is also the standard reply to Rieux's overshoots.  (2) The coding region, which carries about 93% of the sites and most of Eve's depth, shows no fast rate even at the youngest coding calibration (Polynesian B4a1a1, 13 sequences, arrival 4.3 ka, Table 3, p. 221:  0.0464).  (3) Henn's fast rates are confined to hypervariable sites that Lake 2024 shows are unconstrained and saturating.

**Grade.**  Control region:  **moderate** (a real, archaeologically anchored, inside-the-window match to pedigree rates;  quote Henn verbatim).  Coding region:  **fails**, and it is the coding region that decides the date.

## Lead 5.  A matriline documented by records for 500 years:  Richard III

**New;  not in any of the three articles.**  This was tested because it is the one kind of evidence a young-earth reader cannot dismiss on dating grounds:  the generations are counted from written genealogy, not radiocarbon, and the span (Cecily Neville, d. 1495, to living descendants) lies wholly inside any young-earth chronology.  It went against the young age.

**Source.**  King TE, Gonzalez Fortes G, Balaresque P, Thomas MG, Balding D, Maisano Delser P, et al.  "Identification of the remains of King Richard III."  *Nature Communications* 5:5631 (2 December 2014).  doi:10.1038/ncomms6631.  https://pmc.ncbi.nlm.nih.gov/articles/PMC4268703/ .  CC-BY.  Library `sources/papers/king-2014b_identification-remains-richard-iii.{html,txt,yaml}`, read by curl.

**What it measured.**  Complete mitochondrial genomes of Skeleton 1 (Richard III, d. 1485;  hybridisation capture, Illumina) and of two living female-line relatives through his sister Anne of York (long-range PCR, Ion Torrent, every difference from rCRS confirmed by Sanger "on both strands in both ML1 and ML2 in duplicate").

**Quotes (Results, section "Female-line relatives and mtDNA analysis";  quote 3 from Results, "Statistical analysis"):**
1. "we were able to obtain samples for comparison from Michael Ibsen (ML1), 19 generations removed from Richard III on the female line, and from Wendy Duldig (ML2), 21 generations removed from Richard III … Wendy Duldig and Michael Ibsen are female-line 14th cousins, twice removed (32 meioses)."
2. "These revealed a perfect whole-genome sequence match with ML1 and a single difference (position 8,994) with ML2 consistent with these individuals being matrilinear relatives over the time period considered."
3. Statistical analysis:  "The mtDNA sequences of Skeleton 1 and the presumed 19-meiosis matrilinear relative of Richard III, Michael Ibsen, matched completely.  A 21-meiosis relative also matched except at one base (8994)."

**The count.**  Three sampled genomes, joined through their mother line.  From the paper's three path lengths (19, 21, 32) the tree is:  Cecily → Richard (1 meiosis), Cecily → common female ancestor X of the two relatives (3), X → Ibsen (15), X → Duldig (17).  Check:  1+3+15 = 19, 1+3+17 = 21, 15+17 = 32.  **Total 36 documented transmissions;  one whole-genome difference observed.**  Sanger confirmation means a difference had to be at roughly 20% or more to be called, so this counts near-fixed changes, which is the quantity that matters.

**Price.**  Expected differences across the 36 transmissions, and the chance of seeing one or fewer:

| rate per genome per generation | expected | P(≤ 1 observed) |
|---|---|---|
| 0.743 Arbeithuber arising | 26.7 | 7 × 10⁻¹¹ |
| 0.26 Wei arising | 9.4 | 0.0009 |
| **0.238 needed for a 6,000-yr Eve** | **8.6** | **0.0018** |
| 0.158 Jeanson 2015 | 5.7 | 0.023 |
| 0.0703 deCODE raw (5% threshold) | 2.5 | 0.28 |
| 0.023 deCODE ÷ bottleneck | 0.83 | 0.80 |
| 0.0054-0.0114 Connell fixed-only to aDNA tip | 0.19-0.41 | 0.94-0.98 (P(≥1) = 0.18-0.34) |

Rate estimated from this pedigree alone:  1/36 = 0.028 per transmission, exact Poisson 95% interval 0.0007-0.155.  **The upper bound sits below both Jeanson's rate and the young-age requirement.**  It is one family and one mutation, so it cannot discriminate among the slow rates;  it does exclude the fast ones, in a pedigree dated by documents.

**Romanovs (checked, not usable).**  Coble MD et al. 2009, *PLoS ONE* 4:e4838 (PMC2652717):  Tsar Nicholas II carried a heteroplasmy at 16169 that his brother Georgij also carried "in differing ratios", while "his maternal relatives were fixed for 16169 T."  That is a transmitted heteroplasmy, i.e. an arising event in the middle of drifting, which illustrates the joint;  but the living-relative comparisons are control region only (HVI/HVII), so no whole-genome count can be made.

**Skeptic's reply.**  None needed;  this is the skeptic's evidence.

**Grade.**  **Fails** for a young age.  **Worth adding to the merged article** as a short paragraph after Norfolk Island:  it is the only multi-century, document-dated, whole-genome matriline in print, and its result (0 and 1 differences over 19 and 21 meioses) is easy for any reader to check.

## Lead 6.  Germline studies 2020-2026 not already in the articles

**Searched** (WebSearch, standard and extended, 2026-10-27) for human trio or pedigree studies of mtDNA germline mutation published 2020-2026 beyond Connell 2022, Árnadóttir (deCODE) 2024 and Arbeithuber 2025.  **Found nothing that changes the picture.**  Liu C et al. 2021 (*Mitochondrion*;  TOPMed preprint 2020, "Presence and Transmission of Mitochondrial Heteroplasmic Mutations in Human Populations of European and African Ancestry") counts heteroplasmy transmission, i.e. another arising-side measurement;  not read in full, no fixing rate expected.  Gupta 2023 and 2026 (*Nature*) are somatic blood studies (Lead 7).  The newest germline numbers remain the three already in the articles, and every one of them sits on one side or the other of the arising/fixing joint as the articles already describe.

**Grade.**  **Nothing new to add.**  A one-line "searched through October 2026" note in the merged article's pedigree table would pre-empt "you ignored newer studies."

## Lead 7.  Heritable variation in the mtDNA mutation rate

The young-earth burst (Lead 2) needs the mtDNA rate to have been very different in the past.  Is there any measured precedent for large, heritable differences in mtDNA mutation rate within one species, without engineering?

### 7a.  Daphnia magna:  a 90-fold spread among wild genotypes

**Source.**  Ho EKH, Macrae F, Latta LC IV, McIlroy P, Ebert D, Fields PD, Benner MJ, Schaack S.  "High and Highly Variable Spontaneous Mutation Rates in *Daphnia*."  *Molecular Biology and Evolution* 37(11):3258-3266 (2020).  doi:10.1093/molbev/msaa142.  https://pmc.ncbi.nlm.nih.gov/articles/PMC7820357/ .  CC-BY.  Library `sources/papers/ho-2020_high-variable-mutation-rates-daphnia.{html,txt,yaml}`.

**What it measured.**  Mutation-accumulation lines from nine genotypes of *D. magna* (three each from Finland, Germany, Israel), whole-genome sequenced;  mtDNA rate computed "assuming neutrality (Haag-Liautard et al. 2008) as µm,bs = ∑i fi/(g × n)", i.e. the Σf fixation-equivalent count, not a raw heteroplasmy count.

**Quotes (Results):**  "In the mtDNA, the variation among genotypes is even greater (spanning >3 orders of magnitude; fig. 1C and supplementary table S1A …) and the mean base substitution mutation rate is an order of magnitude higher than the previous highest mitochondrial mutation rate reported in animals using an MA approach (µm,bs = 8.7 × 10⁻⁷/bp/generation …)."  (Discussion):  "The intraspecific variation observed, as well as the evidence for nonzero levels of heritability of the mutation rate as a trait, highlights the potential for rapid evolution of the mutation rate among populations and species."

**The numbers (Table 1, mtDNA per bp per generation, × 10⁻⁷):**  FA 0.28, FB 9.37, FC 9.43, GA 10.77, GB 9.76, GC 25.46, IA 0.00, IB 4.63, IC 8.01.  **Honest reading:**  the ">3 orders of magnitude" depends on genotype IA's zero.  Among the eight genotypes with any mutations the spread is 0.28 to 25.46, **91-fold**;  seven of nine fall within 4.6-25.5 (5.5-fold);  and the table note says one MA line, GC8, "was an order of magnitude higher than other MA lines", so GC's top value is carried by a single line that may itself have picked up a mutator during the experiment.

**What it gives the young-earth side.**  A peer-reviewed, non-engineered precedent:  mtDNA mutation rate is a heritable genotype trait that can differ by roughly two orders of magnitude between natural genotypes of one species, and the authors themselves say it can evolve rapidly.  The 21-44x the young age needs is inside that range.

**Limits.**  (1) It is a crustacean, with the highest MA mtDNA rates recorded in any animal.  (2) No human pedigree study (Iceland, Norfolk Island, the UK, the USA, Sardinia) has reported a family-level mutator;  the spread among their fixing-side figures (about 0.0045 to 0.023 per generation) tracks detection threshold and the bottleneck conversion, not family.  (3) A young Eve needs a fast genotype in *every* human lineage at once and then its disappearance everywhere (Lead 2, point 2).  (4) MA lines suppress selection;  in nature a mutator genotype is selected against.

**Grade.**  **Weak to moderate**, as precedent only.

### 7b.  Nuclear genes measurably control human mtDNA heteroplasmy

**Source.**  Gupta R, Kanai M, Durham TJ, et al.  "Nuclear genetic control of mtDNA copy number and heteroplasmy in humans."  *Nature* 620(7975):839-848 (2023).  doi:10.1038/s41586-023-06426-5.  PubMed abstract read via efetch;  full text not saved.

**Quote (abstract):**  "heteroplasmic single nucleotide variants tend to arise somatically and accumulate sharply after the age of 70 years, whereas (2) heteroplasmic indels are maternally inherited as mixtures with relative levels associated with 42 nuclear loci involved in mtDNA replication, maintenance and novel pathways."

**What it gives.**  Human nuclear variation demonstrably shapes how mtDNA variants are inherited and drift, so "the nuclear background was different in the past" is not an empty phrase.  **What it does not give:**  any germline point-mutation effect, any effect size near 20x, or anything on substitution rates;  its point-mutation finding is somatic.

**Grade.**  **Weak.**

## Lead 8.  Creationist primary work not yet engaged:  Wood 2013 states the burst model, and rejects created heteroplasmy

### 8a.  Todd Wood, ICC 2013

**Status.**  `research/critics-of-the-yec-clock.md` (items 1d, 6a, 4 under gaps) records this paper as "not obtained" because the Cedarville repository migrated.  **It is now in the library:**  `sources/papers/wood-2013_mtdna-three-baramins.{pdf,md,yaml}` (retrieved 2026-08-15 from https://cedarville.tind.io/record/21104).  The articles cite it, if at all, only by title.  Its own words follow.

**Source.**  Wood, Todd Charles.  "Mitochondrial DNA Analysis of Three Terrestrial Mammal Baramins (Equidae, Felidae, and Canidae) Implies an Accelerated Mutation Rate Near the Time of the Flood."  *Proceedings of the Seventh International Conference on Creationism*, Pittsburgh, PA:  Creation Science Fellowship, 2013.  The PDF carries no printed page numbers;  quotations below are from the Discussion, PDF page 8 of 16.

**What he measured.**  Published mtDNA from extant and ancient (Pleistocene) horses, cats and dogs, with the ancient specimens' radiocarbon dates recalibrated into a post-Flood chronology.

**Quotes (Discussion, PDF p. 8):**
1. "the similarity of aDNA from Pleistocene specimens to modern mitochondrial DNA implies a rapid sequence diversification from the time of the Flood to the time the fossils were formed.  Since then, DNA divergence has been low, since these ancient sequences closely resemble their extant counterparts.  Thus, these results imply a period of rapid genetic divergence after the Flood that quickly decreased to the low mutation rates characteristic of the present."
2. "Alternatively, these results could be interpreted as the result of heteroplasmy in the animals aboard the Ark, wherein different mtDNA types were sorted into different lineages … Despite being a somewhat ad hoc explanation, this hypothesis could be tested by evaluating the relationships implied by morphological vs. molecular similarity.  In this model, there would be no expectation that the mtDNA similarities would reflect morphological similarity, since mtDNA types would be sorted randomly into different lineages.  Since the broad intrabaraminic relationships inferred from morphology do reflect mtDNA similarity, this relict mtDNA hypothesis can be ruled out, and instead, the hypothesis of rapid mtDNA divergence is supported."
3. On his own chronology:  "The recalibration used here places twelve of the specimens within seven decades of the Flood, but four of those specimens were dated to the Flood itself.  Since the Upper Pleistocene is unquestionably post-Flood, the recalibration used here cannot be precisely correct".

**Why it matters for the articles.**  (1) It is the young-earth burst model stated by a leading creationist biologist from his own analysis, so the merged article can state the burst rescue in a creationist's words rather than as a scenario of its own (the corrections log requires a model's position from its primary source before pricing it).  (2) Wood **concedes the present rate is low** ("the low mutation rates characteristic of the present"):  he does not use pedigree arising rates as the clock.  That is the opposite of Jeanson's move and agrees with the articles' central joint.  (3) Wood **tests and rejects the created-heteroplasmy ("relict mtDNA") alternative** that EVE2 presents as "a coherent starting-variation hypothesis".  A creationist has already argued against that escape.  (4) Lead 2 now supplies what Wood's paper lacks:  a known mechanism of the required size (a germline polymerase mutator), with its costs.

**Skeptic's reply.**  Wood's burst must end before his own Pleistocene specimens, whose radiocarbon ages he admits his recalibration cannot place correctly;  the human aDNA series (Rieux S13) shows the post-sample rate tracking radiocarbon age linearly, not a burst-then-slow curve.

**Grade.**  **Moderate** (as a primary statement of the strongest young-earth model, and as an internal creationist check on the heteroplasmy escape);  not evidence that the burst happened.

### 8b.  Carter, Criswell & Sanford 2008, the Eve consensus

**Source.**  Carter RW, Criswell D, Sanford J.  "The 'Eve' Mitochondrial Consensus Sequence."  *Proceedings of the Sixth International Conference on Creationism*, pp. 111-116 (2008).  Library `carter-2008_eve-mitochondrial-consensus-sequence.{pdf,md}`.  Page locators from the printed page numbers in the PDF.

**Quotes.**  Abstract (p. 111):  "On average, the individuals in our dataset differed from the Eve consensus by 21.6 nucleotides.  Sequences derived from sub-Saharan Africa were considerably more divergent than average."  Results (p. 113):  "On average, individuals differed from the consensus at only 22.6 positions, with sequences from sub-Saharan Africa varying at up to 89 positions."  (The paper gives both 21.6 and 22.6;  the Fig. 1 caption on p. 113 says 21.6.  Quote whichever is needed with its page, and note the discrepancy.)

**What it would give.**  A mean of ~22 from the root instead of Behar's 57.1 would cut the requirement to ~0.09 per generation, 2.6x easier.  **Why it does not:**  the 22 is a mean over a sample dominated by Eurasians, with Africans "up to 89";  a common clock gives every lineage the same time since the root, so the root-to-tip that has to be explained is the African one.  `research/root-topology-test.md` already shows re-rooting at R (this consensus, now also Jeanson 2025's root) raises the maximum root-to-tip from 75 to 104.  At 104 over 240 generations the requirement is 0.43 per generation, worse than Behar's 0.238.

**Grade.**  **Fails** as a young-age aid;  already handled in the root dossier, recorded here so nobody rediscovers "only 22 mutations from Eve."

## Lead 9.  Testing the created-variation escape:  the deepest branches carry the signature of mutation

EVE2 treats inherited (created) haplotype variation as the coherent young-earth option ("Suppose Eve carried two mitochondrial haplotypes … that already differed at many sites").  A created difference need not look like a mutation;  a mutational difference should carry mitochondrial mutation's strong transition bias.  This was checked directly.

**Method (this dossier, scratch script `/home/eric/scratch/strengthen-2026-10-27/mito/an/depth_spectrum.py`, using the project's verified PhyloTree Build 17 parser `research/code/phylotree17.py`).**  Every substitution in PhyloTree Build 17 (indels and parenthesised calls excluded;  PhyloTree writes a transversion with a lower-case derived base) was binned by how far from the RSRS root its branch starts.  A random base change is a transversion two times in three.

| branch starts this many substitutions below the root | substitutions | transversions |
|---|---|---|
| 0-9 | 29 | 3.4% |
| 10-19 | 139 | 6.5% |
| 20-29 | 158 | 3.8% |
| 30-39 | 871 | 4.8% |
| 40 and deeper (most of the tree) | 10,502 | 4.3% |
| all | 11,699 | 4.4% |

Deepest 326 substitutions (first 30 below the root):  coding 3.6% transversions (247), control 8.9% (79);  everything shallower:  coding 4.2%, control 4.8%.

**Reading.**  The differences between the deepest maternal lineages, the ones a created-variation model would have to supply (about 55 of 57.1 per lineage at the measured fixing rate), are 95% transitions, the same as the tips, where every young-earth author agrees the changes are mutations.  Had they been created as arbitrary functional variants, about two-thirds would be transversions.  So created variation is only compatible with the data if the created differences were made to imitate the mutational process in its base-change bias at every depth.  That is a coherent theological claim, but it is no longer a scientific prediction, and it removes the reason the escape was attractive.  (Caveat to print:  a design restricted to synonymous third-position changes would also favour transitions in the coding region, so the coding figure alone is not decisive;  the control region, which codes for no protein, shows the same bias in its deep substitutions, 91%.)

**This agrees with Wood 2013's own rejection of relict heteroplasmy (Lead 8a), reached by a different test.**

**Grade.**  **Fails** for the young-age escape;  new;  short enough for one paragraph and a footnote in the merged article.  **Reproducibility note:**  the script is scratch;  if the article cites it, copy it into `research/code/` and port it to Deno per the project rule.

## Lead 10.  Checked and already covered, so not redone

1. **Ancient-DNA tip-dating assumptions** (Fu 2013, Rieux 2014, Posth 2016):  `research/ancient-dna-calibration.md` and `research/phylogenetic-rates-and-calibration.md` already hold the verbatim methods, Rieux's S13 single-sample consistency (CV 3.2%), Rieux's 11x lower variance for tips than nodes, the Table 2 overshoots, Posth's four model combinations (2.62-2.77e-8), and the memory file's uniform-compression result (k = 26.2 puts Ötzi at about AD 1826).  Nothing new found.
2. **Rate variation among haplogroups:**  our 229-genome clade refit (U2/U5b 3.2x, fragile) and Översti & Palo are in the archaic memory file.  Nothing new found.
3. **Saturation and hotspots:**  Soares S3 (12% of variants carry 50% of mutation), deCODE's 19.7%/65.3% repeat test, Henn's K2P+Γ curves:  all in the articles.  Lead 3 adds only that the metazoan slopes are d-loop-dominated.
4. **Germline bottleneck and purifying selection:**  Wei 2019, Floros 2018, Wilton 2018, Zaidi 2019, Stewart 2008 (selection side) are all in the articles.  Lead 2 adds Stewart's germline counts.
5. **Emerson 2007** (*Syst Biol* 56:337), long recorded as unobtainable, is now in the library from a 2012 Wayback snapshot of the OUP full-text HTML:  `sources/papers/emerson-2007_alarm-bells-molecular-clock.{html,txt,yaml}`.  It argues against Ho 2005's time-dependency, so it bears on Lead 3 against the young-earth side;  not worked up here.  **Debruyne & Poinar 2009** has no usable Wayback copy (CDX shows only 301s).

## Ranked summary

### For a young age, strongest first

1. **A germline polymerase mutator is big enough to make the burst (Lead 2;  moderate).**  Stewart 2008's proofreading-defective mice passed on a mean of 29.8 mtDNA mutations per animal at ≥25% heteroplasmy, made in at most three mutator generations, 38% of them spreading through the descendant line.  By the Σf rule that is 2.5-7.5 eventual fixations per mutator generation, so 8-23 generations (about 200-600 years) would supply the whole 57.1 root-to-tip depth.  The original article's "3-5x" (Trifunovic, somatic) understates this by about a hundredfold, and should be corrected whatever else the merge does.  It shows the mechanism is large enough;  it is not evidence the mechanism operated.
2. **Inside the young-earth window, archaeologically calibrated control-region rates equal the pedigree rate (Lead 4;  moderate for the control region, fails for coding).**  Henn 2009, p. 222:  rates "between 2,500 and 5,000 years ago are indistinguishable from pedigree-based mutation rate estimate";  Table 2 mean 0.90 vs Howell's 0.95 per bp per Myr.  With the Rieux Table 2 overshoots already in `phylogenetic-rates-and-calibration.md`, this is the best inside-the-window calibration evidence on file.  The coding region at the same timescale (Polynesian B4a1a1, 0.0464) gives the slow rate.
3. **A creationist's own burst model, and his own rejection of created heteroplasmy (Lead 8a;  moderate as a primary statement).**  Wood 2013 infers "a period of rapid genetic divergence after the Flood that quickly decreased to the low mutation rates characteristic of the present" and rules out the "relict mtDNA hypothesis."  The merged article can state the burst rescue in a creationist's words and price it with Lead 2.
4. **Jeanson's invertebrate comparisons use the right kind of rate (Lead 1a;  moderate as method).**  Mutation-accumulation lines run for 187-409 generations, so their counts approximate retained changes;  the articles' central objection does not apply to them the way it applies to his human clock.
5. **Heritable mtDNA rate variation within a species (Lead 7a;  weak to moderate).**  *Daphnia magna* genotypes span 0.28-25.46 × 10⁻⁷ per bp per generation (91-fold among non-zero genotypes;  the paper's ">3 orders of magnitude" depends on a zero).
6. **The metazoan time-dependency curve, extrapolated from the chimp calibration, supplies 20-55x (Lead 3;  weak).**  Human data alone give 5x, and direct within-window human calibrations exist and are slow.

### Against a young age (new, and worth a sentence each)

1. **Richard III's documented matriline (Lead 5):**  36 transmissions, one whole-genome difference;  the young-age rate predicts 8.6 (P ≤ 1 observed = 0.0018);  95% upper bound 0.155 per transmission.
2. **The deepest branches carry mutation's transition bias (Lead 9):**  95% transitions in the first 30 substitutions below the root, coding and control alike, against 33% for arbitrary differences.  Created variation must imitate mutation to fit.
3. **The one Jeanson input re-measured since moved against him (Lead 1b):**  Konrad 2017 halves Denver's *C. elegans* rate;  the "match" becomes 1.3-1.5x short.
4. **Carter 2008's "21.6 from Eve" (Lead 8b)** is a Eurasian-weighted mean;  the African root-to-tip under that root is worse, not better.

### Footnote-ready citations (verified against downloaded bytes on 2026-10-27 unless stated)

- `stewart-2008`:  Stewart, James B.;  Freyer, Christoph;  Elson, Joanna L.;  Wredenberg, Anna;  Cansu, Zekiye;  Trifunovic, Aleksandra;  Larsson, Nils-Göran.  "Strong Purifying Selection in Transmission of Mammalian Mitochondrial DNA."  *PLoS Biology* 6(1):e10, January 2008.  doi:10.1371/journal.pbio.0060010.  https://pmc.ncbi.nlm.nih.gov/articles/PMC2214808/ .  Quote (Results, first paragraph):  "We sequenced the entire mtDNA of 190 animals from generations N2 to N6 and identified 1,069 unique mutations (Dataset S1).  The typical animal carried approximately 30 mtDNA mutations (mean = 29.8 mutations, standard deviation = 9.2).  In each line, a large proportion of the identified mtDNA mutations (38.48%) were transmitted to the descendents of that particular N1 female, similar to the propagation of mtDNA haplogroups in human pedigrees."  Detection (Materials and Methods):  "heteroplasmy or substitutions at ≥25% signal intensity on both strands."
- `ross-2013`:  Ross, Jaime M.;  Stewart, James B.;  Hagström, Erik;  et al.  "Germline mitochondrial DNA mutations aggravate ageing and can impair brain development."  *Nature* 501(7467):412-415, 19 September 2013.  doi:10.1038/nature12474.  PMC3820420.  Quote (abstract):  "maternally transmitted mtDNA mutations lead to anticipation of reduced fertility in mice that are heterozygous for the mtDNA mutator allele".
- `vermulst-2007`:  Vermulst, Marc;  Bielas, Jason H.;  Kujoth, Gregory C.;  Ladiges, Warren C.;  Rabinovitch, Peter S.;  Prolla, Tomas A.;  Loeb, Lawrence A.  "Mitochondrial point mutations do not limit the natural lifespan of mice."  *Nature Genetics* 39(4):540-543, April 2007.  doi:10.1038/ng1988.  Quote (abstract, via PubMed):  "a mitochondrial mutator mouse was able to sustain a 500-fold higher mutation burden than normal mice, without any obvious features of rapidly accelerated aging."  (Genotype split from Phillips, *The Scientist*, March 2007, `phillips-2007`;  full paper not read.)
- `henn-2009b`:  Henn, Brenna M.;  Gignoux, Christopher R.;  Feldman, Marcus W.;  Mountain, Joanna L.  "Characterizing the Time Dependency of Human Mitochondrial DNA Mutation Rate Estimates."  *Molecular Biology and Evolution* 26(1):217-230, 2009.  doi:10.1093/molbev/msn244.  Mirror https://web.archive.org/web/2017/http://mbe.oxfordjournals.org/content/26/1/217.full.pdf .  Quotes:  p. 222, "Genealogy-based rate estimates between 2,500 and 5,000 years ago are indistinguishable from pedigree-based mutation rate estimate (tables 1 and 2)";  p. 223, "The Polynesian-specific B4a1a1 haplogroup was the youngest haplogroup in our data set and gave an estimate of 0.046/bp/My";  Table 2 p. 220;  Table 3 p. 221;  units statement p. 218, "All mutation rates discussed here will be in the form of divergence rates".
- `wood-2013`:  Wood, Todd Charles.  "Mitochondrial DNA Analysis of Three Terrestrial Mammal Baramins (Equidae, Felidae, and Canidae) Implies an Accelerated Mutation Rate Near the Time of the Flood."  *Proceedings of the Seventh International Conference on Creationism*.  Pittsburgh, PA:  Creation Science Fellowship, 2013.  https://cedarville.tind.io/record/21104 .  Quotes:  Discussion, PDF p. 8 (no printed folios), as in Lead 8a.
- `molak-2015`:  Molak, Martyna;  Ho, Simon Y. W.  "Prolonged decay of molecular rate estimates for metazoan mitochondrial DNA."  *PeerJ* 3:e821, 2015.  doi:10.7717/peerj.821.  PMC4358697.  Quotes:  Results, slopes −0.28 and −0.43;  Methods, "the randomized data produced an average regression slope of −1";  Supplementary article s001, human slope −0.24 (n = 26, R² = 0.69);  Table S3, slope SE 0.03, H0 s = −1 rejected p = 5.42 × 10⁻¹⁸.
- `ho-2020`:  Ho, Eddie K. H.;  Macrae, Fenner;  Latta, Leigh C., IV;  McIlroy, Peter;  Ebert, Dieter;  Fields, Peter D.;  Benner, Maia J.;  Schaack, Sarah.  "High and Highly Variable Spontaneous Mutation Rates in *Daphnia*."  *Molecular Biology and Evolution* 37(11):3258-3266, 2020.  doi:10.1093/molbev/msaa142.  https://pmc.ncbi.nlm.nih.gov/articles/PMC7820357/ .  Quote (Results):  "In the mtDNA, the variation among genotypes is even greater (spanning >3 orders of magnitude …)";  Table 1.
- `konrad-2017`:  Konrad, Anke;  Thompson, Owen;  Waterston, Robert H.;  Moerman, Donald G.;  Keightley, Peter D.;  Bergthorsson, Ulfar;  Katju, Vaishali.  "Mitochondrial Mutation Rate, Spectrum and Heteroplasmy in *Caenorhabditis elegans* Spontaneous Mutation Accumulation Lines of Differing Population Size."  *Molecular Biology and Evolution* 34(6):1319-1334, 2017.  doi:10.1093/molbev/msx051.  https://pmc.ncbi.nlm.nih.gov/articles/PMC5850408/ .  Quotes as in Lead 1b.
- `king-2014b`:  King, Turi E.;  Gonzalez Fortes, Gloria;  Balaresque, Patricia;  Thomas, Mark G.;  Balding, David;  Maisano Delser, Pierpaolo;  et al.  "Identification of the remains of King Richard III."  *Nature Communications* 5:5631, 2 December 2014.  doi:10.1038/ncomms6631.  https://pmc.ncbi.nlm.nih.gov/articles/PMC4268703/ .  Quotes (Results, "Female-line relatives and mtDNA analysis"):  "Michael Ibsen (ML1), 19 generations removed from Richard III on the female line, and from Wendy Duldig (ML2), 21 generations removed … (32 meioses)";  "a perfect whole-genome sequence match with ML1 and a single difference (position 8,994) with ML2".
- `carter-2008`:  Carter, Robert W.;  Criswell, Dan;  Sanford, John.  "The 'Eve' Mitochondrial Consensus Sequence."  In *Proceedings of the Sixth International Conference on Creationism*, pp. 111-116, 2008.  https://www.icr.org/i/pdf/technical/The-Eve-Mitochondrial-Consensus-Sequence.pdf .  Quotes p. 111 (21.6) and p. 113 (22.6;  "up to 89").
- `jeanson-2013`:  Jeanson, Nathaniel T.  "Recent, Functionally Diverse Origin for Mitochondrial Genes from ~2700 Metazoan Species."  *Answers Research Journal* 6:467-501, 2013.  https://assets.answersresearchjournal.org/doc/v6/mitochondrial-genes.pdf .  Table 6, p. 495.

### Wanted sources (no JSTOR or hoopla used)

1. Stewart 2008 Dataset S1 (per-mutation heteroplasmy levels):  would replace the f ≥ 0.25 floor with an exact Σf.  On PLoS;  the PMC supplementary link needs a browser.
2. Vermulst 2007 full text (*Nature Genetics*):  to confirm the heterozygous 500x / homozygous 2,500x split from the paper rather than the news report.
3. Ross 2013 full text (PMC3820420):  fertility effect sizes per generation of inherited load.
4. Debruyne & Poinar 2009 (*Syst Biol* 58:348):  still unobtained;  no Wayback copy.
5. Penny 2005 (*Nature* 436:183-184, "Relativity for molecular clocks"):  not attempted;  paywalled.
6. Liu et al. 2021 (*Mitochondrion*, TOPMed heteroplasmy transmission):  for completeness of the 2020-2026 pedigree list.
7. Denver 2000 full text (*Science* 289:2342):  to settle which Denver rate Jeanson used (back-calculation says about 9.7 × 10⁻⁸, which Ho 2020 also quotes as Denver's).