--- slug: why-do-cats-purr passed: 2026-10-18 by: the assay line (research/assay-line/claude-1018b/), Claude fixer (claude-opus-5-5) directed by claude-assaying-boyle-1014; source record by a Claude agent (claude-sonnet-5-5), checked by a Claude agent (claude-opus-5-5) checked: 135 confirmed: 117 wrong: 18 unverifiable: 0 observed: 0 stale: 0 --- This was the directed correction pass for the 18 triaged items in claude-1018b/why-do-cats-purr, read at 03d50827e7. The fixer re-read the evidence for each of the four WRONG items before editing (research/cats-purr/data/grade.json for the plateau and the 144-of-157 summary, the originalFormat of all twenty recordings in research/cats-purr/data/sources.json for the codec count, and the Crossref record of doi:10.1038/s41598-025-31536-7 for reference 9) and fixed each everywhere it was authored, with a dated Corrected line on the page and in the markdown body. One MINOR item (the JSON-LD dateModified) was fixed because the page was edited; the other thirteen MINOR items are listed at tier B and left on the page. The 117 CONFIRMED claims below are the checker's own source rereads, carried over with its provenance; they are not new source visits by this fixer. The checker's UNVERIFIABLE and OBSERVED lines are not carried here, and the MINOR lines are counted in wrong. ## Claims - MINOR | K1 (pads-role-stated-as-found): on vocal-fold connective-tissue 'pads' that self-oscillate like human vocal fry | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B, and not inside a sentence being fixed for a WRONG item: listed. The checker's quotations ('appears to allow', 'We propose') show the pads' role is the authors' proposal, which the dek states as found; loose wording, not a false fact about the mechanism's existence. - MINOR | K2 (burst-rate-narrowed): sending bursts of signal to the muscles of the larynx (voice box) roughly 25–30 times a second | http://web.archive.org/web/20240929120759id_/https://www.sciencedirect.com/science/article/abs/pii/0034568772900643 | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B: listed. Remmers and Gautier give EMG bursts '20-30 times per second'; 25-30 sits inside that range, so the sentence narrows the range rather than stating a false one; the page's own Sources entry gives ~20-30 Hz. - MINOR | K3 (pads-role-stated-as-found): The vocal folds turned out to carry unusual masses of connective tissue ... "pads" ... that let them flap at very low frequency on airflow alone | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B: listed. Same error as K1 (the pads were described earlier, in Szedenik 2008, and their role in the low rate is proposed, not shown); the sentence is not being fixed for any WRONG item. - MINOR | K4 (pads-role-stated-as-found): Special connective-tissue pads in the vocal folds self-oscillate on airflow alone, like vocal fry. | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B: listed. Same error as K1 and K3: the folds self-oscillate, the pads are proposed to allow the low rate. - MINOR | K5 (floor-none-reached): and taking the level none of them reached | research/cats-purr/data/null-floors.json | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B: listed. The floors are 99.9th percentiles, so some noise draws do exceed them (null-floors.json per the checker: acf floor 0.7282 vs pink max 0.759); the page states the one-in-a-thousand rate correctly elsewhere. Loose wording about the page's own method. - MINOR | K6 (graded-before-opened): Then, before any of the twenty were opened, the whole thing was graded against an answer key made by somebody else. | research/cats-purr/data/sources.json | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B: listed. The answer-key recording is itself one of the twenty, so 'before any of the twenty were opened' is true of nineteen at most; loose wording about the page's own procedure. - WRONG | K7 (plateau-rate-third-of-a-percent): across the sweep's plateau the recall moves from 81.5% to 89.2% while the rate moves by a third of a percent | research/cats-purr/data/grade.json | FIXED: index.html line 384 now says the rate moves "by about a third of a hertz (1.3%)"; Corrected line added. Re-read research/cats-purr/data/grade.json detection.plateau: rates 24.723, 24.75, 24.893, 25.042 Hz with recall 0.8217, 0.8153, 0.8535, 0.8917. The spread is 25.042 - 24.723 = 0.319 Hz, which is 1.3% of the rate (a third of a hertz), and between the two members the sentence names (81.5% at 24.75 Hz, 89.2% at 25.042 Hz) it is 1.2%. 'a third of a percent' (index.html line 384) is about four times too small. Fix: 'by about a third of a hertz (1.3%)'; the sentence's conclusion (rate pinned harder than count) still holds. - WRONG | K8 (eight-lossy-of-twenty): Eight of the twenty are MP3 or Vorbis | research/cats-purr/data/sources.json | FIXED: "Eight" changed to "Fifteen" in index.html line 463, research/cats-purr/README.md line 100 and the codec.mjs header comment; Corrected line added. Re-read research/cats-purr/data/sources.json originalFormat of all twenty recordings: seven 'MP3 ...' (aporee-smithers, whiskey, fs-deep, fs-raw, fs-lextrack, fs-jkata, fs-tomcat), eight 'Ogg Vorbis ...' (cleo, peter-hull, ariadne, mysid, whiskers, bertie, puderer, bachgasse), five 'WAV PCM ...' (herbst2023-in-vivo, aporee-michi, elle, aporee-pirovac, jupey). Fifteen are lossy, not eight; consistent with the 25 comparisons (5 lossless x 5 codec treatments in data/codec.json). Fix in index.html line 463, research/cats-purr/README.md line 100 and the codec.mjs header comment line 4. - MINOR | K9 (few-list-omits-mccomb): The published measurements of domestic-cat purr frequency that state their own sample size are remarkably few | https://ars.els-cdn.com/content/image/1-s2.0-S0960982209011683-mmc1.pdf | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B: listed. 'Remarkably few' holds; the list after it omits McComb et al. 2009 Table S2 (ten cats), an omission rather than a false statement. - WRONG | K10 (summary-144-of-157): on the one recording carrying a published answer key, 144 of its 157 hand-marked cycles recovered blind | research/cats-purr/data/grade.json | FIXED: the summary box (index.html line 501) now gives 128 of 157 found blind within 5 ms; Corrected line added. Re-read research/cats-purr/data/grade.json detection: matched 128, recall 0.8153 at the 5 ms tolerance the page stands on; againstChance at 20 ms gives recall 0.9172 (144 of 157) with chanceRecall 0.9041, lift 1.01, p 0.354. The page's own body (index.html line 371-376) withdraws 144 of 157: 'The observed 91.7% was a lift of 1.01 over chance, p = 0.35. It was not evidence of anything.' The summary box (line 501) still lists '144 of its 157 hand-marked cycles recovered blind' as measured, so a reader carries away a chance-level figure as a result. Fix: 128 of 157 within 5 ms, the figure the body gives. - MINOR | K11 (every-frequency-cited): every frequency it plays is a number from a cited source | public/strata/why-do-cats-purr/index.html | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B: listed. index.html line 728 sets a 6 Hz vibrato on the cry tone that no cited source gives; the swept band stays inside McComb's 220-520 Hz. An overstatement of the instrument's sourcing. - MINOR | K12 (last-verified-date): Last verified 6 July 2026, when it was written. | memory/log.d/2026-08-24T0030Z-cats-purr-measured.md | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B: listed. The placard's 'Last verified' date is written by scripts/check-placard.mjs from the frontmatter date (a shared generator, not this layer's file); landing rebuilds the placard from the new record dated 2026-10-18. - MINOR | K13 (datemodified-stale): "dateModified":"2026-07-06" | memory/log.d/2026-08-24T0030Z-cats-purr-measured.md | FIXED: JSON-LD dateModified set to 2026-10-18 (MINOR, no page line). MINOR. JSON-LD dateModified 2026-07-06 predates the 2026-08-24 rewrite (checker, from the log and git log). The brief requires dateModified set to today on any edited page, and this pass edits the page for K7, K8, K10 and K15, so it moves to 2026-10-18; no line on the page (MINOR). - MINOR | K14 (contexts-example-mccomb): That cats purr across contentment and distress, illness, parturition, nursing and near death is a robustly documented ethological observation (noted, e.g., in the framing of McComb et al. 2009 and standard veterinary/felid-behaviour literature). | http://web.archive.org/web/20260816162826id_/https://www.cell.com/current-biology/fulltext/S0960-9822(09)01168-3 | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B: listed. The observation is supported (Schoetz and Eklund 2011, von Muggenthaler, per the checker), but the example named, McComb et al. 2009's framing, does not note distress, illness, parturition or dying. A loose example in a reference note. - WRONG | K15 (ref9-wrong-authors): Ferretti, S., et al. (2025), "Meows encode less individual information than purrs," Scientific Reports, | https://api.crossref.org/works/10.1038/s41598-025-31536-7 | FIXED: reference 9 in index.html and reference 7 plus prose in research/cats-purr/README.md now name Russo, Schild and Knörnschild with the full title, volume 15, article 43490; Corrected line added. Re-read Crossref https://api.crossref.org/works/10.1038/s41598-025-31536-7: authors Danilo Russo, Anja Birgit Schild, Mirjam Knoernschild (Crossref: 'Knörnschild'); title 'Meows encode less individual information than purrs and show greater variability in domestic than in wild cats'; Scientific Reports 15, article 43490, published 2025-12-09. No author named Ferretti. Fix the authors in index.html reference 9 and README.md (line 219 prose and reference 7), and restore the full title, which the page cut short inside quotation marks. - MINOR | K16 (both-from-table): (added) Two corrections to how that study is usually retold, both taken from its own supplementary table. | https://ars.els-cdn.com/content/image/1-s2.0-S0960982209011683-mmc1.pdf | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B: listed. The second correction (peak normalisation) is in McComb's Supplemental Experimental Procedures, not in a table; the page's Sources entry says it right. A slip of precision. - MINOR | K17 (fourteen-months): (added) verify-why-do-cats-purr.mjs, the check /strata/why-do-cats-purr/ named for fourteen months and did not have. | verify-why-do-cats-purr.mjs | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B: listed. The verifier's header says 'fourteen months'; the page was first committed 2026-07-06 and the verifier written 2026-08-24, seven weeks. A slip in the verifier's comment, not an assertion it makes; no WRONG fix touches it. - MINOR | K18 (prose-numbers-in-data): (added) 8 the shipped page every number in its prose exists in the data | verify-why-do-cats-purr.mjs | NOT FIXED, listed by triage (MINOR, tier B). MINOR, tier B: listed. Overclaim about the page's own check: section 8 checks that headline numbers from the data appear in the prose, not that every prose number is in the data. Overclaims about the page's own checking are MINOR. - CONFIRMED | the honest answer is that the mechanism was quietly reopened in 2023 | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [md-reopened] Results: "the current consensus opinion that cat purring requires AMC was reached prematurely"; Discussion: "a revised theory of purring is required". "quietly" is the page's characterisation. - CONFIRMED | For fifty years the textbook story was a neural oscillator firing the laryngeal muscles about twenty-five times a second. | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [md-fifty-years] Summary: "cats have long been believed ... neurally driven electromyographic burst patterns (typically at 20–30 Hz)"; the AMC model is Remmers & Gautier 1972, 51 years before 2023. - CONFIRMED | Herbst and colleagues blew air through eight excised cat larynges (no brain, no nerves, no muscle activation) and they purred anyway, at 25–30 Hz | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [md-herbst-eight] Summary: "all eight excised larynges produced self-sustained oscillations at typical cat purring rates" and "fundamental frequencies of 25 to 30 Hz without neural input or muscular contraction". - CONFIRMED | the authors say so themselves: their data do not fully reject it | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [md-authors-say] Summary: "While our data do not fully reject the AMC hypothesis for purring". - CONFIRMED | twenty openly licensed recordings of real cats were measured from scratch | public/strata/why-do-cats-purr/measured.json | [md-20-recordings] "sourceCount":20, 20 rows; the licence of every one of the twenty also checked at its source (see the added licence line). - CONFIRMED | with the confidence thresholds fixed on 3,000 stretches of noise beforehand | research/cats-purr/data/null-floors.json | [md-3000-noise] "draws": 1000 per colour, three colours (white, pink, burst). - CONFIRMED | the 157 purr cycles Herbst's team marked by hand | research/cats-purr/data/grade.json | [md-157-marks] "cycleMarks": 157; Data S2.TextGrid fetched from ars.els-cdn.com mmc3.zip has tier "cycles", "points: size = 157", sha256 582f478e... as in data/sources.json; Figure S1 legend: "manual annotation of acoustic excitation events". - CONFIRMED | The fundamental comes out at a median of 27.15 Hz, every readable recording inside 20–30 Hz | public/strata/why-do-cats-purr/measured.json | [md-median-2715] recomputed from the 19 rows: median 27.153, min 23.976, max 28.866, 19 of 19 in band. - CONFIRMED | the 220–520 Hz 'cry' turns out to be present in ordinary purrs too | public/strata/why-do-cats-purr/measured.json | [md-cry-ordinary] recomputed: median cry excess 4.15 dB vs control 0.695 dB, cry higher in 16 of 20, two-sided sign test p = 0.0118. - CONFIRMED | which is what McComb's own supplementary table says | https://ars.els-cdn.com/content/image/1-s2.0-S0960982209011683-mmc1.pdf | [md-mccomb-table] Table S2: voiced peak frequency 391.70 (solicitation) vs 367.50 Hz (non-solicitation), P = 0.284; height 3.59 vs 0.51 dB. A peak is measured in the ordinary purrs, though only 0.51 dB above the slope on average. McComb told the press the cry "occurs at a low level in cats' normal purring" (https://www.sciencedaily.com/releases/2009/07/090713121348.htm). - CONFIRMED | what the popular retelling loses | https://en.wikipedia.org/wiki/Purr | [md-popular-retelling] The record had no source. Wikipedia, Purr: "This purring has a high-frequency component not present in other purrs." That is the retelling the page corrects. (Wikipedia API extract, fetched 2026-10-19.) - CONFIRMED | the frequency difference between the two halves of the breath is real inside a cat and has no consistent direction across cats | public/strata/why-do-cats-purr/measured.json | [md-two-halves] recomputed from rows: 11 of 19 loud-higher, 8 significant, 5 up and 3 down, median 0.081 Hz; inside the reference cat +1.583 Hz, p = 5e-5 (research/cats-purr/data/grade.json). - CONFIRMED | Median fundamental 27.15 Hz. | public/strata/why-do-cats-purr/measured.json | [md-median-fund] "f0Median":27.153, recomputed from rows. - CONFIRMED | Apparatus, sources and licences in `research/cats-purr/`; the check is `node verify-why-do-cats-purr.mjs`. | verify-why-do-cats-purr.mjs | [md-dir-check] Ran it from a git archive of d9d5d37097: --quick 75/75 passed; --quick --mutate 9 of 9 controls red. - CONFIRMED | running on the out-breath and the in-breath with barely a seam | https://www.ebi.ac.uk/europepmc/webservices/rest/PMC12695941/fullTextXML | [p-continuous] "a continuous low-frequency, low-amplitude sound produced with the mouth closed on alternating egressive and ingressive airstreams". - CONFIRMED | for minutes at a time | https://www.ebi.ac.uk/europepmc/webservices/rest/PMC12695941/fullTextXML | [p-minutes] "Purring can last from less than a second to several minutes". - CONFIRMED | a fundamental down around 25 hertz | https://www.ebi.ac.uk/europepmc/webservices/rest/PMC12695941/fullTextXML | [p-25hz] "The fundamental frequency (F0) of purring, which ranges from 25 to 30 Hz". - CONFIRMED | at the very floor of what a human ear will register as a tone | https://en.wikipedia.org/wiki/Hearing_range | [p-floor-of-hearing] "The human range is commonly given as 20 to 20,000 Hz". - CONFIRMED | And the animal producing it weighs four kilograms. | http://liu.diva-portal.org/smash/get/diva2:539090/FULLTEXT01.pdf | [p-4kg] PDF read with pdftotext, p. 9: "a domestic cat normally weighs between 4 and 5 kilos". - CONFIRMED | that pitch comes out at a median of 27.2 Hz, with every recording the instrument could read landing inside 20 to 30 Hz | public/strata/why-do-cats-purr/measured.json | [p-median-272] 27.153 rounds to 27.2. - CONFIRMED | (The exact seat of that oscillator was inferred from the muscle recordings, never pinned to a single anatomical spot.) | http://web.archive.org/web/20240929120759id_/https://www.sciencedirect.com/science/article/abs/pii/0034568772900643 | [m-seat] R&G place it only "within the central nervous system", inferred from afferent interruption with EMG; no source located it further. - CONFIRMED | Each burst snaps the glottis partly shut; air passing through is chopped into a train of little puffs, and that train is the purr. | http://web.archive.org/web/20240929120759id_/https://www.sciencedirect.com/science/article/abs/pii/0034568772900643 | [m-glottis] "Each burst of muscle discharge causes glottal closure and the development of a trans-glottal pressure which, when dissipated by glottal opening, generates sound." (The source says closure, the page "partly shut": weaker, not stronger.) - CONFIRMED | This is the model built up across the classic work of the 1970s–90s. | https://api.crossref.org/works/10.1111/j.1469-7998.1991.tb04749.x | [m-classic] FS 1991 abstract: "the primary mechanism ... is a centrally driven laryngeal modulation of respiratory flow"; R&G 1972. - CONFIRMED | It was assumed, not shown, that the brain was doing all of it. | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [m-assumed] "currently believed to rely entirely upon neural control" and "Direct empirical evidence for this AMC mechanism is sparse." Sparse is not none, but "all of it" (entirely) was the untested part. - CONFIRMED | In 2023 a team led by Christian T. Herbst took eight larynges from domestic cats (euthanised for unrelated medical reasons) | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [m-herbst-take] "euthanised due to terminal disease unrelated to the respiratory tract"; Herbst is first author and lead contact. - CONFIRMED | blew warm, humid air through them on a bench | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [m-warm-humid] "heated (∼37°C) and humidified (100% humidity) air". - CONFIRMED | no brain, no nerves, no muscle activation at all. | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [m-no-brain] "without neural input or muscular contraction". - CONFIRMED | they self-oscillated at 25–30 Hz, squarely in the natural range. | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [m-selfosc] Summary: "self-sustained oscillations at typical cat purring rates"; Results add "at or below" 25–30 Hz and a 15 to 200 Hz continuum in six of eight. - CONFIRMED | the same myoelastic-aerodynamic principle behind human "vocal fry." | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [m-vocal-fry] "This is comparable to vocal fry phonation in humans". - CONFIRMED | In the living animal, muscle tone almost certainly still sets up and sustains the oscillation | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [m-muscle-tone] Better support than the record's excerpt: in the bench rig the folds had to be "adducted to a medialized position using prongs", so in a live cat muscles must do that; Discussion: "we hypothesize that in vivo cat purring is facilitated by the co-existence of AMC and MEAD". - CONFIRMED | The authors say so themselves: their data, in their words, do not fully reject the active-muscle model, and the purr is potentially augmented by muscle contraction. | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [m-authors-quote] verbatim: "do not fully reject", "potentially augmented by AMC". - CONFIRMED | is still open. | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [m-still-open] "Further research will be needed to elucidate the physiological and physical vocal mechanisms that underlie cat purring". - CONFIRMED | cats also purr when they are injured, frightened, giving birth, being examined by a vet, and dying. | https://www.futura-sciences.com/en/scientists-finally-explain-why-cats-purr-and-its-not-just-happiness_34586/ | [w-contexts] futura: "at the vet, when injured, even while giving birth or near death"; "frightened" is in von Muggenthaler's abstract ("cats purr when they are severely injured or frightened", https://api.crossref.org/works/10.1121/1.4777098); Schötz & Eklund p. 9: "hungry, stressed, in pain or close to dying". - CONFIRMED | In 2009, Karen McComb and colleagues studied the specific purr cats use to get fed | http://web.archive.org/web/20260816162826id_/https://www.cell.com/current-biology/fulltext/S0960-9822(09)01168-3 | [w-mccomb-2009] "some cats also purr at feeding time, actively soliciting food from their owners". - CONFIRMED | Hidden inside its low rumble is a high-frequency peak, around 220–520 Hz with a pooled mean of 380 Hz | http://web.archive.org/web/20260816162826id_/https://www.cell.com/current-biology/fulltext/S0960-9822(09)01168-3 | [w-peak-range] "(range 220–520 Hz, mean 380 Hz)"; "pooled" is the page's word. - CONFIRMED | uncannily close to the frequency band of a human infant's cry | http://web.archive.org/web/20260816162826id_/https://www.cell.com/current-biology/fulltext/S0960-9822(09)01168-3 | [w-infant] "comparable frequencies to the fundamental frequency of a human infant's cry (300–600Hz in a healthy infant". - CONFIRMED | People rated purrs carrying that peak as more urgent and less pleasant, and when the peak was digitally removed and the purr replayed, the same listeners rated it less urgent. | http://web.archive.org/web/20260816162826id_/https://www.cell.com/current-biology/fulltext/S0960-9822(09)01168-3 | [w-urgent] "re-synthesised solicitation purrs to remove only the voiced component ... perceived as being significantly less urgent"; the same 50 participants did part 2 (Supplemental Data). - CONFIRMED | its frequency did not differ significantly between solicitation and non-solicitation purrs (391.70 Hz against 367.50 Hz, P = 0.284) | https://ars.els-cdn.com/content/image/1-s2.0-S0960982209011683-mmc1.pdf | [w-freq-same] PDF read with pdftotext, Table S2: "Voiced peak frequency (Hz) 391.70 367.50 ... Z = -1.07, N – Ties = 10, P = 0.284". - CONFIRMED | What differed was its height, 3.59 dB against 0.51 dB (p = 0.008). | https://ars.els-cdn.com/content/image/1-s2.0-S0960982209011683-mmc1.pdf | [w-height-diff] Table S2: "Voiced peak Height (dB) 3.59 0.51 ... Z = -2.67, N - Ties = 9, p = 0.008**". - CONFIRMED | the playbacks were peak-normalised to 99% of full scale, which equalises amplitude rather than perceived loudness | https://ars.els-cdn.com/content/image/1-s2.0-S0960982209011683-mmc1.pdf | [w-normalised] Supplemental Experimental Procedures: "normalized to 99 % peak amplitude (a process designed to control for any variations in amplitude)". This is in the procedures text, not in Table S2 (see the added line on "both taken from its own supplementary table"). - CONFIRMED | Measured below on twenty recordings, none of them labelled as solicitations, the peak is duly present in ordinary purring too. | public/strata/why-do-cats-purr/measured.json | [w-ordinary-present] as md-cry-ordinary. - CONFIRMED | The instrument at the top of the page plays numbers taken out of four papers | public/strata/why-do-cats-purr/index.html | [r-four-papers] The synth's fixed numbers: 25 Hz default and a 25–30 Hz band (FS 1991, Herbst), 380 Hz (McComb), a 15–150 Hz slider (von Muggenthaler's 25–150). Exception in k-every-freq. - CONFIRMED | by running the estimators over three thousand stretches of noise containing no periodicity at all (flat, 1/f, and randomly spaced bursts) | research/cats-purr/data/null-floors.json | [r-3000-stretches] "draws": 1000 per colour; colours white (flat), pink (1/f), burst. - CONFIRMED | Nineteen of the twenty produced a fundamental, and all nineteen land inside the 20 to 30 Hz the literature claims: median 27.15 Hz, range 23.98 to 28.87. | public/strata/why-do-cats-purr/measured.json | [r-nineteen] recomputed from rows. 20–30 Hz is Herbst's "purring range (20–30 Hz)" and R&G's "20–30 times per second". - CONFIRMED | Frazer Sissom, Rice and Peters put domestic cats at 26.3 ± 1.95 Hz in 1991. | https://api.crossref.org/works/10.1111/j.1469-7998.1991.tb04749.x | [r-fs-263] abstract: "purr at a frequency of 26·3 ±; 1·95 (S.D.) Hz". - CONFIRMED | it is also the only recording whose reading moves when you re-encode it. | research/cats-purr/data/codec.json | [r-twentieth] elle moves 6.01 Hz; the other four re-encoded at most 0.125 Hz. Only the five lossless recordings were re-encoded, so it is the only one of five tested, not of twenty. - CONFIRMED | cycles marked by hand in the source paper | research/cats-purr/data/grade.json | [r-key157] 157, and 157 points in the fetched TextGrid. - CONFIRMED | of them found blind, to within 5 ms | research/cats-purr/data/grade.json | [r-key128] "matched": 128 at "matchToleranceMs": 5. - CONFIRMED | between their rate and ours | research/cats-purr/data/grade.json | [r-key-pct] "differencePct": -0.5; (24.881 - 25.006) / 25.006 = -0.50%. - CONFIRMED | windows where our own headline rule fired | public/strata/why-do-cats-purr/measured.json | [r-key-0of80] "consensus":{"windows":80,"fired":0. - CONFIRMED | a 12.4 second recording of a live domestic cat purring (13 years old, castrated male) | https://ars.els-cdn.com/content/image/1-s2.0-S0960982223012307-mmc1.pdf | [r-herbst-rec] PDF read with pdftotext: "In vivo purring vocalization of a domestic cat (13 years, castrated male)". - CONFIRMED | a 12.4 second recording of a live domestic cat purring (13 years old, castrated male) | research/cats-purr/data/corpus.json | [r-herbst-duration] "totalDurationSec": 12.404; the Data S1.wav fetched from mmc2.zip is 12.404 s, 48 kHz mono, sha256 40bd50a8... as in data/sources.json. - CONFIRMED | a point tier marking 157 individual purr cycles by hand | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [r-herbst-tiers] "Data S2. Praat TextGrid file with glottal cycle annotations of Data S1"; opened it: tiers "phonation" (IntervalTier, 21 intervals, 5 labelled i and 5 e, over 2.1 to 8.8 s) and "cycles" (TextTier, 157 points). - CONFIRMED | put 128 of their 157 marks within 5 milliseconds, with a mean offset of 2.05 ms and a scatter of 0.91 ms, and read the rate as 24.881 Hz against the 25.006 Hz their marks give | research/cats-purr/data/grade.json | [r-128-of-157] all five figures match; verifier section 5 re-scores the key and passes. - CONFIRMED | That is 0.5% apart. | research/cats-purr/data/grade.json | [r-fiveoh] -0.50%. - CONFIRMED | a random key scores 90.4% | research/cats-purr/data/grade.json | [r-random-904] "chanceRecall": 0.9041 at 20 ms. - CONFIRMED | the detections sit about 43 milliseconds apart | research/cats-purr/data/grade.json | [r-43ms] 12.404 s / 286 detections = 43.4 ms. - CONFIRMED | The observed 91.7% was a lift of 1.01 over chance, p = 0.35. | research/cats-purr/data/grade.json | [r-lift101] "lift": 1.01, "p": 0.354. - CONFIRMED | at 5 ms the counter finds 81.5% where chance finds 23.0%, a lift of 3.5 with p = 5 × 10⁻⁴, and at 2 ms it finds 32.5% where chance finds 9.2% | research/cats-purr/data/grade.json | [r-5ms-23] all match; p = 0.0004998 = 1/2001, i.e. 2,000 random keys, as the page says. - CONFIRMED | The abstract of How cats purr contains a sentence that has been sitting behind a paywall since 1991: "The frequency at mid-expiration exceeds that at mid-inspiration by 2·4 ± 1·3 Hz." | https://api.crossref.org/works/10.1111/j.1469-7998.1991.tb04749.x | [r-quote-1991] verbatim in the Crossref abstract. Note: the sentence is in the free abstract; it is the paper that is paywalled. - CONFIRMED | A purr, in other words, is slightly faster on the way out than on the way in. | https://api.crossref.org/works/10.1111/j.1469-7998.1991.tb04749.x | [r-faster-out] higher frequency at mid-expiration. - CONFIRMED | the egressive cycles run at 26.184 Hz (n = 50) against 24.601 Hz ingressive (n = 97), a difference of +1.58 Hz, permutation p = 5 × 10⁻⁵ | research/cats-purr/data/grade.json | [r-egress-1991] theirPhases matches all five numbers. - CONFIRMED | That is within one standard deviation of a number published thirty-five years earlier, on a different cat, by a different group | https://api.crossref.org/works/10.1111/j.1469-7998.1991.tb04749.x | [r-35-years] 2.4 - 1.3 = 1.1 <= 1.58; 2026 - 1991 = 35. - CONFIRMED | the louder half of the breath is the faster half in 11 of 19 | public/strata/why-do-cats-purr/measured.json | [r-11of19] recomputed from rows: 11 of 19 with phaseDeltaHz > 0. - CONFIRMED | The difference reaches p < 0.05 in eight of them: five going up and three going down. The median difference across recordings is 0.08 Hz | public/strata/why-do-cats-purr/measured.json | [r-eight-sig] recomputed from rows: 8, 5 up, 3 down, median 0.081. - CONFIRMED | Schötz and Eklund found both directions among four cats. | http://liu.diva-portal.org/smash/get/diva2:539090/FULLTEXT01.pdf | [r-se-both] PDF read with pdftotext, p. 12: "Two of the cats (D/T) had significantly higher F0 for the egressive phase ... One cat (V) showed the opposite pattern ... no significant difference was found for one cat (R)." - CONFIRMED | the 220 to 520 Hz band carries a median 4.15 dB more energy than the spectrum around it predicts, against 0.70 dB for a matched control band at 600 to 900 Hz, and it is the higher of the two in 16 of 20 recordings (sign test p = 0.012) | public/strata/why-do-cats-purr/measured.json | [r-cry-415] recomputed; control band [600, 900] in research/cats-purr/measure.mjs line 226. - CONFIRMED | the peak sits at a median of 399 Hz | public/strata/why-do-cats-purr/measured.json | [r-399] median of the 16 interior cryPeakHz values = (389.6 + 408) / 2 = 398.8. - CONFIRMED | McComb's pooled mean is 380. | http://web.archive.org/web/20260816162826id_/https://www.cell.com/current-biology/fulltext/S0960-9822(09)01168-3 | [r-mccomb-380] "mean 380 Hz". - CONFIRMED | It fired on 0 of 80 windows. | public/strata/why-do-cats-purr/measured.json | [f-0of80] "fired":0. - CONFIRMED | a window counts only if at least three of four independent estimators clear their noise floor and agree within 5% | public/strata/why-do-cats-purr/measured.json | [f-rule] "decisionRule". - CONFIRMED | beaten by 1/f room noise more than three times in four | public/strata/why-do-cats-purr/measured.json | [f-acf] acf clears 18 of 80, so 62 of 80 (77.5%) do not; its floor was set by pink (1/f) noise ("floorSetBy"). - CONFIRMED | never clears a null made of randomly spaced noise bursts | public/strata/why-do-cats-purr/measured.json | [f-cep] cep 0 of 80; floor set by burst. - CONFIRMED | counts events in time; the only one with an external grade | research/cats-purr/data/grade.json | [f-pulse] grade.json grades the pulse counter only. - CONFIRMED | taken that way the autocorrelation agrees with the counter within 5% in 59% of windows, the cepstrum in 44%, the grid in 21% | public/strata/why-do-cats-purr/measured.json | [f-face-value] 44/75 = 58.7%, 33/75 = 44.0%, 16/75 = 21.3%. - CONFIRMED | (96.9% of their egressive cycles fall in our loud class) | research/cats-purr/data/grade.json | [f-969] e/loud 31, e/quiet 1: 31/32 = 0.9688. - CONFIRMED | So the lossless recordings were pushed through the same encoders the lossy ones arrived in and re-measured: same cat, same second, only the encoder changed. | research/cats-purr/data/codec.json | [c-same-encoders] five treatments (MP3 128 and 190 kbps, Vorbis q4, Vorbis at 16 kHz, MP3 twice) on the five lossless recordings. - CONFIRMED | Across 25 comparisons the median shift in the fundamental is 0.014 Hz. | research/cats-purr/data/codec.json | [c-25] 5 x 5 = 25; "medianAbsDeltaHz": 0.0138. - CONFIRMED | The single recording it destabilises is the one the confidence floor had already refused. | research/cats-purr/data/codec.json | [c-destab] elle 6.01 Hz; elle has f0Hz null in measured.json. - CONFIRMED | four cats in Schötz and Eklund | http://liu.diva-portal.org/smash/get/diva2:539090/FULLTEXT01.pdf | [l-four-cats] PDF read: "a comparative analysis of purring in four domestic cats". - CONFIRMED | one in Eklund, Peters and Duthie | http://liu.diva-portal.org/smash/get/diva2:539494/FULLTEXT01.pdf | [l-one-eklund] PDF read: "Data were collected from two felids, one domestic cat, and one cheetah"; Table 1 domestic cat 22.6 Hz. - CONFIRMED | Frazer Sissom gives 26.3 ± 1.95 Hz with no recoverable n. | https://api.crossref.org/works/10.1111/j.1469-7998.1991.tb04749.x | [l-fs-n] the abstract states no n; the log records every route to the full text failing. - CONFIRMED | The largest modern corpus, 557 purrs from 21 cats, measured duration and spectral centroid and did not measure the fundamental at all. | https://www.ebi.ac.uk/europepmc/webservices/rest/PMC12695941/fullTextXML | [l-557] "We measured the duration and the spectral centroid of meows ... and purrs"; MFCCs; F0 appears only as a literature citation. - CONFIRMED | The specific "purr as healing mechanism" claim traces to a single 2001 conference presentation by Elizabeth von Muggenthaler, not a controlled, peer-reviewed study showing that purring heals cats. | https://api.crossref.org/works/10.1121/1.4777098 | [h-single] a JASA meeting abstract, 2001, asserting frequency overlap. That every retelling descends from this one talk ("single") is a genealogy no source shows; the abstract also carries "[Published with permission from the New Zealand Veterinary Journal...]". - CONFIRMED | there is real bone-biology research showing that certain low-magnitude, high-frequency mechanical vibrations can influence bone in the lab | https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=11493908&retmode=xml | [h-rubin-bone] sheep, trabecular density +34.2%. - CONFIRMED | But that work is not about cats, not about purring | https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=11493908&retmode=xml | [h-not-cats] "adult sheep". - CONFIRMED | The "25–150 Hz" band it reports is a cross-species range (44 felids of several species), not the domestic-cat fundamental. | https://api.crossref.org/works/10.1121/1.4777098 | [h-hundred-fifty] "Forty-four felids were recorded including cheetahs, ocelots, pumas, domestic cats, and servals ... Every felid in the study generated strong frequencies between 25 and 150 Hz." - CONFIRMED | the fundamental itself genuinely present in the file in 18 of 19 | public/strata/why-do-cats-purr/measured.json | [k-18of19] recomputed: fundamentalPresent true in 18 of the 19 measured rows (cleo false). - CONFIRMED | cats purr on both inhale and exhale, near-continuously | https://www.ebi.ac.uk/europepmc/webservices/rest/PMC12695941/fullTextXML | [k-both-breaths] as p-continuous. - CONFIRMED | excised cat larynges self-oscillate at 25–30 Hz on airflow alone (Herbst et al. 2023, 8 larynges) | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [k-8-larynges] as md-herbst-eight. - CONFIRMED | listeners rate purrs carrying the voiced peak as more urgent, and rate the same purrs less urgent once it is removed (McComb et al. 2009). | http://web.archive.org/web/20260816162826id_/https://www.cell.com/current-biology/fulltext/S0960-9822(09)01168-3 | [k-urgent] "stimuli with the VP removed were consistently judged by participants as less urgent". - CONFIRMED | The claim rests on a 2001 conference talk, not a controlled study in cats. | https://api.crossref.org/works/10.1121/1.4777098 | [k-2001-talk] as h-single. - CONFIRMED | the published literature disagrees with itself about it | http://liu.diva-portal.org/smash/get/diva2:539090/FULLTEXT01.pdf | [k-lit-disagrees] Schötz & Eklund's cats go both ways; Eklund 2010's one cat runs opposite to FS 1991 (not significant). - CONFIRMED | a harmonic-rich low tone at the frequency you set, plus an optional 380 Hz component standing in for the voiced peak | public/strata/why-do-cats-purr/index.html | [k-harmonic-380] seven sawtooth harmonics; CRY_HZ = 380. - CONFIRMED | with --mutate to corrupt the inputs on purpose and confirm the checks go red | verify-why-do-cats-purr.mjs | [k-mutate] ran --quick --mutate: "9 of 9 controls went red as required." - CONFIRMED | Frazer Sissom, D.E., Rice, D.A. & Peters, G. (1991), "How cats purr," Journal of Zoology 223(1):67–78. | https://api.crossref.org/works/10.1111/j.1469-7998.1991.tb04749.x | [ref1-fs] Crossref: given "Dawn E. Frazer", family "Sissom"; Rice, Peters; vol 223 issue 1, 67-78, 1991. - CONFIRMED | The ~25 Hz fundamental and the continuous inhale-and-exhale production: | https://api.crossref.org/works/10.1111/j.1469-7998.1991.tb04749.x | [ref1-attrib] 26.3 Hz, and frequencies given at mid-expiration and mid-inspiration; "continuous" is not in the abstract (it is in Russo 2025's definition). - CONFIRMED | Remmers, J.E. & Gautier, H. (1972), "Neural and mechanical mechanisms of feline purring," Respiration Physiology 16(3):351–361 | https://api.crossref.org/works/10.1016/0034-5687(72)90064-3 | [ref1-rg] Crossref matches title, vol 16 issue 3, 351-361, 1972-12. Herbst calls it "the AMC model proposed by Remmers & Gautier". - CONFIRMED | (an EMG study: intrinsic laryngeal muscles fire in regular ~20–30 Hz bursts, inferred to be centrally driven) | http://web.archive.org/web/20240929120759id_/https://www.sciencedirect.com/science/article/abs/pii/0034568772900643 | [ref1-emg] abstract. - CONFIRMED | Herbst, C.T., Prigge, T., Garcia, M., Hampala, V., Hofer, R., Weissengruber, G.E., Švec, J.G. & Fitch, W.T. (2023), "Domestic cat larynges can produce purring frequencies without neural input," Current Biology 33(21):4727–4732.e4. | https://api.crossref.org/works/10.1016/j.cub.2023.09.014 | [ref2-herbst] all eight authors in order, vol 33 issue 21, 4727-4732.e4. The tool's "DOI MISMATCH" is only initials against "Christian T."; not a fault. - CONFIRMED | Open access under CC BY 4.0. | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [ref2-ccby] "User License: Creative Commons Attribution (CC BY 4.0)". - CONFIRMED | Its Data S1 (a 12.4 s in-vivo purr) and Data S2 (the authors' own Praat annotation of it, 157 hand-marked cycles) are the answer key the instrument below is graded against. | http://web.archive.org/web/20250730173344id_/https://www.cell.com/current-biology/fulltext/S0960-9822(23)01230-7 | [ref2-datas] both files fetched and opened (12.404 s; 157 points); hashes match data/sources.json. - CONFIRMED | McComb, K., Taylor, A.M., Wilson, C. & Charlton, B.D. (2009), "The cry embedded within the purr," Current Biology 19(13):R507–R508. | https://api.crossref.org/works/10.1016/j.cub.2009.05.033 | [ref4-mccomb] matches. - CONFIRMED | Ten cats, fifty listeners. | http://web.archive.org/web/20260816162826id_/https://www.cell.com/current-biology/fulltext/S0960-9822(09)01168-3 | [ref4-ten-fifty] "purrs from 10 cats ... to 50 human participants". - CONFIRMED | The "range 220–520 Hz, mean 380 Hz" is from the main text | http://web.archive.org/web/20260816162826id_/https://www.cell.com/current-biology/fulltext/S0960-9822(09)01168-3 | [ref4-range-quote] verbatim. - CONFIRMED | von Muggenthaler, E. (2001), "The felid purr: A healing mechanism?" presented at | https://api.crossref.org/works/10.1121/1.4777098 | [ref5-vm] JASA 110(5_Supplement):2666, 2001-11. - CONFIRMED | the 142nd Meeting of the Acoustical Society of America (Fort Lauderdale, Dec 2001) | http://web.archive.org/web/2003id_/http://www.acoustics.org:80/press/142nd/lay_lang.html | [ref5-142nd] The record had no source. ASA's archived index: "lay language papers being presented at the 142nd Meeting of the Acoustical Society of America to be held December 3-7 in Fort Lauderdale, Florida", including von Muggenthaler's rhino paper ("Presented Tuesday afternoon, December 4, 2001", .../142nd/vonmuggenthaler.html), whose JASA abstract (10.1121/1.4777095) is on the same supplement page, 110(5_Supplement):2666, as the purr abstract (10.1121/1.4777098). - CONFIRMED | A conference abstract reporting frequency overlap and posing healing as a question | https://api.crossref.org/works/10.1121/1.4777098 | [ref5-question] title ends "A healing mechanism?"; the abstract argues frequency correspondence. - CONFIRMED | Rubin, C., et al. (2001), "Anabolism: Low mechanical signals strengthen long bones," Nature 412:603–604 | https://api.crossref.org/works/10.1038/35088122 | [ref6-rubin] Crossref: Rubin, Turner, Bain, Mallinckrodt, McLeod, Nature 412(6847):603-604, 2001; "et al." covers the tool's DOI MISMATCH. - CONFIRMED | and with mixed results | https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=21849376,27145947&retmode=text&rettype=abstract | [ref6-mixed] The record had no source. Lau et al. 2011 meta-analysis: "no overall treatment effect on bone mineral density in older women"; Oliveira et al. 2016: potential for lumbar spine aBMD, "the ideal intervention is not yet clear". - CONFIRMED | Twenty recordings, their licences and attributions in research/cats-purr/NOTICE.md | research/cats-purr/NOTICE.md | [ref7-notice] verifier: "every recording is credited in NOTICE.md" ok. - CONFIRMED | the full-length originals are re-obtainable with node research/cats-purr/fetch.mjs and hash-verified. | research/cats-purr/fetch.mjs | [ref7-hash] fetch.mjs compares sha256 to data/sources.json; spot check: Herbst Data S1 and S2 fetched today hash to 40bd50a8... and 582f478e..., the recorded values. - CONFIRMED | Schötz, S. & Eklund, R. (2011), "A comparative acoustic analysis of purring in four cats," Proc. Fonetik 2011, TMH-QPSR 51:9–12 | http://liu.diva-portal.org/smash/get/diva2:539090/FULLTEXT01.pdf | [ref8-schotz] PDF: "TMH - QPSR Vol. 51", "Fonetik 2011", pages 9 to 12. - CONFIRMED | Eklund, R., Peters, G. & Duthie, E.D. (2010), "An acoustic analysis of purring in the cheetah and in the domestic cat," Proc. Fonetik 2010, 17–22 (one domestic cat, 22.6 Hz). | http://liu.diva-portal.org/smash/get/diva2:539494/FULLTEXT01.pdf | [ref8-eklund] PDF: one domestic cat, mean 22.6 Hz; S&E reference list: "Proceedings of Fonetik 2010 ... 17–22". - CONFIRMED | 557 purrs from 21 cats: the largest purr corpus published, measuring duration, spectral centroid and MFCCs. It does not report a fundamental frequency. | https://www.ebi.ac.uk/europepmc/webservices/rest/PMC12695941/fullTextXML | [ref9-557] counts and measures confirmed; "largest" as l-largest. - CONFIRMED | (added) the one owners describe as impossible to ignore | https://www.sussex.ac.uk/broadcast/read/1208 | University of Sussex release, 2009: "a special purr that humans just can't ignore", "cat owners find this 'solicitation' purr irresistible"; McComb, an owner: "so difficult to ignore" (https://www.sciencedaily.com/releases/2009/07/090713121348.htm). - CONFIRMED | (added) twenty openly licensed recordings | https://archive.org/metadata/aporee_66401_76770 | Every licence checked at its source on 2026-10-19 matches data/sources.json: six archive.org items (PD Mark x3, CC BY-SA 3.0 x2, publicdomain x1, via the metadata API), eight Wikimedia Commons files (CC BY-SA 4.0 x3, CC BY-SA 3.0 x2, Public domain x3, via extmetadata), five Freesound pages (all CC0 1.0), Herbst Data S1 (CC BY 4.0). ## Findings - [declined] K1 (pads-role-stated-as-found, MINOR, listed at tier B): MINOR, tier B, and not inside a sentence being fixed for a WRONG item: listed. The checker's quotations ('appears to allow', 'We propose') show the pads' role is the authors' proposal, which the dek states as found; loose wording, not a false fact about the mechanism's existence. - [declined] K2 (burst-rate-narrowed, MINOR, listed at tier B): MINOR, tier B: listed. Remmers and Gautier give EMG bursts '20-30 times per second'; 25-30 sits inside that range, so the sentence narrows the range rather than stating a false one; the page's own Sources entry gives ~20-30 Hz. - [declined] K3 (pads-role-stated-as-found, MINOR, listed at tier B): MINOR, tier B: listed. Same error as K1 (the pads were described earlier, in Szedenik 2008, and their role in the low rate is proposed, not shown); the sentence is not being fixed for any WRONG item. - [declined] K4 (pads-role-stated-as-found, MINOR, listed at tier B): MINOR, tier B: listed. Same error as K1 and K3: the folds self-oscillate, the pads are proposed to allow the low rate. - [declined] K5 (floor-none-reached, MINOR, listed at tier B): MINOR, tier B: listed. The floors are 99.9th percentiles, so some noise draws do exceed them (null-floors.json per the checker: acf floor 0.7282 vs pink max 0.759); the page states the one-in-a-thousand rate correctly elsewhere. Loose wording about the page's own method. - [declined] K6 (graded-before-opened, MINOR, listed at tier B): MINOR, tier B: listed. The answer-key recording is itself one of the twenty, so 'before any of the twenty were opened' is true of nineteen at most; loose wording about the page's own procedure. - [fixed] K7 (plateau-rate-third-of-a-percent, WRONG): FIXED: index.html line 384 now says the rate moves "by about a third of a hertz (1.3%)"; Corrected line added. - [fixed] K8 (eight-lossy-of-twenty, WRONG): FIXED: "Eight" changed to "Fifteen" in index.html line 463, research/cats-purr/README.md line 100 and the codec.mjs header comment; Corrected line added. - [declined] K9 (few-list-omits-mccomb, MINOR, listed at tier B): MINOR, tier B: listed. 'Remarkably few' holds; the list after it omits McComb et al. 2009 Table S2 (ten cats), an omission rather than a false statement. - [fixed] K10 (summary-144-of-157, WRONG): FIXED: the summary box (index.html line 501) now gives 128 of 157 found blind within 5 ms; Corrected line added. - [declined] K11 (every-frequency-cited, MINOR, listed at tier B): MINOR, tier B: listed. index.html line 728 sets a 6 Hz vibrato on the cry tone that no cited source gives; the swept band stays inside McComb's 220-520 Hz. An overstatement of the instrument's sourcing. - [declined] K12 (last-verified-date, MINOR, listed at tier B): MINOR, tier B: listed. The placard's 'Last verified' date is written by scripts/check-placard.mjs from the frontmatter date (a shared generator, not this layer's file); landing rebuilds the placard from the new record dated 2026-10-18. - [fixed] K13 (datemodified-stale, MINOR): FIXED: JSON-LD dateModified set to 2026-10-18 (MINOR, no page line). - [declined] K14 (contexts-example-mccomb, MINOR, listed at tier B): MINOR, tier B: listed. The observation is supported (Schoetz and Eklund 2011, von Muggenthaler, per the checker), but the example named, McComb et al. 2009's framing, does not note distress, illness, parturition or dying. A loose example in a reference note. - [fixed] K15 (ref9-wrong-authors, WRONG): FIXED: reference 9 in index.html and reference 7 plus prose in research/cats-purr/README.md now name Russo, Schild and Knörnschild with the full title, volume 15, article 43490; Corrected line added. - [declined] K16 (both-from-table, MINOR, listed at tier B): MINOR, tier B: listed. The second correction (peak normalisation) is in McComb's Supplemental Experimental Procedures, not in a table; the page's Sources entry says it right. A slip of precision. - [declined] K17 (fourteen-months, MINOR, listed at tier B): MINOR, tier B: listed. The verifier's header says 'fourteen months'; the page was first committed 2026-07-06 and the verifier written 2026-08-24, seven weeks. A slip in the verifier's comment, not an assertion it makes; no WRONG fix touches it. - [declined] K18 (prose-numbers-in-data, MINOR, listed at tier B): MINOR, tier B: listed. Overclaim about the page's own check: section 8 checks that headline numbers from the data appear in the prose, not that every prose number is in the data. Overclaims about the page's own checking are MINOR.