---
slug: the-newsquake
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: 106
confirmed: 91
wrong: 11              # 1 WRONG, 10 MINOR
unverifiable: 0
observed: 4
stale: 0
---

This was a directed repair pass over the 12 triage items in research/assay-line/claude-1018b/the-newsquake/triage.json, read at 03d7ff0071. The claim lines below are the checker's (research/assay-line/claude-1018b/the-newsquake/N2-claude.md, 106 claims against the N1 source record plus claims it left out), carried forward unchanged except the runtime-footer line (see below); the fixer re-read the evidence for both WRONG items (app.js and the verifier for the frozen figures, and the served page and /privacy for the runtime footer), fixed both on the page with dated Corrected lines, and listed the ten MINOR items, which tier B leaves on the page. The directing instance then reverted the footer fix and filed that claim CONFIRMED, under the human's ruling of 2026-09-28 (board note 706d1d) that the host is not a third party.

## Claims
- CONFIRMED | [hypo-citation] W. H. K. Lee and J. C. Lahr, HYPO71 (Revised), USGS Open-File Report 75-311, June 1975. | https://pubs.usgs.gov/of/1975/0311/report.pdf | Cover (PDF p.1): "HYP071 (REVISED) ... by W. H. K. Lee and J. C. Lahr ... OPEN-FILE REPORT 75-311 ... June, 1975". The PDF I downloaded hashes to 1ec78d52...f7cd, the reportSha256 in data.js.
- CONFIRMED | [hypo-appendices] The calibration uses the Appendix 2 card deck and the Appendix 3 printed results, file page 95. | https://pubs.usgs.gov/of/1975/0311/report.pdf | PDF p.90 "APPENDIX 2 A Listing of a Test Run of HYP071" (deck on PDF pp.91-92); PDF p.93 "APPENDIX 3 Printed Results of the HYP071 Test Run"; PDF p.95, printed "- 92 -", is the 11:12 event's output.
- CONFIRMED | [hypo-tolerance] The report asks adopters to reproduce its results to one or two counts of the last figure, which here is 0.02 s. | https://pubs.usgs.gov/of/1975/0311/report.pdf | PDF p.93: "Users adapting HYP071 for their computers should reproduce these results to ± 1 or 2 counts of the last significant figures." TPCAL is printed to 0.01 s, so two counts is 0.02 s.
- CONFIRMED | [hypo-deck] It gets the first event in Lee and Lahr's 1975 HYPO71 test deck: twenty station coordinates, twenty station delays, a five-layer velocity model, and nineteen picked arrival seconds. | https://pubs.usgs.gov/of/1975/0311/report.pdf | PDF p.91 page image: 20 station cards (SR01 to SR19 and SR8A), velocity cards "3.30 0.0 / 5.00 1.0 / 5.70 4.0 / 6.70 15.0 / 8.00 25.0", 17 P and 2 S readings for 11:12, so 19; every coordinate, delay and pick in data.js matches the image. The delay count is the separate line hypo-station-delays-note.
- CONFIRMED | [hypo-s-cards] S cards: SR04 62.45 · SR06 62.59 | https://pubs.usgs.gov/of/1975/0311/report.pdf | PDF p.91 image: "SR04IPU0 691005111258.05 62.45ISU0" and "SR06IPD0 691005111258.53 62.59IS 0". The S readings sit on the SR04 and SR06 phase cards (one card carries P and S); the values are exact.
- CONFIRMED | [hypo-travel-column] HYPO71 printed its calculated P travel time for every station. This recomputes all seventeen | https://pubs.usgs.gov/of/1975/0311/report.pdf | PDF p.95 image read at 400 dpi: a TPCAL value on all 17 P rows (SR13 1.69 ... SR02 6.44, SR01 6.90); all 17, with their DLY and P-RES, equal data.js printedStations.
- CONFIRMED | [hypo-depth-printed] The printed depth on that page reads 0.41 km. | https://pubs.usgs.gov/of/1975/0311/report.pdf | PDF p.95 image, final row "691005 1112 52.83 38-28.59 122-41.94 0.41", and both iteration-3 rows "0.41".
- CONFIRMED | [hypo-depth-841] The report's own columns say it is 8.41 km, and the page recomputes the arithmetic that decides it. | https://pubs.usgs.gov/of/1975/0311/report.pdf | PDF p.95 image: iteration 1 depth 5.00 with ADJUSTMENTS TAKEN DZ 3.41; iteration 2 depth 8.41 with ADJUSTMENTS TAKEN "0.67 0.0 0.0", so iteration 3 cannot sit at 0.41; SR13's printed TPCAL 1.69 s needs 8.41 as well (next line).
- CONFIRMED | [hypo-depth-consequence] a 0.41 km source and an 8.41 km source predict completely different travel times, and the first would have made the calibration fail for the wrong reason. | public/strata/the-newsquake/engine.js | Recreated: travelTimeAtDepth(8.41)(1.2301) = 1.694 s and travelTimeAtDepth(0.41)(1.2301) = 0.393 s, against the printed 1.69 s.
- CONFIRMED | [hypo-jbig2] The scan is a bilevel JBIG2 image and it loses the waist of an 8. | https://pubs.usgs.gov/of/1975/0311/report.pdf | Recreated: pdfimages -list gives page 95 "4338 3241 gray 1 1 jbig2 ... 8814B". The same image also prints SR07 P-SEC "54.01" (deck 54.81), SR10 "54.00" (54.80), SR06 "50.53" (58.53), SR01 "59.70" (59.78) and SR09 TPOBS "2.03" (2.83): more places than the three the page shows, which collision.json's "at least three" allows.
- CONFIRMED | [hypo-sr10-sr07] the same page prints station SR10's distance as 0.1 km where the geometry gives 8.11, and SR07's as 6.0 where it gives 6.80. | https://pubs.usgs.gov/of/1975/0311/report.pdf | PDF p.95 image, DIST column "SR07 6.0" and "SR10 0.1". Recreated with engine.js from the printed coordinates and epicentre: SR10 8.106 km, SR07 6.800 km; an independent haversine gives the same.
- CONFIRMED | [hypo-weights] because HYPO71 weights its readings by quality and by residual and this does not. | https://pubs.usgs.gov/of/1975/0311/report.pdf | P-WT: "This weight is a combination of quality weight specified in the data and other selected weightings"; TEST(01): "cutoff value for RMS below which Jeffreys' weighting of residuals is not used" (final RMS 0.16 is above 0.1). The method section adds distance weighting, which the page does not name. The "because" is the page's own explanation.
- MINOR | [hypo-station-delays-note] twenty station delays | https://pubs.usgs.gov/of/1975/0311/report.pdf | PDF p.91 image: the SR17 card's delay field is blank ("SR173845.95 12248.35" then "0.40 0.25 8"), so the deck prints 19 delays and data.js holds SR17's as null. Fortran reads a blank field as 0.0 and SR17 has no pick in this event, so no number moves. Fix at public/strata/the-newsquake/index.html line 95.
- CONFIRMED | [vt-cluster] Viral Texts cluster 400732, used for discovery and for checking witness identifiers, page fragment counts and the Coverage coordinate proposals. | https://clusters.viraltexts.org/cluster/400732 | Header "Cluster size: 733". The CSV (?download_csv=400732) hashes to 92e8aa9f...c0b3, 1,303,693 bytes, 733 rows, the cluster_csv_sha256 in target-extract.json; its Witness_id, Coverage and repeated URL rows are the ones the page uses.
- CONFIRMED | [vt-unlicensed] The unlicensed cluster CSV is not shipped. | https://clusters.viraltexts.org/about | No licence on the about page, the cluster page or the CSV ("licen" 0 hits; positive control: "Suggested Citation" and "permission" found on the same page); the site asks only to be cited. No CSV under research/the-newsquake or public/strata/the-newsquake at 7f3b133f1f (git ls-tree).
- CONFIRMED | [loc-public-domain] Library of Congress, Chronicling America, twelve public-domain newspaper issue images. | https://raw.githubusercontent.com/LibraryOfCongress/chronam/master/README.md | "The NDNP data is in the Public Domain and is itself available on the Web for anyone to use." The twelve crops are NDNP images from tile.loc.gov (research/the-newsquake/data/image-sources.txt); the Rutland crop, re-fetched today, hashes to the pinned 82b29dc0...
- CONFIRMED | [census-gazetteer] 2025 United States Census places gazetteer, representative coordinates for reviewed publication places. | https://www2.census.gov/geo/docs/maps-data/data/gazetteer/2025_Gazetteer/2025_Gaz_place_national.zip | Zip sha256 49644173...ee29, as pinned in target-extract.json; the coordinates used are its INTPTLAT and INTPTLONG internal points.
- CONFIRMED | [data-coordinates-comment] Coordinates are the 2025 Census places gazetteer, except Gold Hill from GNIS feature 858000. | https://www2.census.gov/geo/docs/maps-data/data/gazetteer/2025_Gazetteer/2025_Gaz_place_national.zip | All twelve Census rows (eleven towns plus San Francisco city) equal data.js to the last digit, e.g. "Chicago city ... 41.837045 -87.684939" and "St. Johnsbury CDP ... 44.430445 -72.008582".
- CONFIRMED | [gnis-gold-hill] GNIS populated places, public-domain coordinate for Gold Hill, Nevada, feature 858000. | https://prd-tnm.s3.amazonaws.com/StagedProducts/GeographicNames/Topical/PopulatedPlaces_National_Text.zip | Text/PopulatedPlaces_National.txt row 858000, Gold Hill, Populated Place, Nevada, Storey, 39.2904678, -119.6574011 (zip sha256 919fac88...). The zip's own metadata file (PopulatedPlaces_National_Text.xml) states useconst "None." and accconst "None", the statement the source record said it had not found.
- CONFIRMED | [cordell-citation] Ryan Cordell, Two (of Three) Ways of Looking at C19 Newspaper Exchange Networks, the closest target-side precedent. | https://ryancordell.org/research/two-of-three/ | Title "Two (of Three) Ways of Looking at C19 Newspaper Exchange Networks", Ryan C. Cordell, "Created on June 22, 2016". "Closest target-side precedent" is the page's judgment, the same as the spec's "Closest precedents" item 1.
- CONFIRMED | [pinto-citation] Pedro C. Pinto, Patrick Thiran and Martin Vetterli, Locating the source of diffusion in large-scale networks, 2012, a stronger network-aware source-localization precedent. | https://api.crossref.org/works/10.1103/PhysRevLett.109.068702 | Crossref: "Locating the Source of Diffusion in Large-Scale Networks", Pedro C. Pinto, Patrick Thiran, Martin Vetterli, Physical Review Letters, issued 2012-08-10; arXiv 1208.2534 has the same three authors. The tool's "DOI MISMATCH (gives Pinto the initial C.)" is a false alarm in assay/sources.mjs, whose initials regex reads the middle initial "C." before "Pinto".
- OBSERVED | [prior-art-search] We searched the general web, Crossref and OpenAlex on 10 September 2026 for combinations of hypocenter inversion, earthquake location, arrival time residuals, newspaper reprinting and information diffusion ... no direct application of iterative seismic hypocentre least squares to nineteenth-century newspaper reprint dates. | oversight/waves/2026-09-10-foreign-instruments/specs/the-newsquake.md | PRIOR ART: "Live search date: 10 September 2026", the three indexes, the queries verbatim, and "The five Crossref results and five OpenAlex results were seismology or historical-earthquake papers, not newspaper-reprinting applications."
- OBSERVED | [semantic-scholar-429] A Semantic Scholar API attempt returned HTTP 429. | oversight/waves/2026-09-10-foreign-instruments/specs/the-newsquake.md | "Semantic Scholar: https://api.semanticscholar.org/graph/v1/paper/search?query=hypocenter%20inversion%20newspaper%20reprinting... HTTP 429, 174 bytes".
- CONFIRMED | [row-intelligencer] Daily National Intelligencer, Washington, D.C., 1868-03-02, fit | https://clusters.viraltexts.org/cluster/400732 | Cluster: "Daily national intelligencer. 1868-03-02", "Washington City [D.C.]"; the IIIF identifier carries 1868030201 and the page masthead above the crop reads "MONDAY, MARCH 2, 1868".
- CONFIRMED | [row-crescent] New Orleans Crescent, New Orleans, Louisiana, 1868-03-08, baseline | https://clusters.viraltexts.org/cluster/400732 | Cluster: "The New Orleans crescent. 1868-03-08", "New Orleans, La."; IIIF identifier 1868030801.
- CONFIRMED | [row-argus] Evening Argus, Rock Island, Illinois, 1868-03-09, fit | https://clusters.viraltexts.org/cluster/400732 | Cluster: "The evening Argus. 1868-03-09", "Rock Island, Ill."; IIIF identifier 1868030901.
- CONFIRMED | [row-rutland] Rutland Weekly Herald, Rutland, Vermont, 1868-03-19, holdout | https://clusters.viraltexts.org/cluster/400732 | Cluster: "Rutland weekly herald. 1868-03-19", "Rutland, Vt." (three rows, one URL); IIIF identifier 1868031901.
- CONFIRMED | [row-davenport] Daily Davenport Democrat, Davenport, Iowa, 1868-04-04, fit | https://clusters.viraltexts.org/cluster/400732 | Cluster: "Daily Davenport Democrat. 1868-04-04", "Davenport, Iowa" (two rows); IIIF identifier 1868040401.
- CONFIRMED | [row-richmond] Richmond Palladium, Richmond, Indiana, 1868-04-09, fit | https://clusters.viraltexts.org/cluster/400732 | Cluster: "The Richmond palladium. 1868-04-09", "Richmond, Ind."; IIIF identifier 1868040901.
- CONFIRMED | [row-stjohnsbury] St. Johnsbury Caledonian, St. Johnsbury, Vermont, 1868-04-17, holdout | https://clusters.viraltexts.org/cluster/400732 | Cluster: "St. Johnsbury Caledonian. 1868-04-17", "St. Johnsbury, Vt."; IIIF identifier 1868041701.
- CONFIRMED | [row-bellows] Bellows Falls Times, Bellows Falls, Vermont, 1868-04-24, holdout | https://clusters.viraltexts.org/cluster/400732 | Cluster: "Bellows Falls times. 1868-04-24", "Bellows Falls, Vt."; IIIF identifier 1868042401.
- CONFIRMED | [row-goldhill] Gold Hill Daily News, Gold Hill, Nevada, 1868-04-25, fit | https://clusters.viraltexts.org/cluster/400732 | Cluster: "Gold Hill daily news. 1868-04-25", "Gold Hill, N.T. [Nev.]"; IIIF identifier 1868042501.
- CONFIRMED | [row-findlay] Hancock Jeffersonian, Findlay, Ohio, 1868-05-01, fit | https://clusters.viraltexts.org/cluster/400732 | Cluster: "The Hancock Jeffersonian. 1868-05-01", "Findlay, Ohio"; IIIF identifier 1868050101.
- CONFIRMED | [row-chicago] Workingman's Advocate, Chicago, Illinois, 1868-05-30, fit | https://clusters.viraltexts.org/cluster/400732 | Cluster: "The workingman's advocate. 1868-05-30", "Chicago [Ill.]"; IIIF identifier 1868053001.
- CONFIRMED | [row-clarksville] Clarksville Chronicle, Clarksville, Tennessee, 1868-06-26, fit | https://clusters.viraltexts.org/cluster/400732 | Cluster: "Clarksville chronicle. 1868-06-26", "Clarksville, Tenn." (two rows); IIIF identifier 1868062601.
- CONFIRMED | [printed-intelligencer] … lines from thirty-eight English poets. The names of the authors are given below: | research/the-newsquake/data/daily-national-intelligencer.jpg | The frozen crop begins "lines from thirty-eight English poets The names of the authors are given below:". Above the crop the same page prints "Mrs. H. A. Deming, of San Francisco, is said to have occupied a year..." (tile.loc.gov region 180,330,960,300 of ...1868030201:0622); see the added line on pages versus crops.
- CONFIRMED | [printed-crescent] the San Francisco Times, from the pen of Mrs. H. A. Deming. | https://tile.loc.gov/image-services/iiif/service:ndnp:lu:batch_lu_psylocke_ver01:data:sn82015775:00200299899:1868030801:0467/4150,200,850,280/full/0/default.jpg | Full-resolution region: "The following remarkable little poem is a contribution to the San Francisco Times, from the pen of Mrs. H. A. Deming." The frozen crop starts at the second line of that sentence.
- CONFIRMED | [printed-argus] Mrs. H. A. Deming, of San Francisco, is said to have occupied a year in hunting up and fitting together the following thirty-eight lines from thirty-eight English poets. The names of the authors are given below: | https://tile.loc.gov/image-services/iiif/service:ndnp:iune:batch_iune_article_ver01:data:sn84038628:00295879932:1868030901:0492/2870,4800,780,420/full/0/default.jpg | Read: "Mrs. H A Deming, of San Francisco, is said to have occupied a year in hunting up and fitting together the following thirty-eight lines from thirty-eight English poets. The names of the authors are given below:" (the print drops the stop after H).
- CONFIRMED | [printed-rutland] The following remarkable little poem is a contribution to the San Francisco Times, from the pen of Mrs. H. A. Deming. | https://tile.loc.gov/image-services/iiif/service:ndnp:vtu:batch_vtu_leek_ver01:data:sn84022367:00415628353:1868031901:0097/3620,400,700,420/full/0/default.jpg | Read at full resolution, word for word, "Mrs. H. A. Deming".
- CONFIRMED | [printed-davenport] Mrs. H. A. Demming, of San Francisco, is said to have occupied a year in hunting up and fitting together the following thirty-eight lines from thirty-eight English poets. The names of the authors are given below : | research/the-newsquake/data/daily-davenport-democrat.jpg | Crop read: "Mrs. H. A. Demming, of San Francisco, is said to have occupied a year in hunting up and fitting together the following thirty-eight lines from thirty-eight English poets. The names of the authors are given below :".
- CONFIRMED | [printed-richmond] … eight English poets. The names of the authors are given below : | research/the-newsquake/data/richmond-palladium.jpg | Crop begins "eight English poets. The names of the authors are given below :". Above it the page prints "Mrs. H. A. Demming, of San Francisco, is said to have occupied a year..." (tile.loc.gov region 1150,800,940,400 of ...1868040901:0155).
- CONFIRMED | [printed-stjohnsbury] The following remarkable little poem is a contribution to the San Francisco Times, from the pen of Mrs. H. A. Deming. | research/the-newsquake/data/st-johnsbury-caledonian.jpg | Crop read: "The following remarkable little poem is a contribution to the San Francisco Times, from the pen of Mrs. H. A. Dem-ing."
- CONFIRMED | [printed-bellows] The following remarkable little poem is a contribution to the San Francisco Times, from the pen of Mr. H. A. Deming. | research/the-newsquake/data/bellows-falls-times.jpg | Crop read: "The following remarkable little poem is a contribution to the San Francisco Times, from the pen of Mr. H. A. Dem-ing."
- CONFIRMED | [printed-goldhill] Mrs. H. A. Deming, of San Francisco, is said to have occupied a year in hunting up and fitting together the following thirty-eight lines from thirty-eight English poets. The names of the authors are given below each line. | research/the-newsquake/data/gold-hill-daily-news.jpg | Crop read word for word, ending "given below each line."
- MINOR | [printed-findlay] The following remarkable little poem is a contribution to the San Francisco Times, from the pen of Mrs. H. A. Deming. | https://tile.loc.gov/image-services/iiif/service:ndnp:ohi:batch_ohi_iago_ver01:data:sn85042591:00280774972:1868050101:0150/930,1480,760,130/full/0/default.jpg | The Hancock Jeffersonian prints "is a contributation to the San Francisco Times, from the pen of Mrs. H. A. Deming." (a compositor's slip, legible in the frozen crop too). data.js gives this row the normalised SF_TIMES string and the record's excerpt copies the normalised wording, so the hand transcription is not as printed. Nothing numeric moves and the lineage is right.
- CONFIRMED | [printed-chicago] The following remarkable little poem is a contribution to the San Francisco Times, from the pen of Mrs. H. A. Deming. | research/the-newsquake/data/workingmans-advocate.jpg | Crop read: "[The following remarkable little poem is a contribu-tion to the San Francisco Times, from the pen of Mrs. H. A. Deming."
- CONFIRMED | [printed-clarksville] Mrs. H. A. Deming, of San Francisco, is said to have occupied a year in hunting up and fitting together the following thirty-eight lines from thirty-eight English poets. The names of the authors are given below: | research/the-newsquake/data/clarksville-chronicle.jpg | Crop read word for word ("Mrs. H. A. DEMING" in small capitals).
- CONFIRMED | [wording-bellows-mr] Bellows Falls prints Mr. H. A. Deming where Rutland and St. Johnsbury print Mrs. | research/the-newsquake/data/bellows-falls-times.jpg | Bellows Falls crop "from the pen of Mr. H. A. Dem-ing"; Rutland (full resolution) and St. Johnsbury (crop) "Mrs. H. A. Deming".
- CONFIRMED | [holdouts-one-sentence] All three name the San Francisco Times, in the same sentence, with one word different. | research/the-newsquake/data/rutland-weekly-herald.jpg | The three held-out sentences read identically except Mr./Mrs. (Rutland, St. Johnsbury and Bellows Falls lines above).
- CONFIRMED | [six-times] Six pages say the poem is a contribution to the San Francisco Times, which names a publication. | https://tile.loc.gov/image-services/iiif/service:ndnp:lu:batch_lu_psylocke_ver01:data:sn82015775:00200299899:1868030801:0467/4150,200,850,280/full/0/default.jpg | New Orleans, Rutland, St. Johnsbury, Bellows Falls, Findlay ("contributation") and Chicago, each read above.
- CONFIRMED | [four-of-sf] Four say Mrs. H. A. Deming, of San Francisco, which names a person's city. | research/the-newsquake/data/evening-argus.jpg | In the crops: Rock Island, Davenport ("Demming"), Gold Hill, Clarksville. The Washington and Richmond pages print the same wording just above their crops (see the added line), so six pages in all carry it.
- CONFIRMED | [two-partial] Two crops begin below the naming clause and show only the tail of the second sentence. | research/the-newsquake/data/richmond-palladium.jpg | Richmond and Intelligencer crops both begin below "of San Francisco" (crop text quoted above); "the second sentence" reads as the second wording.
- OBSERVED | [sf-times-unseen] No issue of the San Francisco Times has been seen. | oversight/waves/2026-09-10-foreign-instruments/specs/the-newsquake.md | A statement of what the author saw; the spec records the targeted searches ("H. A. Deming" "San Francisco Times" poem; site:cdnc.ucr.edu ...). I could not search further: CDNC and the Chronicling America title search both answer scripts with a Cloudflare "Just a moment..." page (HTTP 403).
- CONFIRMED | [hawaiian-unreviewed] the Hawaiian Gazette of 26 February | https://clusters.viraltexts.org/cluster/400732 | First witness: "The Hawaiian gazette. 1868-02-26", "Honolulu [Oahu, Hawaii]", "This text comes from a proprietary database and cannot be displayed due to copyright restrictions."; it is the earliest date among the 733 CSV rows.
- CONFIRMED | [d-cells-earthquake] The same threshold, run on the earthquake below, admits three cells out of 1,517. | public/strata/the-newsquake/results.js | home.grid cells 1517, originSetCells 3; the verifier recomputed it (PASS, 336 assertions).
- CONFIRMED | [d-chicago-41] locate a planted Chicago origin to 41 km | public/strata/the-newsquake/results.js | chicago-1day errorKm 40.50, state held; collision.json controls[0] "recovered_km": 41.
- CONFIRMED | [d-chicago-held] a planted Chicago source is recovered to 41 km with an origin set 187 km across, and the rule returns HELD. | public/strata/the-newsquake/results.js | chicago-1day setCells 9, diameterKm 186.57, state held.
- CONFIRMED | [d-sf-176-175] at one-hour precision the San Francisco set is 176 km across and misses the truth by 175 km. | public/strata/the-newsquake/results.js | sf-1h diameterKm 176.43, errorKm 175.16, containsPlant false, state broke.
- CONFIRMED | [d-sf-eight-east] eight of the nine sit east of the Rockies and the ninth is Gold Hill | public/strata/the-newsquake/data.js | Fitted longitudes: Gold Hill -119.66; the other eight lie between -77.02 and -90.60, all well east of the Rockies' eastern front (about 105 W).
- CONFIRMED | [d-nine-dates] A seismic location surface run on nine verified 1868 newspaper issue dates cannot find where a poem started | public/strata/the-newsquake/results.js | primary rows 9, originSetCells 12545 of 12545, verdict UNDERDETERMINED; each date checked against the cluster and the page's IIIF identifier.
- MINOR | [d-nine-towns-baseline] The ordinary baseline uses the earliest reviewed source line, in the New Orleans Crescent on 8 March. | https://tile.loc.gov/image-services/iiif/service:ndnp:dlc:batch_dlc_buford_ver01:data:sn83026172:00211678446:1868030201:0622/180,330,960,300/full/0/default.jpg | The Daily National Intelligencer of 2 March, a fitted page the layer reviewed, prints "Mrs. H. A. Deming, of San Francisco, is said to have occupied a year..." directly above where its crop begins. New Orleans is the earliest source line inside a frozen crop, not on a reviewed page. The baseline's answer (San Francisco) is the same either way.
- CONFIRMED | [d-grid] The map is 18 to 50 N and 161 to 65 W at 0.5 degree spacing. | public/strata/the-newsquake/data.js | CONFIG.grid latMin 18, latMax 50, lonMin -161, lonMax -65, step 0.5; 65 x 193 = 12,545 cells.
- CONFIRMED | [d-hawaii-atlantic] Your browser profiles origin day and nonnegative slowness at every half-degree cell from Hawaii to the Atlantic. | public/strata/the-newsquake/data.js | 161 W and 18 N take in all eight main Hawaiian islands (Niihau about 160.2 W, Ka Lae 18.9 N); 65 W lies in the western Atlantic.
- CONFIRMED | [d-conventions] Seed 400732 fixes 2,000 title-block bootstrap replicates. | public/strata/the-newsquake/data.js | CONFIG seed 400732, bootstrapReplicates 2000; results.js bootstrap replicates 2000.
- CONFIRMED | [d-onset] The onset window is 180 days from 26 February 1868. | public/strata/the-newsquake/collision.json | hand_conventions "The primary window is 180 days from 26 February 1868"; CONFIG.primaryWindowDays 180; 26 February is the Hawaiian Gazette date, the cluster's first.
- CONFIRMED | [d-nine-titles-once] Each of the nine fitted titles appears exactly once, so a title block and a row are the same object here | public/strata/the-newsquake/data.js | Nine distinct fitted titles in TARGET; the title ablation moves nothing (results.js shiftKm 0).
- CONFIRMED | [d-two-not-runnable] Two rows cannot be run at all, and say so instead of printing a figure. | public/strata/the-newsquake/app.js | NOT_COMPUTABLE holds cluster_id and date; results.js and collision.json mark both "computable": false.
- CONFIRMED | [d-30day] The 30-day window admits three issue pages and the fit has four free quantities, so there is no residual scale and no region to report. | public/strata/the-newsquake/results.js | windows[0] rows 3 (Washington 2 March, New Orleans 8 March, Rock Island 9 March), sufficient false; four quantities: latitude, longitude, origin day, slowness.
- CONFIRMED | [d-iter] The same optimizer reaches the earthquake minimum from all five of its starting points | public/strata/the-newsquake/results.js | home.solved startsAtBest 5, starts 5; newspaper iterative converged 1 of 25.
- WRONG | [d-bootstrap-only-frozen] This is the only figure on the page that is not recomputed in your browser | public/strata/the-newsquake/app.js | The onset-window chart and table are drawn from FROZEN_RESULT.windows (app.js lines 126 and 131, and 320 on resize) and the synthetic ladder from FROZEN_RESULT.synthetic (line 208); neither is recomputed in the browser. The ladder is labelled precomputed; the onset-window section is not. Fix at index.html line 199.
- CONFIRMED | [d-gold-miss] The predeclared date-fault rule is not relaxed after inspection. A miss is a miss, and it is reported as one in the transfer ledger. | oversight/waves/2026-09-10-foreign-instruments/specs/the-newsquake.md | Spec: "Detection succeeds only if Gold Hill becomes the largest absolute externally studentized residual and its absolute value exceeds 3.0 ... Do not redefine success after inspection." Ledger planted_fault_caught false; verifier --ledger prints planted_fault_caught=false.
- CONFIRMED | [d-uncertain-6471] admitting the eleven in-window rows the eligibility rules throw away moves the minimum 6,471 km, and admitting only the earliest of them, the Hawaiian Gazette of 26 February, moves it 9,952 km. | public/strata/the-newsquake/results.js | eligibility rows 20, shiftKm 6470.58; the verifier asserts Math.round(honoluluShift)===9952 and passed. The CSV has exactly eleven in-window issue pages outside the twelve (six closed, five open but not image-reviewed).
- CONFIRMED | [d-seven-names] seven of the nine fitted issues name San Francisco outright in their frozen crop ... and the two that do not are the two whose crop begins below the naming clause. | research/the-newsquake/data/richmond-palladium.jpg | Crops read: San Francisco appears in New Orleans, Rock Island, Davenport, Gold Hill, Findlay, Chicago, Clarksville; not in the Washington and Richmond crops. True of the crops; see the added line for the pages.
- CONFIRMED | [d-eleven-synth] Eleven planted origins are regenerated and re-adjudicated. | public/strata/the-newsquake/data.js | SYNTHETIC_PLANS has 11 entries; results.js synthetic has 11 rows.
- CONFIRMED | [d-six-cell-ledger] The check that stands behind this page recomputes the visible figures and the six-cell transfer ledger. | research/the-newsquake/verify-the-newsquake.mjs | Recreated: node verify-the-newsquake.mjs printed "PASS the-newsquake: 336 assertions"; --ledger printed six booleans.
- CONFIRMED | [d-images-twice] Every one of twelve image files exists twice with a pinned SHA-256, and the module and the frozen manifest are compared row by row. | public/strata/the-newsquake/data.js | Recreated: all 12 files in research/the-newsquake/data/ and public/strata/the-newsquake/images/ hash to IMAGE_HASHES (0 mismatches).
- CONFIRMED | [d-mutations] A wrong Earth radius, a moved date, an admitted holdout, a shifted depth digit and a swapped transcription must each make the check fail. | research/the-newsquake/verify-the-newsquake.mjs | Recreated: --mutate printed "CAUGHT a wrong Earth radius", "CAUGHT a moved issue date", "CAUGHT a holdout admitted to the fit", "CAUGHT the scan depth 0.41 km", "CAUGHT a transcription copied onto the wrong paper", exit 0.
- CONFIRMED | [d-twenty-five] The exhaustive profile adjudicates, and it needs no starting point, so this audit is reported beside the verdict and not folded into it. | public/strata/the-newsquake/engine.js | adjudicate(): the unstable test reads sufficiency, diameter, boundary, speed and deletion, not iterative.passes.
- CONFIRMED | [d-everything-else-passes] with dates generated from a real origin by this page's own model, everything else passes and only this check fails. | public/strata/the-newsquake/results.js | chicago-1day: 9 cells, 187 km, deletion 69.6 km, no boundary, contains the plant, held, but converged 23 of 25, and engine.js passes only at 25 of 25; chicago-1h is the same with 22 of 25. The source record looked only at the San Francisco plants and doubted the sentence; the Chicago rows bear it out.
- CONFIRMED | [f-chicago-cincinnati] Modern links can be wrong. This one points Chicago at Cincinnati. | https://clusters.viraltexts.org/cluster/400732?download_csv=400732 | CSV row: "The workingman's advocate.", "Chicago [Ill.]", 1868-05-30, Coverage "http://dbpedia.org/resource/Cincinnati".
- CONFIRMED | [f-fragments] A page is not an article, and the cluster holds several fragments of the same page. | https://clusters.viraltexts.org/cluster/400732?download_csv=400732 | Same issue-page URL on three Rutland rows, two Davenport, two Clarksville and two New Orleans rows (one by a chroniclingamerica URL for seq-3); data.js fragments 3, 2, 2, 2.
- OBSERVED | [d-last-verified] Last verified 10 September 2026, when it was written. | git show 7f3b133f1f:memory/log.d/2026-09-10T0115Z-nothing-survives-by-analogy.md | The wave log dated 2026-09-10 built the layer; no assay record for the slug exists at 7f3b133f1f (assay/records has none).
- CONFIRMED | [r-error-bar] Error-bar coverage asks whether intervals survive later recommendations | src/content/strata/error-bar-coverage.md | dek: "operate a declared audit of historical CODATA and PDG recommended-value intervals without treating today's recommendation as known truth".
- CONFIRMED | [r-leonardo] The Mona Lisa layer measures a coverage spike with rival paintings beside it | public/strata/a-minor-leonardo/index.html | "How often U.S. newspapers named each artwork, 1905-1920 ... The two rivals are the ones that outranked the Mona Lisa in pre-1911 Louvre guidebooks"; "a spike of dated coverage".
- CONFIRMED | [r-dead-reckoning] Dead Reckoning shows an estimate walking away when its motion model omits a current. | public/strata/dead-reckoning/index.html | "deliberately excludes the systematic current, so the actual error comes to exceed the cone radius".
- MINOR | [f-santa-rosa] Residual surface for the 1969 Santa Rosa earthquake | https://en.wikipedia.org/wiki/1969_Santa_Rosa_earthquakes | The home surface's accessible label (app.js line 57) names "the 1969 Santa Rosa earthquake", the name of the M5.6 and M5.7 pair of 1 October 1969 (2 October UTC, per the source). The event used is the small 5 October 1969 11:12 event of the report's "SOME SANTA ROSA QUAKES FOR TESTING HYP071" (PDF p.95 heading, "69/10/ 5"; collision.json "first test event, 5 October 1969").
- CONFIRMED | [d-box] a 0.4 by 0.36 degree box | public/strata/the-newsquake/data.js | HOME.box lon -122.90 to -122.50 (0.40), lat 38.30 to 38.66 (0.36); 37 x 41 = 1,517 cells.
- CONFIRMED | [d-no-single-point] the method cannot rule out a single point on the map it searched | public/strata/the-newsquake/results.js | primary originSetCells 12545 of cells 12545.
- CONFIRMED | [d-all-cells] Every cell is inside the origin set. | public/strata/the-newsquake/results.js | primary originSetFraction 1.
- CONFIRMED | [d-sf-corner] BOUNDARY CONTACT: the best cell is the south-east corner, so the search box truncates the estimate. | public/strata/the-newsquake/results.js | primary lat 18, lon -65 (south-east corner), boundary true.
- CONFIRMED | [d-baseline-hidden] Ordinary baseline hidden. The verdict does not change, because the baseline never enters the fit. | public/strata/the-newsquake/engine.js | The baseline is the San Francisco coordinate read from the source line; runPipeline uses it only to score, and the fit reads no text. The New Orleans row's date does enter the fit (ROLE_IN_FIT admits role baseline), which the sentence does not deny.
- CONFIRMED | [d-sf-1day-entirely] the same method finds a Chicago one to 41 km and loses a San Francisco one entirely | public/strata/the-newsquake/results.js | sf-1day: minimum 193 km from the plant but an origin set of 1,120 cells and 4,614 km, deletion shift 1,115 km, underdetermined. "Entirely" is the page's word for a rule that cannot confine it, not for the point error.
- CONFIRMED | [d-westward-flat] the distance pattern westward is nearly flat, so the surface has nothing to bite on | public/strata/the-newsquake/results.js | The page's explanation, borne out by the control: plants at San Francisco and Gold Hill (one day) give origin sets of 4,614 and 4,295 km; Chicago gives 187 km.
- CONFIRMED | [d-thirty-days] One fault changes an arrival by thirty days. | public/strata/the-newsquake/engine.js | mutationTests: Gold Hill "day:o.day+30".
- CONFIRMED | [d-text-fault] The other changes a name inside the transcribed text that travels with every fitted row | public/strata/the-newsquake/engine.js | mutationTests: ocr.replace('Deming','Darning') on the first fitted row with Deming (New Orleans), with carryOcr true for every row.
- CONFIRMED | [d-limits] 500 km predeclared maximum diameter | public/strata/the-newsquake/data.js | compactDiameterKm 500, deletionLimitKm 500; the spec's Held rule uses 500 km for both.
- MINOR | [d-atlantic] The 0.5 degree grid limits spatial resolution, and the search box excludes the Atlantic, so an English origin could never have been found. | public/strata/the-newsquake/data.js | lonMax -65 puts hundreds of kilometres of the western North Atlantic inside the box (the minimum cell itself, 18 N 65 W, is at sea), and the page says elsewhere "from Hawaii to the Atlantic". It excludes the eastern Atlantic and Europe; the English conclusion stands. Fix at index.html line 231.
- CONFIRMED | [d-free-choices] home tolerances 1.0 km, 0.02 s and one percent of the box. | public/strata/the-newsquake/data.js | HOME.tolerances epicentreKm 1.0 (twice the printed ERH 0.5), originSec 0.02, regionFraction 0.01.
- CONFIRMED | [d-urls-recorded] Exact page URLs and image-service URLs are recorded in the frozen extract. | research/the-newsquake/data/target-extract.json | Page URLs are in target-extract.json; the image-service URLs are in image-sources.txt beside it in the same frozen data directory, not in the JSON itself.
- CONFIRMED | [d-ablation-count] All ten adapter ablations and the combined eligibility ablation are re-run from the data and compared with the collision manifest | research/the-newsquake/verify-the-newsquake.mjs | Lines 252-269 call runAblation for all eleven and compare with collision.json; the two non-computable rows are checked to stay "not computable".
- CONFIRMED | Nothing is uploaded, tracked or fetched from a third party at runtime. | src/pages/privacy.astro | [f-live-beacon] Checker filed WRONG: the served page (curl 2026-10-18) loads "https://static.cloudflareinsights.com/beacon.min.js" and POSTs its path to /api/pulse, which the committed index.html does not carry (deploy injects them). Overturned by the directing instance: the human ruled on 2026-09-28 (board note 706d1d, coordination/tending.tsv) that the host is not a third party and that "no third-party request" sentences stand, read through /privacy/, which names both the counter and Cloudflare's analytics script.
- MINOR | (added) Both put digitized American newspapers on an operable geographic and temporal surface. | public/strata/a-minor-leonardo/index.html | Nearby layers, from src/content/strata/the-newsquake.md line 14. The Leonardo page has a time series of mentions per million pages, 1905-1920, and no geographic element: "geograph", "latitude", "longitude", "map" 0 hits in its HTML, against "coverage" 13 and "spike" 4 (positive control).
- MINOR | (added) Re-run on the cluster's own unreviewed coordinates, including the Chicago row that points at Cincinnati, 480 km away. | public/strata/the-newsquake/collision.json | Served manifest, ablations coverage note (line 87). From data.js Chicago 41.837045, -87.684939 to its Coverage point 39.1, -84.5125 is 405.7 km (engine greatCircleKm and an independent haversine agree).
- MINOR | (added) Admits the five open in-window witnesses that were never image-reviewed here, four of them on Archive.org rather than at the Library of Congress. | https://clusters.viraltexts.org/cluster/400732?download_csv=400732 | Served manifest, line 91. All five (Friends' Intelligencer, The Albion, Louisville Daily Journal, The Commonwealth, Sphinx) have archive.org URLs in the CSV, and data.js describes all five as "open on Archive.org".
- MINOR | (added) source attributions and author-place phrases, which name San Francisco on seven of the nine fitted pages | https://tile.loc.gov/image-services/iiif/service:ndnp:in:batch_in_indianapolisolympians_ver02:data:sn86058250:00271744808:1868040901:0155/1150,800,940,400/full/0/default.jpg | Served manifest, lines 35, 38 and 110 ("seven of the nine fitted pages"). The Richmond page prints "Mrs. H. A. Demming, of San Francisco" and the Intelligencer page "Mrs. H. A. Deming, of San Francisco" directly above their crops, so all nine fitted pages name San Francisco; seven is the count in the crops, as the visible page correctly says.
- MINOR | (added) the printed result card, transcribed AS SCANNED: '691005 1112 52.83 38-28.59 122-41.94 0.41 MAG 0.81 NP 19 ...' | https://pubs.usgs.gov/of/1975/0311/report.pdf | data.js lines 71 and 87 and research/the-newsquake/README.md line 94 give MAG 0.81. The PDF p.95 image prints the magnitude with a leading 1: glyph widths 10 px like the other 1s on the row (the 0s are 18 px), so "1.81" (the second digit has a filled upper loop, an 8 by shape). Magnitude is not shown on the visible page or used by the engine.

## Findings
- [fixed] only-figure-not-recomputed (K4, WRONG d-bootstrap-only-frozen): public/strata/the-newsquake/index.html said the frozen bootstrap was "the only figure on the page that is not recomputed in your browser", but app.js draws the onset-window chart and table from FROZEN_RESULT.windows (lines 126, 131, 320) and the control ladder from FROZEN_RESULT.synthetic (line 208). The sentence now names all three as read from the frozen results and recomputed by the independent check (verify-the-newsquake.mjs lines 221-238 and 275); a dated Corrected line is on the page and in the markdown body.
- [declined] station-delay-count (K1, MINOR hypo-station-delays-note): "twenty station delays" against a deck that prints 19 (SR17's field is blank, read as 0.0); no number moves. Listed by tier B.
- [declined] findlay-transcription-normalised (K2, MINOR printed-findlay): the Hancock Jeffersonian prints "contributation"; data.js stores the normalised wording. Lineage and numbers are right. Listed by tier B.
- [declined] earliest-reviewed-source-line (K3, MINOR d-nine-towns-baseline): New Orleans on 8 March is the earliest source line inside a frozen crop; the 2 March Intelligencer prints an of-San-Francisco line above its crop. The baseline's answer is the same. Listed by tier B.
- [declined] santa-rosa-event-name (K5, MINOR f-santa-rosa): the surface's accessible label names "the 1969 Santa Rosa earthquake", the usual name of the 1 October pair; the event used is the small 5 October 11:12 test event. Listed by tier B.
- [declined] search-box-excludes-atlantic (K6, MINOR d-atlantic): the box runs to 65 W, so it holds part of the western North Atlantic; it excludes the eastern Atlantic and Europe, and the English conclusion stands. Listed by tier B.
- [declined] leonardo-relates-geographic (K8, MINOR, relates note): the Leonardo page is a time series with no geographic element, against "an operable geographic and temporal surface". Listed by tier B.
- [declined] chicago-cincinnati-distance (K9, MINOR, collision.json line 87): the manifest says Cincinnati is "480 km away" from the Chicago row; the great-circle distance from data.js Chicago to the Coverage point is 405.7 km (recomputed by the fixer). Nothing the page computes uses it. Listed by tier B.
- [declined] five-witnesses-four-on-archive (K10, MINOR, collision.json line 91): "four of them on Archive.org" where data.js describes all five as open on Archive.org. Listed by tier B.
- [declined] seven-of-nine-fitted-pages (K11, MINOR, served manifest): the manifest drops the visible page's qualifier "in their frozen crop"; on the full pages all nine name San Francisco. Listed by tier B.
- [declined] home-card-magnitude-digit (K12, MINOR, data.js and README): the scan prints the magnitude as 1.81 by glyph width, against "MAG 0.81"; magnitude is neither shown nor used by the engine. Listed by tier B.