Tousey-Sinclair: Quality Engineering (Step 4b)

QE: Opus, 2026-10-09. Inputs: _text/manifest.json, all 62 digests (read in full, including the index chunks 58-62), the deep read, facts.json, claims.md, sites.md, research/ notes, physicist.md, mathematician.md, contrarian.md, arbiter.md, primer.md, example.py. Spot checks were made against the extracted text (_text/...42fec2.txt) and are tagged [read: p.N] with PDF page numbers.

Verdict: MINOR

The central verdicts survive, but four findings must change how the arbiter and the primer present claims 1, 2, 3 and 6. All four come from passages in the book that the fact-finder read only by "sampling" or as truncated digest lines:

1. The book contains its own dose calibrations, and the arbiter's dose analysis (D1) does not use them. The book's figures put the "author's dose" anywhere from about one quarter of an erythema dose to about one and a half erythema doses.
2. The book itself rejects the condenser (electrostatic) theory of x-ray burns. It cites Kienböck and others against it (p.1158) and says Tousey "does not regard this as probably correct" (p.1162). This answers one of the arbiter's open questions.
3. The arbiter misquotes the book on ultraviolet from the vacuum electrode. The book says the ultraviolet comes from sparks outside the glass and that it "is hardly to be supposed" that it plays a therapeutic part (p.557, 578). The "Refuted by calculation" verdict therefore answers a claim the book mostly disowns.
4. Tousey ran his own "output greater than input" experiment and explained it himself (p.155-156: hot-wire meter, apparent 6000 W from about 600 W, explained by the voltage collapse in an arc). The primer's worked example leaves this out, and its conclusion for an averaging meter needs revising.

None of these moves a claim to a different place on the scale for its main form. All of them can be fixed by updating facts.json, claims.md, a "QE revision" section in arbiter.md, and primer.md, so the right level is MINOR rather than a redo of Steps 2-4.

1. Coverage

• The manifest lists one document: the 1921 third edition of Medical Electricity, Rontgen Rays and Radium, 1361 pages, every page OCR'd. All 62 chunks have a digest; none is missing. Blank and figure-only pages are marked as such in the digests (for example p.62, 64-65, 227, 281, 381, 408, 1046, 1054, 1058, 1157, 1297). This is acceptable.
• Pages whose text layer is unreadable and that matter:
• p.1176 is a rotated table, "Erythema Exposures of x-Ray of Various Radiographic Strengths" in Tousey units. It has never been read from the page image. It bears on claims 1 and 5. Deep read 1 covers it.
• p.408 was handled by the deep reader from the page image (correct).
• Missing printed pages 282-283, 286-287 and 290-295 of the folder's scan: the second scan (IA medicalelectrici00tousiala) supplies them. They have been read only as noisy OCR, and pp.290-295 not at all. The second scan is not in the manifest, so it cannot carry a manifest slug. See the remediation list.
• No skipped files, so there is no skipped text.

2. Shortcuts

• Process defect (Step 1b). STEP 1b says the fact-finder reads ALL digests. facts.json records something less:
• Part 3 (PDF 80-240): "chunks 4-11 sampled: section headings and author-statement lines only".
• Parts 5, 9, 14, 15 and 17: "digest lines truncated at about 165-260 characters".

Every substantive item in section 3 below sits in one of these parts, or in Part 11's digest-only range (749-817). The shortcut was not harmless. - Claim 2 rests partly on a digest-only chapter. The x-ray burns chapter (PDF 1153-1162) holds the book's own verdict on the condenser theory. Only the digest of it was read. - Claim 1's dose reconstruction rests on outside data alone. The dosage chapter (PDF 1163-1180) and the Cabot generator pages (PDF 756-763) were digest-only. Those pages contain the book's conversions from its settings to erythema doses. - Claim 3 is acknowledged by the arbiter as digest-level (low confidence), and that is honest. But the arbiter's ultraviolet verdict quotes a sentence from the high-frequency chapter that nobody had deep-read, and the quote is out of context (see section 4). - Acceptable digest-only coverage: - static electricity (19-79); - most of dynamic electricity, apart from the items in section 3; - development, exposure tables, chest, abdomen and bone radiography (865-904, 975-1129); - the index.

None of these carries a central claim.

3. Missed material (read in the digests; absent from facts.json, claims.md, the reviews, the arbiter or the primer unless stated)

Ordered by importance. Each item was checked against the extracted text where marked [read].

3.1 The book's own dose calibrations (claims 1 and 5)

The "author's dose" is 6-inch spark, 3 mA, 10 inches, and either 5 minutes unfiltered (900 mA·s) or 10 minutes through 3 mm aluminium (1800 mA·s) [read: p.1184 note, "The author's own dose is 6-inch spark, 3 ma., five minutes at 10 inches distance"]. Converted to the book's own erythema figures, it comes out as follows:

Other dose passages in the same chapters bear on the same question and also go unused: - the Tousey-unit dose table, in which 500 Tousey meter-seconds is an ordinary erythema dose (p.1173-1175, digest); - the hyperhidrosis calibration of p.1190-1191 (already cited by the arbiter for the injury, not for the calibration); - the "least dose giving a visible skin effect on the face is about 3 H" (p.1148, p.1160-1161, digest).

Consequences:

• The arbiter's D1 says the meaning of the "(5)" H could be settled "only [by] his earlier editions or papers" and that no period calibration is on record. That is wrong: the book holds four conversions.
• The two independent period calibrations (Cabot and Dodd; the 57 kV figure) agree with the arbiter's modern estimate (sub-erythema, roughly 0.15-0.55 of the 2 Gy threshold). They support "rarely a visible reaction".
• Tousey's own heavy-dose conversion (p.1179) puts the same settings above an erythema dose. That contradicts "never causes dermatitis" in the book's own terms more strongly than the "(5) H" label does.
• The "Internally inconsistent (minor)" note on the H label should now say the book's own conversions span about 0.26 to 1.5 erythema doses for one set of settings.
• The verdict on "never" stays Never properly tested.

3.2 The book's own rejection of the condenser theory (claim 2)

• p.1158 [read]: "The experiments of Strater, Kienböck, Oudin, and Scholtz seem to leave no room for doubt that radiodermatitis is due to the x-rays and not to electric discharges and other physical phenomena which take place around the tube." The same page cites Bordier and Salvador on electrolytic action in x-ray erythema.
• PDF p.1162 (digest p.1161) [read]: "Geyser has suggested that the cause of dermatitis is not the x-ray, but an electrostatic condition ... The author does not regard this as probably correct, and recommends avoidance of overdosage."
• p.1231 (already cited): the theory "has not been proved" but "probably contains a certain element of truth".

So the condenser theory is Geyser's. Tousey reports it, cites the experimental record against it in one chapter, and concedes "an element of truth" in another. This changes three things: - The wording of claims.md claim 2. - The primer's lesson in section 7, which says "a plausible mechanism was held after a 1900 experiment had already set it aside". - The arbiter's open question ("Does the book anywhere cite Kienböck 1900 against the condenser theory? Not found"). The answer is yes, p.1158.

Also missed: p.681 [read], "A burn appearing within twenty-four hours after an x-ray exposure is doubtless due to ultraviolet rays generated by the x-ray tube. This source of injury may be suppressed by interposing any screen opaque to ordinary light." This is a third burn mechanism the book proposes. It is testable and not assessed.

3.3 What the book says about ultraviolet from the vacuum electrode (claim 3)

• p.557 [read]: "Violet-ray Treatment ... there is really no such treatment." "As the ultraviolet ray will not penetrate the thinnest piece of mica or celluloid, the results ... are due to the form of current applied, and not so much to the ultraviolet rays, which are given off from the sparks outside."
• p.578 [read]: the visible light "has no demonstrable physiologic or therapeutic effect". The ultraviolet comes from "minute sparks ... passing along the outside of the tube". "It is hardly to be supposed that the ultraviolet rays absorbed by the body during a treatment by means of vacuum electrodes takes any more direct part in the physiologic or therapeutic effect."
• Against this, p.611 and p.631 [read] recommend electrodes that give "the greatest amount of ultraviolet radiation, as detected by the fluorescence caused in a piece of Willemite". p.631 credits a germicidal effect in acne partly to "the ultraviolet ray produced around the vacuum tube". p.635 says the same for skin disease (digest).
• p.657-658 and p.677-678 (deep read, Part 10): Tousey used Willemite fluorescence as a detector, and checked it with glass, which makes ultraviolet fluorescence "disappear entirely".

So the book's position is that the ultraviolet arises in the air outside the glass, and that glass and mica block it. That is the physicist's own physics. The strong claim the arbiter refutes ("germicidal UV from a soda-lime glass vacuum electrode") is not the book's claim. What the book does claim, a germicidal or tonic effect of ultraviolet and ozone from external sparks in acne and skin disease, is untested.

3.4 Tousey's own over-unity experiment and meter discussion (claim 6, the primer's worked example)

• p.155 [read]: a D'Arsonval milliammeter "indicates in a general way the average strength and direction of the secondary current ... Such a meter may indicate 0 when the x-ray tube is giving a radiance by which a moderately good picture may be made".
• p.156 [read]: Tousey's experiment. A hot-wire milliammeter rated to 650 mA, 12-inch coil, Wehnelt interrupter, 6 A primary at 110 V, a 4-inch flame: "a voltage of 40,000 and the milliamperemeter indicated an intensity of 150 milliamperes. Here there was an apparent power of 6000 watts produced by the application of about 600 watts ... The apparent discrepancy is due to the fact that the voltage is greatly reduced during the passage of the continuous flame, just as 100 volts will maintain an electric arc 1 inch long which it would take 10,000 volts to originate."
• p.214 (digest): hot-wire meters on alternating current "read somewhat more than the true average", and "volts times amperes can exceed the power used to run the dynamo".

Consequences:

• The book does discuss which meters read what.
• Tousey's own explanation is the right physics: the voltage needed to sustain an arc is far below the voltage needed to strike it.
• Tousey's own analogy implies a voltage collapse by up to about 100 times. That closes the ledger even with an averaging meter.
• The arbiter's D2 and the primer's Step 5 ("with an averaging meter, Tousey's 'less than one-half' voltage drop is not enough") rest only on the p.740 Cabot-converter measurement. They should add p.156.
• This is a second, fully documented "apparent output above input" passage. It belongs in claims.md claim 6 and in the primer.

3.5 Claim 3: supporting evidence omitted while contrary evidence is kept

• p.603 [read]: "Apostoli, at the Twelfth International Medical Congress, reported success in over 500 cases [diabetes; cage or couch, average twenty-five treatments]. Following Apostoli, several of his pupils ... Williams, Allen, and the present author." Against: "Boedeker, Cohn, Loewy, Doumer, and more recently Halfon ... no reduction in the amount of urine".
• The arbiter and the primer list the contrary authors (Cohn, Boedeker, Vinaj, Denoyés, German authors) but not Apostoli's 500-case report. SOURCE HONESTY requires both sides.
• p.604 [read]: Tousey's own caveat, that his diabetic patients also had medicines (arsenic), so his cases do not show the electric treatment alone works. This fair hedge by the inventor is also absent.
• p.595 (digest): Gay's 12 autoconduction cases with pulse, pressure, urea and uric-acid figures (for example pressure 180 to 135 mm Hg in fifteen treatments). Also absent.

3.6 Claim 7: the book's own distinction between signal speed and ion drift

• p.289 [read]: "The rate of progress of the wave of low potential, not over 30 meters a second ... does not indicate the rate of transmission of the ... current"; "The velocity at which the current flows, possibly 180,000 miles a second".
• p.296 [read] attributes the 39 m/s view to "many observers".

So Tousey states the correct distinction himself, three pages before he reports the observers' view. The primer's section 7 lesson ("take a true premise ... one step too far") presents the 39 m/s error as Tousey's. The primer's own correction 3 is right, but the body text was not changed to match.

• p.330 (digest) gives the book's mechanism of stimulation: Nernst's view of semipermeable membranes, ionization above a threshold, and osmotic pressure. This supports the arbiter's "loose ionic" reading of "electrolysis" (D5) from the book's own text, not only from the second scan.

3.7 Tousey's own experiments not recorded anywhere

• Pregnant-cat experiment [read: p.1142-1143]: mild radiographic exposures from 6 Oct to 17 Dec 1904, with a dated settings table. The result was "three mature and perfectly formed kittens ... born dead". The exposures stopped, and four healthy kittens followed. The author's hands were in the beam ("this was a dangerous experiment"). This is a primary observation of embryo radiosensitivity by the inventor.
• Shoe-conductivity experiment, 12 May 1919 (p.360, digest; facts.json mentions it only as an OCR fix). This is an electrical-safety measurement.
• Radiated-saline experiment of 1904 (p.1282-1283, digest): no radioactivity was detected, and there were no effects in animals.
• Oyster-pearl radiography (p.1141): 121 oysters per plate; the method was used in Ceylon.
• Millener's 22 power-house cases (p.302-303, digest), with Tousey's own seven-year exposure note. These are health effects attributed to stray fields.

These items are peripheral to the verdicts, but they are pertinent and are in the inventor's own text.

3.8 Investigator leads dropped from the arbiter and primer

• Tousey's own papers are held at two archives: the New-York Historical Society (a typescript from 1911) and the University of Pennsylvania Dental Library (Hellman reprints, box 29, folder 40, "Tousey, Sinclair, 1906-1920") [web via Investigator, sites.md]. Only mathematician.md mentions them. They are the nearest records that could decide the contact-tube and dental-arch priority questions, and they belong in the arbiter's "What would settle it" and in the primer's further reading.
• Thiosinamin, which Tousey introduced in 1894 (book p.634, 1189-1190; the 1896 monograph found by the Investigator), is his own pharmaceutical contribution, and the book claims cures with it. It is absent.

4. Accuracy spot checks (against the extracted text)

5. Retention check

The question here is whether pertinent material was left out of the arbiter or the primer because a reviewer judged it negatively.

• Evidence for claim 3 dropped while evidence against was kept. Apostoli's "over 500 cases" (p.603) and Gay's 12 measured cases (p.595) are omitted; the contrary authors from the same page are listed. This is a one-sided condensation and is a defect under SOURCE HONESTY ("evidence against a claim is also stated" cuts both ways). Fix: add both, as reports by named workers with no controls, beside the contrary reports.
• Tousey's own hedges dropped. These are:
• the condenser theory "not ... probably correct" (p.1162);
• ultraviolet "hardly to be supposed" to act (p.578);
• "did not stop medicines" (p.604);
• the 39 m/s view being the observers' (p.296, with p.289).

The omissions make him look more committed to the rejected positions than his text is. This is the mirror image of the "keep pertinent material" rule. The arbiter has already set the pattern for correcting this, in its correction of the contrarian on "deep action of ions". - Investigator leads. Most are retained in the primer's "Material that stays in the record" list (patents, the NYMJ 1913 generator article, the Tousey unit papers, the hand keratoses, the AJMS 1911 review, the 1916 orthodontia review, the name confusion, Geyser and the AMA). The archival holdings (NYHS, University of Pennsylvania) and thiosinamin are missing; see 3.8. These were not dropped for a negative judgement, but they should be added. - Negative material kept, correctly. The primer keeps the material that cuts against Tousey: - the hyperhidrosis injury; - the AJMS 1911 criticism; - Geyser's AMA trouble, labelled as bearing on weight and not as refutation.

This is done properly.

6. Process

• Source tags. Present and mostly honest throughout the arbiter and primer. Exceptions:
• The misquote in arbiter D4 (section 4, row 3).
• Arbiter D2(ii) cites p.740 as a coil measurement; the primer's correction 2 already notes it was the Cabot converter.
• Primer section 4 cites "PDF p.585" for the 39 m/s neighbourhood correctly, but section 7 does not carry the attribution to "many observers".
• Web checks.
• Reviewers: physicist 7, mathematician 3, contrarian about 3.
• Arbiter: 4 of 8, which covers tube output, the IAEA erythema threshold and glass UV cut-off.
• Still [memory, medium] where a verdict leans on it: backscatter ×1.3 (D1); the air attenuation coefficients behind "slightly faster than inverse square: Plausible, untested"; the Oudin power factor of 0.1; Moyers-type mixed-dentition analysis. The arbiter states that none changes a verdict. For the air-attenuation verdict this is arguable: the verdict rests on that calculation. One web check of the NIST mass-attenuation value for air at 20-30 keV would settle it.
• Verdict scale. Mostly consistent. Four lapses:
• "Pulmonary tuberculosis little permanent cure: Tested, not supported (as treatment)" scores x-ray as a TB treatment, not Tousey's claim. His claim, that x-ray gives little permanent cure, is consistent with the outcome and would read Established, as the leukaemia line does. State both.
• "Arch formula: geometry consistent" is not a scale verdict. Say "Internally consistent; Plausible, untested as anatomy; Never properly tested as a predictor", or similar.
• "Germicidal UV from a soda-lime vacuum electrode: Refuted by calculation" targets a claim the book disowns (3.3).
• "39 m/s: Internally inconsistent" attributes the inconsistency to Tousey, although the book separates his view (p.289) from the observers' (p.296). "Refuted by calculation (the observers' view as reported)" is the accurate label.
• Primer against STEP 4 and the reader profile. All nine sections are present in order. Symbols are defined at first use, there are worked numbers, five check-yourself questions with answers, further reading, and example.py. Undergraduate level is respected. Bremsstrahlung, Duane-Hunt, power factor, mobility and SAR are all introduced from scratch. The primer's own "Corrections" section is good practice.
• example.py. It runs cleanly (exit 0, project venv) and reproduces Parts A-D of the primer.

7. Remediation actions

Deep reads (most important first; at most 8)

1. Opus, ...42fec2, PDF pp.1130-1180 (x-ray effects, burns, precautions, dosage), with page images for pp.1173-1177 (the Tousey-unit erythema table; p.1176 is rotated). Question: - Set out every dose statement that converts settings (spark, mA, time, distance, filter) or Tousey units into H, Sabouraud tints or "erythema dose", with page. Reconcile them with the author's dose (6-inch spark, 3 mA, 10 in, 5 min unfiltered or 10 min through 3 mm Al). - Quote the book's position on causes of x-ray dermatitis: Strater, Kienböck, Oudin and Scholtz (p.1158); Bordier and Salvador; Geyser (p.1161-1162). - Quote the operator injuries and the pregnant-cat experiment (p.1142-1143), with the settings table.
2. Sonnet, ...42fec2, PDF pp.557-558, 577-579, 611, 631, 635, 657-659, 677-678, 681. Question: what exactly the book says about violet and ultraviolet light from vacuum electrodes and x-ray tubes (where it arises, whether glass or mica pass it, the Willemite and glass tests, the therapeutic role claimed or denied), and the p.681 claim that early burns are due to tube ultraviolet. Quote each statement in context.
3. Sonnet, ...42fec2, PDF pp.749-773. Question: - The Cabot and Shaw generator history as the book tells it: the first-edition reference (p.663 of the 1910 edition), the heterodyne, the constant-voltage generator, its settings. - The erythema calibrations: Cabot and Dodd's 3400 mA·s at 10 in (p.759 at about 90 kV and p.763 at No. 4 Benoist; resolve which), and the 57 kV, 4 mA erythema figure (p.762). - Tousey's remarks on mA·s as a dose measure.
4. Sonnet, ...42fec2, PDF pp.144-162 and 213-217. Question: how the book describes the secondary current of an induction coil and the meters that read it (D'Arsonval average, hot-wire). Give Tousey's own apparent-power experiment (p.155-156: 40 kV, 150 mA, 6000 W against about 600 W) and his arc explanation, and the p.214 statement on hot-wire meters and power. Quote with figures.
5. Sonnet, ...42fec2, PDF pp.544-647 (high-frequency currents and diathermy; requested by the arbiter too). Question: tabulate every clinical result with its source (named author or "the author"), number of cases, measurement (blood pressure, sugar, urea), co-treatments, follow-up, and any contrary report. Cover Apostoli's 500 cases and the contrary authors (p.603), Gay's 12 cases (p.595), Guilleminot, Moutier, Gidon, Delherm and Laquerrière, and Tousey's own cases and caveats (p.585, 597-607).
6. Sonnet, ...42fec2, PDF pp.298-349 (physiologic effects, muscle and nerve). Question: what the book gives as the mechanism of electrical stimulation (electrolysis, membrane and Nernst view, osmotic pressure at p.330; the energy law for condenser discharges at p.306; high-frequency non-excitation at p.302, 326, 339), and how this squares with the "almost all effects are electrolysis" statement on p.297.
7. Sonnet, ...42fec2, PDF pp.1252-1297 (radium). Question: the Tousey unit applied to radium (p.1259-1261), the book's radium-to-H conversions (p.1281-1282), Tousey's forearm self-experiment (p.1270, Plate 15) and the 1904 radiated-saline experiment (p.1282-1283). Quote figures (bears on claim 5).

Not a manifest document, so not in the structured list: deep read of printed pp.282-296 of the second 1921 scan (research/pages/ia-1921-ucla.txt, from line 17042), with page images from the Internet Archive copy if the owner approves the download. Question: the electrotonus and concentration-cell account, and whatever pp.290-295 contain (claim 7).

facts.json

• Part 3 entry:
• Add the p.155-156 hot-wire experiment and arc explanation, and the p.155 D'Arsonval note.
• Correct "how_read" to state honestly that digests were sampled; after deep read 4, update it.
• Part 9:
• Add Apostoli (500 cases) and the full contrary list (p.603), Gay's 12 cases, and the p.604 caveat.
• Replace "ozone and ultraviolet from the glass electrode as the active 'tonic'" with the book's actual position (3.3).
• Part 10: add p.681 (ultraviolet as the cause of early burns) and the Willemite and glass test.
• Part 11: add the p.759, 762 and 763 calibrations.
• Part 15:
• Add p.1158 and p.1161-1162 (rejection of the condenser theory), p.1167 and p.1179 (dose conversions), and the cat experiment.
• Update "how_read" after deep read 1.
• Part 17: add the radium-to-H conversions and Tousey's forearm self-experiment.

claims.md

• Claim 1: add the book's own conversions (3.1), from about 0.26 to 1.5 erythema doses for the author's settings.
• Claim 2: restate. The condenser theory is Geyser's. Tousey cites the experiments against it (p.1158) and judges it "not ... probably correct" (p.1162), while elsewhere allowing "a certain element of truth" (p.1231). The claim that is actually his is the contact tube's superiority and priority. Add p.681.
• Claim 3: remove "acting through ... ultraviolet from the glass vacuum electrode". Replace it with the book's position: ultraviolet and ozone come from sparks outside the electrode, the therapeutic role of ultraviolet is denied in one place (p.578) and claimed for acne and skin in another (p.631, 635).
• Claim 6: add the p.156 experiment (apparent 6000 W from about 600 W) and its arc explanation.
• Claim 7: add that p.289 gives the correct distinction between signal and ion speed in Tousey's own voice.

arbiter.md (add a "QE revision" section; do not delete the earlier text)

• D1: add the in-book calibrations. Conclude that two independent period calibrations put the dose at about a quarter to a half of an erythema dose, and Tousey's own heavy-dose conversion puts it above one. Say so in the claim 1 verdict note ("Internally inconsistent" in the book's own units, now not "minor"). "Never" remains Never properly tested.
• Claim 2:
• Answer the open question (Kienböck is cited, p.1158).
• Attribute the theory to Geyser.
• The verdict on the theory (Tested, not supported) stands. Add that the book itself reports that test.
• Add p.681 as a further proposed mechanism: Never properly tested.
• D4 and claim 3:
• Correct the citation (p.578, not 585) and the misquote.
• Re-target the verdict. "UV leaves through the glass": the book says no, which agrees with physics; no verdict against Tousey.
• "Germicidal or tonic UV and ozone from external sparks (acne, skin)": Never properly tested.
• Keep the soda-lime calculation as explanation.
• D2 and claim 6: add p.156. Tousey's arc explanation implies a voltage collapse large enough to close the ledger with any meter. The quantitative reconciliation is therefore in the book in principle, though not with numbers for the 100 kV, 15 mA case.
• Claim 3, evidence: add Apostoli's 500 cases and Gay's 12 cases beside the contrary authors, and Tousey's p.604 caveat.
• Claim 7:
• Change "39 m/s: Internally inconsistent" to "the observers' view as reported: Refuted by calculation; Tousey's own statement (p.289) is correct".
• Add p.330 to D5.
• Scale fixes: TB line (state Tousey's claim and the treatment separately); arch formula (use scale names).
• "What would settle it": add the NYHS Tousey papers and the University of Pennsylvania Hellman reprints (box 29, folder 40).

primer.md

• Section 5 (worked example):
• Add Tousey's own experiment (p.156) and his arc analogy as Step 3b.
• Revise the conclusion "does not by itself work for an averaging galvanometer" to say that his p.156 explanation does work, because an arc sustains itself at a small fraction of its striking voltage.
• Keep the duty-cycle analysis as the hot-wire case.
• Note that p.155 shows Tousey knew the D'Arsonval meter reads the average.
• Section 5b: add the book's own conversions beside the modern estimate (a small table), and say they disagree with each other by a factor of about 6.
• Section 4 table and section 7:
• Re-attribute the condenser theory (Geyser's; Tousey reports the evidence against it).
• Re-attribute the 39 m/s view ("many observers"; Tousey's p.289 is correct).
• Restate the ultraviolet row.
• Section 6 evidence table:
• Revise row 3c (ultraviolet).
• Add Apostoli and Gay to 3b.
• Add a row for p.681 (early burns from tube ultraviolet: Never properly tested).
• "Material that stays in the record": add the pregnant-cat experiment, the archival holdings (NYHS, University of Pennsylvania) and thiosinamin.
• example.py: optionally add a Part E converting the author's dose with the book's four in-book calibrations (3.1).

Corrections to earlier documents

1. arbiter.md, D4: "[the book says] that 'ultraviolet rays absorbed by the body' matter [read: PDF p.585 region, text lines 32139-32166 ...]". The lines are on PDF p.578. The sentence reads: "It is hardly to be supposed that the ultraviolet rays absorbed by the body during a treatment by means of vacuum electrodes takes any more direct part in the physiologic or therapeutic effect" [read: p.578]. p.557 adds that ultraviolet "will not penetrate the thinnest piece of mica or celluloid" and comes "from the sparks outside" [read]. Correct the quote and the attribution.
2. arbiter.md, open questions, "Does the book anywhere cite Kienböck 1900 against the condenser theory? Not found". It does: "The experiments of Strater, Kienböck, Oudin, and Scholtz seem to leave no room for doubt that radiodermatitis is due to the x-rays and not to electric discharges" [read: p.1158]. Tousey also writes of Geyser's electrostatic theory: "The author does not regard this as probably correct" [read: PDF p.1162].
3. arbiter.md, D1, and open question "What did the '(5)' H in his table mean ... Only his earlier editions or papers could say": the book itself converts settings to erythema doses on p.759, 762, 763, 1167 and 1179 [read] (section 3.1).
4. arbiter.md, D2, and primer.md section 5, Step 5: the claim that the ledger does not close with an averaging meter on the book's evidence omits Tousey's own experiment and explanation on p.156 [read], where the voltage is "greatly reduced during the passage of the continuous flame, just as 100 volts will maintain an electric arc 1 inch long which it would take 10,000 volts to originate".
5. claims.md, claim 2: "X-ray dermatitis is partly a 'condenser' (electrostatic) effect ... ; Tousey holds the theory 'has not been proved' but 'probably contains a certain element of truth'". This is incomplete. Add p.1158 and p.1162 (above). The theory is Geyser's.
6. claims.md, claim 3, and facts.json Part 9: "acting through ozone and ultraviolet from the glass vacuum electrode". The book says the ultraviolet comes from sparks outside the electrode and largely denies it a therapeutic role (p.557, 578), while crediting it in acne and skin disease (p.631, 635). Restate.
7. primer.md, section 7: - "The condenser theory of burns. A plausible mechanism was held after a 1900 experiment had already set it aside." The book reports the experiment against it and rejects it (p.1158, 1162); Geyser held it. - "39 m/s and 'almost all effects are electrolysis.' Both take a true premise ... one step too far." The 39 m/s inference is reported from "many observers" (p.296); Tousey's own p.289 states the correct distinction.
8. arbiter.md and primer.md, claim 3 evidence: add Apostoli's report of "success in over 500 cases" and the named successes (Williams, Allen, the author), stated beside the contrary authors from the same paragraph [read: p.603].
9. facts.json, Part 3 "how_read": it states the digests were sampled. STEP 1b requires all digests to be read. The fact-finder's summary of Part 3 therefore missed p.155-156 and p.214. This is noted so that later runs do not treat Part 3 as fully read.

Recheck

Recheck by Sonnet, 2026-10-09, against the current facts.json (valid JSON), claims.md, arbiter.md, primer.md and example.py. Method: grep of each cited page and phrase, plus reading the changed passages and running example.py.

Residual notes, not defects: - The first-pass wording is intentionally retained in places in arbiter.md and the primer's correction log; each is marked superseded. - claims.md claim 3 still opens with the clause "acting through ozone made by the glass vacuum electrode". That is the book's own ozone claim (p.579), and the ultraviolet part is now correctly separated, so it stands. - Primer 5b says the book's conversions disagree by "a factor of about 6". example.py shows 5.7 unscaled and 9.4 scaled, so the figure is correct for the unscaled case.

Final: PASS