---
slug: there-is-no-white
passed: 2026-10-14
by: the assay line (research/assay-line/cited-1014b/), Codex fixer directed by claude-assaying-boyle-1014; source record by Codex, checked by Codex
checked: 96
confirmed: 59
wrong: 36
unverifiable: 0
observed: 0
stale: 1
---

This pass accounts for all 39 directing triage decisions for there-is-no-white, with 18 WRONG items fixed and three WRONG items open for remaining copies outside this layer's edit scope (X7, X15 and X37), 15 listed MINOR items retained, two findings rejected with evidence, and one JSON-LD date updated under the stated exception. It also retains the 57 CONFIRMED claims the checker re-read in research/assay-line/cited-1014b/there-is-no-white/N2-codex.md. The fixer independently re-read the CIE, NASA, ICC, Adobe and lightness-paper evidence for the edits, compared the embedded data with the official CIE table, reproduced the original solver failures and E/D65 mismatch, and checked the corrected live and offline calculations. MINOR lines count in wrong, giving 21 WRONG plus 15 MINOR; passed is the date of this pass, not a claim that the listed issues vanished.

## Claims
- WRONG | X1: “white is no colour at all (true of paint).” | https://cie.co.at/eilvterm/17-22-049 | I re-read https://cie.co.at/eilvterm/17-22-049, definition and Note 1: "perceived colour devoid of hue" includes white. I also read https://www.adobe.com/creativecloud/design/discover/is-black-a-color.html, the section on whether black and white are colours: it reports a shade convention, not a physical rule that paint has no colour. At f5e5f84f34:public/strata/there-is-no-white/index.html:176 the parenthesis "true of paint" endorses the assertion rather than merely reporting an internet answer. White paint has achromatic surface colour; lack of hue is not lack of colour. Correct that endorsement and its affirmative copies, and add a dated Corrected line. X1 and X2 name the same sentence and error; X25 is a different, non-endorsing report. Fixer reread and disposition: Corrected the affirmative paint claim in the page and catalogue, the OG endorsement, and the nearby card via this layer's own relates note. White paint is an achromatic colour, without hue. The dated paint correction is mirrored in the markdown.
- WRONG | X2: “white is no colour at all (true of paint).” | https://cie.co.at/eilvterm/17-22-049 | I re-read https://cie.co.at/eilvterm/17-22-049, definition and Note 1: "perceived colour devoid of hue" includes white. I also read https://www.adobe.com/creativecloud/design/discover/is-black-a-color.html, the section on whether black and white are colours: it reports a shade convention, not a physical rule that paint has no colour. At f5e5f84f34:public/strata/there-is-no-white/index.html:176 the parenthesis "true of paint" endorses the assertion rather than merely reporting an internet answer. White paint has achromatic surface colour; lack of hue is not lack of colour. Correct that endorsement and its affirmative copies, and add a dated Corrected line. X1 and X2 name the same sentence and error; X25 is a different, non-endorsing report. Fixer reread and disposition: Corrected the affirmative paint claim in the page and catalogue, the OG endorsement, and the nearby card via this layer's own relates note. White paint is an achromatic colour, without hue. The dated paint correction is mirrored in the markdown.
- MINOR | X3: “it is the brightest neutral your eye has currently decided on.” | https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/ | I read the abstract and introduction of https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/, retrieved through the search index after the direct page returned a browser challenge. The abstract says "highest luminance surface appears nearly white" in the tested simple surface displays. The summary at f5e5f84f34:public/strata/there-is-no-white/index.html:176 turns that heuristic into a general definition. This is MINOR scope/precision wording. Tier C: list; its sentence is separate from the paint assertion being fixed in the same paragraph.
- WRONG | X4: “combine every wavelength and you get white.” | https://cie.co.at/eilvterm/17-23-023 | I re-read https://cie.co.at/eilvterm/17-23-023, equi-energy spectrum definition: "constant throughout the visible region". At f5e5f84f34:public/strata/there-is-no-white/index.html:181 the unquoted explanation omits relative power, whereas its actual recipe at line 561 says "all wavelengths, equal energy". Presence of every wavelength alone does not specify white: adding sufficient power at one wavelength changes the tristimulus ratios. Qualify the unquoted explanation with the required spectral balance, including equal energy for the E example. Do not rewrite the words of the schoolbook saying. Add a dated Corrected line. Fixer reread and disposition: Qualified the light recipe by spectral power balance and separated real white surfaces from the perfect-reflector ideal. The dated light-and-surface correction is mirrored in the markdown.
- WRONG | X5: “absorbs nothing, reflecting all of it back.” | https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/ | I re-read the introduction of Radonjic and Gilchrist, Lightness perception in simple images, https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/, through the search index after the direct page returned a browser challenge. It says a typical white surface "reflects ∼90% of incident light". The unqualified definition at f5e5f84f34:public/strata/there-is-no-white/index.html:181 instead requires no absorption and complete reflection. That is an ideal reflector, not a requirement for real white surfaces. Correct the physical claim while distinguishing the ideal neutral-paper model, and add a dated Corrected line. Fixer reread and disposition: Qualified the light recipe by spectral power balance and separated real white surfaces from the perfect-reflector ideal. The dated light-and-surface correction is mirrored in the markdown.
- WRONG | X6: “Both are real facts about two different systems” | https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/; https://cie.co.at/eilvterm/17-23-023 | I re-read f5e5f84f34:public/strata/there-is-no-white/index.html:181 and both cited scientific sources. The introduction of https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/ says a typical white surface "reflects ∼90% of incident light"; https://cie.co.at/eilvterm/17-23-023 requires an equal-energy spectrum to be "constant throughout the visible region". Thus the sentence endorses the two unqualified assertions identified in X4 and X5, which are not both physical facts as written. Correct the endorsement consistently with those explanations and include it in the dated Corrected account of those errors. Fixer reread and disposition: Qualified the light recipe by spectral power balance and separated real white surfaces from the perfect-reflector ideal. The dated light-and-surface correction is mirrored in the markdown.
- WRONG | X7: “No beam of light is white.” | https://spaceplace.nasa.gov/review/magic-windows/vis.html | I re-read NASA, The Magic Visible Light Window, https://spaceplace.nasa.gov/review/magic-windows/vis.html, final paragraph: "White light goes in" when describing prism separation. The source explicitly allows white light made of multiple wavelengths. The statement at f5e5f84f34:public/strata/there-is-no-white/index.html:184 denies white beams altogether, although the computation concerns single wavelengths and specified reference whites. Correct that inference everywhere it is asserted, limiting the wavelength claim to the stated model, and add a dated Corrected line. This shares the error with X21, X27, X31, X32, X33 and X35. Fixer reread and disposition: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- MINOR | X8: “Every point on the rim is one wavelength.” | https://cie.co.at/eilvterm/17-23-058; git show f5e5f84f34:public/strata/there-is-no-white/index.html | I read https://cie.co.at/eilvterm/17-23-058, which defines the purple boundary using mixtures, and f5e5f84f34:public/strata/there-is-no-white/index.html:476-481, which draws the spectral curve and the faint dashed purple closure separately. The caption at line 195 uses rim loosely for a diagram whose closed boundary includes both. The intended spectral-curve statement is supported, so retain MINOR precision, not a false claim about the actual drawn curve. Tier C: list; no WRONG fix shares this sentence.
- MINOR | X9: “"White is all the colours mixed together"” | https://www.adobe.com/creativecloud/design/discover/is-black-a-color.html | I read https://www.adobe.com/creativecloud/design/discover/is-black-a-color.html: its colour/shade section says "white light comprises all hues", not the exact schoolbook wording. I also read f5e5f84f34:public/strata/there-is-no-white/index.html:218, which presents an unattributed familiar saying rather than an Adobe quotation. The defect is N1 schoolbook-quotation being marked as a sourced verbatim quote. Tier C: list this MINOR record issue; classify or source the saying accurately in the record. Do not edit the words inside the quotation, and do not characterize it as a fabricated Adobe quotation.
- WRONG | X10: “Below are three recipes for the identical white.” | git show f5e5f84f34:public/strata/there-is-no-white/index.html; https://cie.co.at/eilvterm/17-23-008 | I re-read f5e5f84f34:public/strata/there-is-no-white/index.html:393-395 and 554-580. The full-spectrum branch assigns "result=E"; the screen branch assigns "result=D65"; the two-line branch targets E. Evaluating those pinned helpers gives E=(0.3330989865, 0.3333710786), D65=(0.31272, 0.32903), separated by 0.0208362198 in xy. I also read https://cie.co.at/eilvterm/17-23-008, definition: metamers have "the same tristimulus values". The three recipes promised at f5e5f84f34:public/strata/there-is-no-white/index.html:218 therefore do not produce the claimed identical white. Match their target chromaticity and luminance before calling them metamers, or accurately describe them as different reference whites. Correct all copies of the identical-white claim and add a dated Corrected line shared with X13, X14 and X28. Fixer reread and disposition: Kept the demonstrated reference whites and named them accurately: full spectrum and two-line recipes target E, screen primaries target D65. Metamerism is conditional on equal XYZ, including luminance. The offline check now compares unrounded XYZ at Y=1; the dated recipe correction is mirrored in the markdown.
- WRONG | X11: “fall exactly on the white point.” | git show f5e5f84f34:public/strata/there-is-no-white/index.html | I re-read f5e5f84f34:public/strata/there-is-no-white/index.html:422-438. The coarse gate is "if(bestB===null || bestD>0.01) return {ok:false};" and the final return sets ok:true without testing its residual. Running these extracted arithmetic helpers gives 569 nm -> B=399.5000026 nm, residual 0.00851168405, and 494 nm -> residual 0.00147424175, both accepted. The 485 nm positive control lands within 6.4e-10. Thus the machine does not always find the exact match asserted at line 255. Require a verified spectral intersection and an explicit final tolerance before reporting a complement, and derive the displayed gap consistently. Correct the false success claim and add a dated Corrected line. Fixer reread and disposition: Replaced coarse near-match acceptance with a verified intersection of the interpolated spectral locus and a final residual <=1e-9 in xy, in the live and offline solvers. Both now reject 494 nm and 569 nm; their 0.5 nm scan gives a gap of 494 to 569.5 nm. The dated solver correction is mirrored in the markdown.
- CONFIRMED | X12: “but everywhere else, two is enough.” | git show f5e5f84f34:public/strata/there-is-no-white/index.html | I re-read the full paragraph at f5e5f84f34:public/strata/there-is-no-white/index.html:257: it explicitly says "roughly 495 to 568 nm" before describing the green gap. I independently intersected opposite rays through E with all 471 rows of the official 360-830 nm CIE table, https://files.cie.co.at/Publications-datasets/CIE_xyz_1931_2deg.csv (MD5 17cca777db64b17170f06f67ce9d3ab7). In the boundary region the integer wavelengths without partners are 494-569 nm; 485 nm has the expected partner near 586.6 nm. This supports the approximate interval, not an exact 568 nm cutoff. The final phrase refers to wavelengths outside that green gap, so the 569 nm example does not refute it. Reject this separate prose finding. The real false-success bug and its misleading gap output remain covered by X11.
- WRONG | X13: “collapsing to one identical white” | git show f5e5f84f34:public/strata/there-is-no-white/index.html; https://cie.co.at/eilvterm/17-23-008 | I re-read f5e5f84f34:public/strata/there-is-no-white/index.html:393-395 and 554-580. The full-spectrum branch assigns "result=E"; the screen branch assigns "result=D65"; the two-line branch targets E. Evaluating those pinned helpers gives E=(0.3330989865, 0.3333710786), D65=(0.31272, 0.32903), separated by 0.0208362198 in xy. I also read https://cie.co.at/eilvterm/17-23-008, definition: metamers have "the same tristimulus values". The sentence at f5e5f84f34:public/strata/there-is-no-white/index.html:261 explicitly calls these demonstrated recipes metamers. Different xy values already rule out matching XYZ; luminance is not normalized either. Correct the actual matching or the claim, including the identical-white descriptions, under the same dated Corrected line as X10, X14 and X28. Fixer reread and disposition: Kept the demonstrated reference whites and named them accurately: full spectrum and two-line recipes target E, screen primaries target D65. Metamerism is conditional on equal XYZ, including luminance. The offline check now compares unrounded XYZ at Y=1; the dated recipe correction is mirrored in the markdown.
- WRONG | X14: “are, to you, the same colour and a different universe of light.” | git show f5e5f84f34:public/strata/there-is-no-white/index.html; https://cie.co.at/eilvterm/17-23-008 | I re-read f5e5f84f34:public/strata/there-is-no-white/index.html:393-395 and 554-580. The full-spectrum branch assigns "result=E"; the screen branch assigns "result=D65"; the two-line branch targets E. Evaluating those pinned helpers gives E=(0.3330989865, 0.3333710786), D65=(0.31272, 0.32903), separated by 0.0208362198 in xy. I also read https://cie.co.at/eilvterm/17-23-008, definition: metamers have "the same tristimulus values". The next sentence at f5e5f84f34:public/strata/there-is-no-white/index.html:261 applies the identical-colour claim to the midday sky, this screen and two laser lines. The demonstrated screen and laser recipes already differ, and the page specifies no matched sky spectrum. Explain a possible match with specified stimuli, or construct the claimed match; do not assert that these actual examples are identical. Include this in the same dated Corrected line as X10, X13 and X28. Fixer reread and disposition: Kept the demonstrated reference whites and named them accurately: full spectrum and two-line recipes target E, screen primaries target D65. Metamerism is conditional on equal XYZ, including luminance. The offline check now compares unrounded XYZ at Y=1; the dated recipe correction is mirrored in the markdown.
- WRONG | X15: “you call it white every time” | https://cie.co.at/eilv/140 | I re-read the current CIE chromatic-adaptation definition at https://cie.co.at/eilvterm/17-22-013: "approximate compensation". Its Note 2 identifies the old term number 17-140 used by the source link. I also read https://cie.co.at/eilvterm/17-23-009, Note 1, which excludes highly chromatic sources from the usual neutral-diffuser assumption. The real-eye claim at f5e5f84f34:public/strata/there-is-no-white/index.html:264 says every time, not just within an ideal transform. Qualify the universal claim about actual observers and keep the exact simulated mapping identified as a model. Add a dated Corrected line. Fixer reread and disposition: Qualified the real-observer claim as approximate chromatic adaptation and explicitly identified the unchanging neutral paper as an ideal Bradford-model result, including the catalogue copy. The dated adaptation correction is mirrored in the markdown.
- MINOR | X16: “the way you actually see” | https://cie.co.at/eilv/140 | I read the CIE definition at https://cie.co.at/eilvterm/17-22-013, which describes approximate chromatic compensation, and f5e5f84f34:public/strata/there-is-no-white/index.html:291 and 318, which discuss the simulated paper and ideal transform. Comparing that ideal complete compensation with the way the reader actually sees is a MINOR qualification issue. Tier C: list. This is a different sentence from the universal real-surface claim being fixed for X15.
- MINOR | X17: “"White" is the brightest thing your visual system has decided is neutral.” | https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/ | I read the abstract of https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/: the evidence concerns anchoring lightness in simple surface displays and a near-white appearance, not a universal definition of all whites. At f5e5f84f34:public/strata/there-is-no-white/index.html:305 the brightest-neutral wording is a useful but overgeneralized heuristic. Retain MINOR and list under tier C; no WRONG finding requires editing this sentence.
- MINOR | X18: “The same checks run offline in research/there-is-no-white/verify.mjs (15/15).” | git show f5e5f84f34:public/strata/there-is-no-white/index.html; git show f5e5f84f34:research/there-is-no-white/verify.mjs | I read runCheck at f5e5f84f34:public/strata/there-is-no-white/index.html:705-724 and the assertions at f5e5f84f34:research/there-is-no-white/verify.mjs:139-215. The browser has eight assertions; the standalone program has fifteen, including extra identity, rim and metamer assertions and different solver precision. The advertised 15/15 result need not be changed into a failure. The issue is the overstatement that the check sets are the same. Under the supplied rule for claims about the page's own checking this is MINOR. Tier C: list; no WRONG fix shares the sentence at page line 308.
- MINOR | X19: “The geometry is exact:” | git show f5e5f84f34:public/strata/there-is-no-white/index.html; git show f5e5f84f34:research/there-is-no-white/verify.mjs | I re-read f5e5f84f34:public/strata/there-is-no-white/index.html:318 and 422-438, and f5e5f84f34:research/there-is-no-white/verify.mjs:181-190. The standalone metamer assertion accepts "dist(cTwo,cFlat)<0.02" after rounding the complementary wavelength to a 5 nm bin, while the browser uses a different solver and leaves final acceptance residuals unchecked. The sentence promises recomputation to the digit offline and live. Relevel WRONG to MINOR under the user's explicit rule that overclaims about the page's own checking are MINOR: this item overstates the precision and equivalence of that checking. This does not downgrade the actual false metamer equality or false successful complements, which remain WRONG in X10/X11/X13/X14/X28. Tier C: list; those WRONG assertions are in other sentences.
- MINOR | X20: “"White is not a wavelength" is unconditional” | https://cie.co.at/eilvterm/17-23-009 | I read https://cie.co.at/eilvterm/17-23-009: achromatic appearance is defined under the prevailing adaptation conditions. The page's reference-white locus calculation does not establish its unconditional scope wording at f5e5f84f34:public/strata/there-is-no-white/index.html:318. Keep this MINOR scope qualification and list under tier C. The quoted phrase can remain intact; any later clarification should qualify its scope outside the quotation. This is a separate sentence from the checking-precision claim.
- WRONG | X21: “no light is white.” | https://spaceplace.nasa.gov/review/magic-windows/vis.html | I re-read NASA, The Magic Visible Light Window, https://spaceplace.nasa.gov/review/magic-windows/vis.html, final paragraph: "White light goes in" when describing prism separation. The source explicitly allows white light made of multiple wavelengths. The conclusion at f5e5f84f34:public/strata/there-is-no-white/index.html:322 infers a denial of white light from the absence of a single white wavelength. Those are different propositions. Restrict the assertion to single-wavelength stimuli for the stated observer/reference whites, and include this copy in the dated Corrected line for the white-light denial. Fixer reread and disposition: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- MINOR | X22: “CIE standard illuminants [...] D65, D50, A, and equal-energy E” | https://cie.co.at/eilvterm/17-23-021; https://cie.co.at/eilvterm/17-23-023 | I read https://cie.co.at/eilvterm/17-23-021 and 17-23-023, and checked the newer official standard at https://www.cie.co.at/publications/colorimetry-part-2-cie-standard-illuminants-0. ISO/CIE 11664-2:2022 explicitly defines "D50 as a new standard illuminant". The checker's implication that D50 is not standard is outdated and must not be adopted. E remains an equi-energy spectrum treated as an illuminant, so the collective heading at f5e5f84f34:public/strata/there-is-no-white/index.html:327 still has a MINOR terminology issue. List only that grouping issue under tier C; preserve the standard status of A, D65 and D50.
- MINOR | X23: “The page recomputes the same from the same data.” | git show f5e5f84f34:public/strata/there-is-no-white/index.html; git show f5e5f84f34:research/there-is-no-white/verify.mjs | I read f5e5f84f34:public/strata/there-is-no-white/index.html:331 and 705-724 and f5e5f84f34:research/there-is-no-white/verify.mjs:139-215. They use the same CMF table and related calculations, but the eight live assertions do not reproduce all fifteen offline assertions, and the offline metamer check additionally bins the complementary wavelength. Treat the broad same-checking implication as MINOR, not as a failed numerical check. Tier C: list, sharing the label with X18 and X30; no WRONG fix shares this sentence.
- WRONG | X24: “No tracking, no database.” | git show f5e5f84f34:worker.js; https://artwaste.land/strata/there-is-no-white/ | I re-read f5e5f84f34:worker.js:3104-3108 and 3145: HitCounter uses ctx.storage.sql and "CREATE TABLE IF NOT EXISTS pulseday". Lines 2435-2437 generate a POST to /api/pulse. I also read 9560d8e7f963623965457330cce236c2d3fbf1e2:worker.js:708, at the layer's original 2026-06-29 publication commit: "CREATE TABLE IF NOT EXISTS daily" already existed. Thus no database was false when published, not merely overtaken since. Correct the database claim at page line 340 and add a dated Corrected line. Describe aggregate counting accurately without inferring personal identification or cross-site tracking from these tables. Fixer reread and disposition: Replaced no-database wording with aggregate page-view and browser-pulse counting in a database. Read worker.js at f5e5f84f34 and at publication commit 9560d8e7f963623965457330cce236c2d3fbf1e2; SQLite page counters already existed then. The dated database correction is mirrored in the markdown; no claim of personal identification or cross-site tracking was inferred.
- CONFIRMED | X25: “white is no color at all (pigment).” | https://cie.co.at/eilvterm/17-22-049 | I re-read f5e5f84f34:public/strata/there-is-no-white/index.html:7 and https://www.adobe.com/creativecloud/design/discover/is-black-a-color.html, the section on whether black and white are colours. Adobe actually says "black and white are not colors, they’re shades". The metadata begins "The internet answers this two opposite ways at once" and immediately follows the reported answers with "Both miss it." It does not include the parenthetical true-of-paint endorsement found in X1/X2/X26. A colourimetric objection to the reported view does not make the report that the internet gives that answer false. Reject this finding on the metadata; retain the affirmative paint corrections separately.
- WRONG | X26: “white is no colour at all (true of paint).” | https://cie.co.at/eilvterm/17-22-049 | I re-read https://cie.co.at/eilvterm/17-22-049, definition and Note 1: "perceived colour devoid of hue" includes white. I also read https://www.adobe.com/creativecloud/design/discover/is-black-a-color.html, the section on whether black and white are colours: it reports a shade convention, not a physical rule that paint has no colour. I also re-read f5e5f84f34:src/content/strata/there-is-no-white.md:4: the catalogue dek explicitly adds "true of paint". This is the same affirmative colour-versus-hue error as X1/X2, not merely the report of conflicting internet answers in X25. Correct the catalogue and other affirmative copies consistently and use the same dated Corrected line. Fixer reread and disposition: Corrected the affirmative paint claim in the page and catalogue, the OG endorsement, and the nearby card via this layer's own relates note. White paint is an achromatic colour, without hue. The dated paint correction is mirrored in the markdown.
- WRONG | X27: “no beam of light is white” | https://spaceplace.nasa.gov/review/magic-windows/vis.html | I re-read NASA, The Magic Visible Light Window, https://spaceplace.nasa.gov/review/magic-windows/vis.html, final paragraph: "White light goes in" when describing prism separation. The source explicitly allows white light made of multiple wavelengths. I also re-read f5e5f84f34:src/content/strata/there-is-no-white.md:4. Its denial of white beams is broader than its correct statement about a single wavelength and is not supported by the nearest-locus calculation. Correct this catalogue copy together with the body and generated check output, using the same dated Corrected line. Fixer reread and disposition: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- WRONG | X28: “are one white and three different universes of light (metamerism).” | git show f5e5f84f34:public/strata/there-is-no-white/index.html; https://cie.co.at/eilvterm/17-23-008 | I re-read f5e5f84f34:public/strata/there-is-no-white/index.html:393-395 and 554-580. The full-spectrum branch assigns "result=E"; the screen branch assigns "result=D65"; the two-line branch targets E. Evaluating those pinned helpers gives E=(0.3330989865, 0.3333710786), D65=(0.31272, 0.32903), separated by 0.0208362198 in xy. I also read https://cie.co.at/eilvterm/17-23-008, definition: metamers have "the same tristimulus values". I also re-read f5e5f84f34:src/content/strata/there-is-no-white.md:4, which promises one white for the sky, screen and laser examples. The demonstrated screen and two-line recipes differ and no matched sky stimulus is specified. Correct this catalogue copy consistently with X10/X13/X14, using the same dated Corrected line. Fixer reread and disposition: Kept the demonstrated reference whites and named them accurately: full spectrum and two-line recipes target E, screen primaries target D65. Metamerism is conditional on equal XYZ, including luminance. The offline check now compares unrounded XYZ at Y=1; the dated recipe correction is mirrored in the markdown.
- MINOR | X29: “it is the brightest neutral your visual system has currently decided on” | https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/ | I read the abstract of https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/ and the relevant sentence in f5e5f84f34:src/content/strata/there-is-no-white.md:4. The anchoring account is supported for the tested surface displays, but it is a loose universal definition of white. Retain MINOR and list under tier C, sharing the label with X3/X17. This sentence is separate from the paint, beam and metamer sentences being corrected in the same dek.
- MINOR | X30: “the same checks (15/15) run offline.” | git show f5e5f84f34:public/strata/there-is-no-white/index.html; git show f5e5f84f34:research/there-is-no-white/verify.mjs | I read the last sentence of the dek at f5e5f84f34:src/content/strata/there-is-no-white.md:4 and compared f5e5f84f34:public/strata/there-is-no-white/index.html:705-724 with f5e5f84f34:research/there-is-no-white/verify.mjs:139-215. Eight live assertions and fifteen related offline assertions are different sets. The 15/15 offline report is not disproved; the same-checks wording overstates equivalence. This is MINOR under the explicit checking-overclaim rule. Tier C: list, sharing the error with X18/X23; the WRONG claims occur in other dek sentences.
- WRONG | X31: “no light is white” | https://spaceplace.nasa.gov/review/magic-windows/vis.html | I re-read NASA, The Magic Visible Light Window, https://spaceplace.nasa.gov/review/magic-windows/vis.html, final paragraph: "White light goes in" when describing prism separation. The source explicitly allows white light made of multiple wavelengths. I also re-read the generated transcript template at f5e5f84f34:public/strata/there-is-no-white/index.html:710. Its condition nearD>0.15 checks a monochromatic-locus distance, but its message concludes "no light is white". Preserve the valid numerical calculation and correct that unsupported inference everywhere it is printed, including the standalone verifier copy at f5e5f84f34:research/there-is-no-white/verify.mjs:147. Add the shared dated Corrected line; this is the same generated sentence as X35. Fixer reread and disposition: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- WRONG | X32: “The screen never emits white; it emits three lights.” | https://registry.color.org/rgb-registry/srgb | I re-read ICC's sRGB encoding characteristics at https://registry.color.org/rgb-registry/srgb: "White point chromaticity: x = 0.3127, y = 0.3290 (D65)". I also read f5e5f84f34:public/strata/there-is-no-white/index.html:569-572 and 713-714: the page assigns the primaries' combined result to D65 and checks their summed XYZ against that white. The assertion that the screen never emits white therefore mistakes a mixture of lights for an inability to emit white light. Correct the generated sentence to describe the combined emission accurately and include it in the dated Corrected line for white-light denial. X32 and X33 identify the same sentence, not two errors. Fixer reread and disposition: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- WRONG | X33: “The screen never emits white; it emits three lights.” | https://registry.color.org/rgb-registry/srgb | I re-read ICC's sRGB encoding characteristics at https://registry.color.org/rgb-registry/srgb: "White point chromaticity: x = 0.3127, y = 0.3290 (D65)". I also read f5e5f84f34:public/strata/there-is-no-white/index.html:569-572 and 713-714: the page assigns the primaries' combined result to D65 and checks their summed XYZ against that white. The assertion that the screen never emits white therefore mistakes a mixture of lights for an inability to emit white light. Correct the generated sentence to describe the combined emission accurately and include it in the dated Corrected line for white-light denial. X32 and X33 identify the same sentence, not two errors. Fixer reread and disposition: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- MINOR | X34: “make the white you are reading this on” | https://registry.color.org/rgb-registry/srgb | I read https://registry.color.org/rgb-registry/srgb, which specifies an sRGB reference white, and f5e5f84f34:public/strata/there-is-no-white/index.html:572. It cannot establish the actual calibration, viewing settings or white point of every reader's display. This is a MINOR qualification of the reference example. Tier C: list. The full-drive assertion is one sentence; the screen-never-emits-white assertion fixed for X32/X33 is the next sentence. Sharing one JavaScript string does not trigger the same-sentence exception.
- WRONG | X35: “no light is white” | https://spaceplace.nasa.gov/review/magic-windows/vis.html | I re-read NASA, The Magic Visible Light Window, https://spaceplace.nasa.gov/review/magic-windows/vis.html, final paragraph: "White light goes in" when describing prism separation. The source explicitly allows white light made of multiple wavelengths. I also re-read f5e5f84f34:public/strata/there-is-no-white/index.html:710: the test is nearD>0.15 but the printed conclusion is "no light is white". The test establishes separation of single wavelengths from the reference white, not nonexistence of white mixed light. Correct this generated sentence and its standalone copy at f5e5f84f34:research/there-is-no-white/verify.mjs:147 under the same dated Corrected line as X31 and the other white-light denials. Fixer reread and disposition: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- MINOR | X36: “three lines (R+G+B) land at D65” | git show f5e5f84f34:public/strata/there-is-no-white/index.html; https://registry.color.org/rgb-registry/srgb | I read f5e5f84f34:public/strata/there-is-no-white/index.html:570, which explicitly says "not pure wavelengths", and lines 713-714, which sum sRGB primary tristimulus vectors while labeling them lines. ICC https://registry.color.org/rgb-registry/srgb supplies primary chromaticities, not monochromatic spectral lines. This is MINOR terminology inconsistent with the page's own qualification. Tier C: list. It is a different generated check sentence from the no-light-is-white error at line 710.
- WRONG | X37: “both are right once you name the frame” | https://cie.co.at/eilvterm/17-22-049 | I re-read https://cie.co.at/eilvterm/17-22-049, definition and Note 1: "perceived colour devoid of hue" includes white. I also read https://www.adobe.com/creativecloud/design/discover/is-black-a-color.html, the section on whether black and white are colours: it reports a shade convention, not a physical rule that paint has no colour. I also re-read f5e5f84f34:public/strata/there-is-no-white/index.html:757. Unlike the metadata's report in X25, this nearby-card sentence expressly endorses both answers as right under the light-versus-pigment framing. That framing does not establish the paint claim as a scientific fact; an artistic hue/shade convention needs to be named as a convention. Correct the affirmative endorsement consistently with X1/X2/X26 and use the same dated Corrected line. This nearby-card text is not one of the shared placard or plain-button classes. Fixer reread and disposition: Corrected the affirmative paint claim in the page and catalogue, the OG endorsement, and the nearby card via this layer's own relates note. White paint is an achromatic colour, without hue. The dated paint correction is mirrored in the markdown.
- STALE | X38: “"dateModified":"2026-06-29"” | git log f5e5f84f34 -- public/strata/there-is-no-white/index.html; git show 2e60bcb5dd1:public/strata/there-is-no-white/index.html | I read the JSON-LD at f5e5f84f34:public/strata/there-is-no-white/index.html:167 and the layer's git history. Publication commit 9560d8e7f963623965457330cce236c2d3fbf1e2 is dated 2026-06-29; commit 2e60bcb5dd1 changes the page's reader-facing OG metadata on 2026-10-01 while dateModified still says 2026-06-29. Retain STALE. Apply the explicit JSON-LD exception: fix dateModified to the actual modification date when the correction lands, with no Updated or Corrected line for this metadata-only item, and preserve datePublished. This is the layer's only STALE item, so there are no non-exempt STALE items to group into an Updated line; any such items would share one Updated line. dateModified is now 2026-10-14; datePublished remains 2026-06-29. The Corrected lines date the revision under the explicit JSON-LD exception.
- MINOR | X39: (added) “600 + 489” | git show f5e5f84f34:public/strata/there-is-no-white/index.html | I read the 600 + 489 button at f5e5f84f34:public/strata/there-is-no-white/index.html:247, its 600 nm handler at lines 618-622, and complementOf at lines 422-438. Running the extracted pinned arithmetic gives B=490.00485583442924 nm, which the displayed readout rounds to 490. The static label is a small numerical precision/preset slip. Retain MINOR and list under tier C; no WRONG fix shares this button label.
- CONFIRMED | “There is no white wavelength.” | https://spaceplace.nasa.gov/review/magic-windows/vis.html | Checker reread retained from N2-codex.md: C003 `white-no-wavelength`. NASA says "A prism separates white light into its different wavelengths." The standard CIE spectrum contains no unique monochromatic white; the pinned 65-row computation independently confirms its separation from E.
- CONFIRMED | “You don't need all the colours to make it.” | https://cie.co.at/eilvterm/17-23-063 | Checker reread retained from N2-codex.md: C004 `white-not-all`. The CIE complementary-wavelength definition and NASA report support suitable two-wavelength white mixtures. Manually read NASA PDF page 29: "by two monochromatic sources of proper wavelengths". This is an existence claim, not a claim that every arbitrary pair works.
- CONFIRMED | “White is a property of you.” | https://cie.co.at/eilvterm/17-22-040 | Checker reread retained from N2-codex.md: C009 `white-property-observer`. CIE defines perceptual colour as a "characteristic of visual perception" and explicitly includes adaptation and surround. Read as the page discussion of perception, not a denial that light has measurable spectral properties.
- CONFIRMED | “the same CIE 1931 colour-matching data that underpins every screen you own” | https://cie.co.at/datatable/cie-1931-colour-matching-functions-2-degree-observer | Checker reread retained from N2-codex.md: C010 `cie-every-screen`. The CIE source identifies the 1931 2 degree colour-matching functions. Independent comparison of all 65 embedded rows with its linked official CSV gives a maximum coefficient difference of 1.0e-7; the ICC sRGB specification uses the corresponding chromaticity system. The record excerpt alone is metadata, so the table and ICC source were also checked.
- CONFIRMED | “Split sunlight with a prism and the colours fall out in order [...] red, orange, yellow, green, blue, violet [...] each a single wavelength of light.” | https://science.nasa.gov/ems/09_visiblelight/; git show f5e5f84f34:public/strata/there-is-no-white/index.html | Checker reread retained from N2-codex.md: C012 `prism-colour-sequence`. NASA Visible Light explicitly gives violet around 380 nm, red around 700 nm, and prism separation in wavelength order. The page slider and endpoint labels use those bounds.
- CONFIRMED | “the spectral locus, the curve of everything a single wavelength can be.” | https://cie.co.at/datatable/cie-1931-colour-matching-functions-2-degree-observer | Checker reread retained from N2-codex.md: C013 `spectral-locus`. The CIE source identifies the 1931 2 degree colour-matching functions. Independent comparison of all 65 embedded rows with its linked official CSV gives a maximum coefficient difference of 1.0e-7; the ICC sRGB specification uses the corresponding chromaticity system. The record excerpt alone is metadata, so the table and ICC source were also checked.
- CONFIRMED | “The whites live in the middle [...] no single wavelength touches them.” | https://cie.co.at/eilvterm/17-23-009 | Checker reread retained from N2-codex.md: C019 `white-middle`. The reproduced CIE locus places E, D65 and A in the interior. Nearest E distance is 0.235; the result is for the page standard observer and specified white points, not every adaptation state.
- CONFIRMED | “A single wavelength is always a saturated colour on the rim [...] never the neutral centre.” | https://spaceplace.nasa.gov/review/magic-windows/vis.html | Checker reread retained from N2-codex.md: C020 `single-wavelength-always`. The reproduced CIE locus places E, D65 and A in the interior. Nearest E distance is 0.235; the result is for the page standard observer and specified white points, not every adaptation state.
- CONFIRMED | “White behaves exactly like magenta in this one respect: it is a colour with no wavelength, a colour the eye assembles rather than receives.” | https://cie.co.at/eilvterm/17-23-057 | Checker reread retained from N2-codex.md: C021 `white-magenta-no-wavelength`. The CIE purple-stimulus and purple-boundary entries distinguish mixtures across the short- and long-wave ends from spectral stimuli. The 530 nm green example has no spectral complement to E. The description of the eye closing the spectrum is explanatory language, not a separate physiological mechanism demonstrated by that definition.
- CONFIRMED | “magenta is the hue the eye invents to close the open spectrum; white is the neutral the eye builds at its centre.” | https://cie.co.at/eilvterm/17-23-058 | Checker reread retained from N2-codex.md: C022 `magenta-closes-spectrum`. The CIE purple-stimulus and purple-boundary entries distinguish mixtures across the short- and long-wave ends from spectral stimuli. The 530 nm green example has no spectral complement to E. The description of the eye closing the spectrum is explanatory language, not a separate physiological mechanism demonstrated by that definition.
- CONFIRMED | “But magenta at least needs the two ends of the spectrum to make it.” | https://cie.co.at/eilvterm/17-23-058 | Checker reread retained from N2-codex.md: C023 `magenta-needs-ends`. The CIE purple-stimulus and purple-boundary entries distinguish mixtures across the short- and long-wave ends from spectral stimuli. The 530 nm green example has no spectral complement to E. The description of the eye closing the spectrum is explanatory language, not a separate physiological mechanism demonstrated by that definition.
- CONFIRMED | “White is not a roll-call of every wavelength; it's a balance across your three cone types.” | https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=EXT_ID%3A9214503%20AND%20SRC%3AMED&resultType=core&format=json | Checker reread retained from N2-codex.md: C025 `three-cone-balance`. Engel, Zhang and Wandell abstract: "three classes of cones, the L, M and S cones"; the full abstract also describes neural processing of their signals. This supports the normal trichromatic-observer account.
- CONFIRMED | “Anything that lands the balance at the centre is white, and there are infinitely many ways to do it.” | https://cie.co.at/eilvterm/17-23-008 | Checker reread retained from N2-codex.md: C026 `infinite-recipes`. CIE requires spectrally different stimuli with matching tristimulus values. Infinitely many recipes follow from linearity: any continuously varying positive mixture of two distinct, luminance-normalised matching spectra remains a match. The code uses linear XYZ addition.
- CONFIRMED | “Two single wavelengths, mixed in the right ratio, land dead on white.” | https://ntrs.nasa.gov/api/citations/19700024467/downloads/19700024467.pdf | Checker reread retained from N2-codex.md: C029 `two-wavelengths-white`. The CIE complementary-wavelength definition and NASA report support suitable two-wavelength white mixtures. Manually read NASA PDF page 29: "by two monochromatic sources of proper wavelengths". This is an existence claim, not a claim that every arbitrary pair works.
- CONFIRMED | “Two lines of light, nothing between them, and your eye reports white.” | https://ntrs.nasa.gov/api/citations/19700024467/downloads/19700024467.pdf | Checker reread retained from N2-codex.md: C031 `two-lines-eye-white`. The CIE complementary-wavelength definition and NASA report support suitable two-wavelength white mixtures. Manually read NASA PDF page 29: "by two monochromatic sources of proper wavelengths". This is an existence claim, not a claim that every arbitrary pair works.
- CONFIRMED | “The schoolbook picture [...] white as a dense sum of all the colours [...] is one recipe out of infinitely many, and far from the cheapest.” | https://cie.co.at/eilvterm/17-23-008 | Checker reread retained from N2-codex.md: C032 `schoolbook-infinite-recipes`. CIE requires spectrally different stimuli with matching tristimulus values. Infinitely many recipes follow from linearity: any continuously varying positive mixture of two distinct, luminance-normalised matching spectra remains a match. The code uses linear XYZ addition.
- CONFIRMED | “A pure green has no single-wavelength partner that makes white; its complement is a magenta, which (as the sibling page shows) is itself not a wavelength.” | https://cie.co.at/eilvterm/17-23-057 | Checker reread retained from N2-codex.md: C034 `pure-green-magenta-complement`. The CIE purple-stimulus and purple-boundary entries distinguish mixtures across the short- and long-wave ends from spectral stimuli. The 530 nm green example has no spectral complement to E. The description of the eye closing the spectrum is explanatory language, not a separate physiological mechanism demonstrated by that definition.
- CONFIRMED | “White is not a property of the light at all [...] it's a property of your adaptation.” | https://cie.co.at/eilv/140 | Checker reread retained from N2-codex.md: C038 `adaptation-property`. CIE chromatic adaptation explicitly describes "approximate compensation" for changes of illuminant. Combined with the verified complementary-pair example, this supports the general non-fixed-white account; it does not establish perfect constancy.
- CONFIRMED | “Switch adaptation off and you see the raw retinal truth: the paper is the colour of the fire.” | https://cie.co.at/eilv/140; git show f5e5f84f34:public/strata/there-is-no-white/index.html | Checker reread retained from N2-codex.md: C047 `raw-retinal-fire`. For the stated ideal neutral surface, the unadapted reflected light has the illuminant chromaticity. The pinned simulation and its adaptation toggle reproduce that relationship. Read raw retinal truth as the illustrated unadapted stimulus, not a physiological measurement of a reader retina.
- CONFIRMED | “This is white balance [...] the thing your phone camera fakes in software and your eye does for free, every waking second, without telling you.” | https://arxiv.org/abs/1507.00410 | Checker reread retained from N2-codex.md: C048 `phone-white-balance`. Barron describes inferring the illuminating colour "so that the illumination color can be removed" and discusses white-balanced images. This supports the software white-balance analogy; it is not evidence that the eye literally executes the same algorithm.
- CONFIRMED | “Yet the one on the dark field looks white, and the one on the light field looks grey.” | https://cie.co.at/eilvterm/17-22-040 | Checker reread retained from N2-codex.md: C050 `grey-looks-white`. CIE explicitly includes stimulus surround in perceived colour. The two target patches are identical RGB values; the unchanged-photon statement is about the target, while the surround changes. The claimed direction is the standard lightness-contrast demonstration, not a guarantee of identical reports by all viewers.
- CONFIRMED | “Change what surrounds it, and the decision changes [...] without one photon changing.” | https://cie.co.at/eilvterm/17-22-040 | Checker reread retained from N2-codex.md: C053 `surround-no-photon`. CIE explicitly includes stimulus surround in perceived colour. The two target patches are identical RGB values; the unchanged-photon statement is about the target, while the surround changes. The claimed direction is the standard lightness-contrast demonstration, not a guarantee of identical reports by all viewers.
- CONFIRMED | “That is the whole of it: white is not in the light, and not on the page. It's a verdict your eye is always quietly re-issuing.” | https://cie.co.at/eilvterm/17-22-040 | Checker reread retained from N2-codex.md: C054 `white-not-in-light`. CIE defines perceptual colour as a "characteristic of visual perception" and explicitly includes adaptation and surround. Read as the page discussion of perception, not a denial that light has measurable spectral properties.
- CONFIRMED | “What's conventional is the machinery itself [...] the CIE 1931 2° observer is a standard average human, not a law of physics (real observers vary; the 1964 10° observer differs);” | https://cie.co.at/publications/proceedings-cie-symposium-advanced-colorimetry-8-10-june-1993-vienna-austria | Checker reread retained from N2-codex.md: C058 `standard-average-observer`. CIE symposium proceedings explicitly discuss "large differences between actual matches performed by one observer" and the "average Standard Observer", and separately discuss the CIE 1964 observer.
- CONFIRMED | “D65, illuminant A, and the Bradford transform are agreed standards, not nature.” | https://www.color.org/ICC1V42old.pdf | Checker reread retained from N2-codex.md: C059 `standards-not-nature`. Manually read ICC.1:2004-04 Annex E, PDF pages 97-98. It recommends linear Bradford and gives the diagonal white-ratio scaling between the Bradford matrix and its inverse. The page coefficients and the three-stage procedure match the displayed equations.
- CONFIRMED | “a perfectly neutral surface reflects the lamp's spectrum” | https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/ | Checker reread retained from N2-codex.md: C060 `neutral-reflects-lamp`. For the explicitly ideal neutral surface, spectrally constant reflectance scales the illuminant spectrum without changing chromaticity. The PMC 90% excerpt establishes typical reflectance magnitude, not spectral flatness by itself; the ideal-neutral premise is essential.
- CONFIRMED | “It is not all the colours: two will do, and the recipe is never unique.” | https://cie.co.at/eilvterm/17-23-063 | Checker reread retained from N2-codex.md: C065 `conclusion-two-infinite`. The CIE complementary-wavelength definition and NASA report support suitable two-wavelength white mixtures. Manually read NASA PDF page 29: "by two monochromatic sources of proper wavelengths". This is an existence claim, not a claim that every arbitrary pair works.
- CONFIRMED | “And it is not a fixed colour at all: it is the neutral your visual system is forever re-deriving from whatever light it finds itself in.” | https://cie.co.at/eilv/140 | Checker reread retained from N2-codex.md: C066 `conclusion-neutral-rederived`. CIE chromatic adaptation explicitly describes "approximate compensation" for changes of illuminant. Combined with the verified complementary-pair example, this supports the general non-fixed-white account; it does not establish perfect constancy.
- CONFIRMED | “CIE 1931 2° standard-observer colour-matching functions (x̄, ȳ, z̄)” | https://cie.co.at/datatable/cie-1931-colour-matching-functions-2-degree-observer | Checker reread retained from N2-codex.md: C067 `cite-cie1931`. The CIE source identifies the 1931 2 degree colour-matching functions. Independent comparison of all 65 embedded rows with its linked official CSV gives a maximum coefficient difference of 1.0e-7; the ICC sRGB specification uses the corresponding chromaticity system. The record excerpt alone is metadata, so the table and ICC source were also checked.
- CONFIRMED | “Digitised tables from the Colour & Vision Research Laboratory, UCL [...] the standard machine-readable form of the CIE 1931 standard observer.” | https://web.archive.org/web/20250000000000id_/http://cvrl.ioo.ucl.ac.uk/cmfs.htm | Checker reread retained from N2-codex.md: C069 `cvrl-provenance`. Opened the archived CVRL page: "CIE 1931 2-deg, XYZ CMFs" and downloadable CSV formats are present. Its UCL host and laboratory provenance match the citation. The official CIE dataset independently supports the copied coefficient table.
- CONFIRMED | “cvrl.ioo.ucl.ac.uk/cmfs.htm” | https://web.archive.org/web/20250000000000id_/http://cvrl.ioo.ucl.ac.uk/cmfs.htm | Checker reread retained from N2-codex.md: C070 `cite-cvrl`. Opened the archived CVRL page: "CIE 1931 2-deg, XYZ CMFs" and downloadable CSV formats are present. Its UCL host and laboratory provenance match the citation. The official CIE dataset independently supports the copied coefficient table.
- CONFIRMED | “D65 (0.31272, 0.32903) and D50 (daylight phases), illuminant A (0.44758, 0.40745; tungsten, ~2856 K), and equal-energy E at (1/3, 1/3).” | https://raw.githubusercontent.com/colour-science/colour/develop/colour/colorimetry/datasets/illuminants/chromaticity_coordinates.py | Checker reread retained from N2-codex.md: C073 `reference-white-coordinates`. A, D65, E and the stated approximate temperature are supported by the CIE/ICC references at the shown precision in ordinary use. Source-record limitation: its Colour excerpt actually gives D65 [0.31270, 0.32900], not [0.31272, 0.32903]; it cannot certify the final two quoted digits. This tiny difference does not repair the much larger E-versus-D65 mismatch.
- CONFIRMED | “CIE 15:2004, Colorimetry.” | https://cie.co.at/publications/colorimetry-3rd-edition | Checker reread retained from N2-codex.md: C074 `cite-cie15`. CIE publication page explicitly gives "Colorimetry, 3rd edition" and "CIE 015:2004". The record accurately names this edition, without claiming it is the latest edition.
- CONFIRMED | “Bradford chromatic-adaptation transform (CAT)” | https://www.color.org/ICC1V42old.pdf | Checker reread retained from N2-codex.md: C075 `cite-bradford`. Manually read ICC.1:2004-04 Annex E, PDF pages 97-98. It recommends linear Bradford and gives the diagonal white-ratio scaling between the Bradford matrix and its inverse. The page coefficients and the three-stage procedure match the displayed equations.
- CONFIRMED | “The standard von-Kries-type transform used in Instrument III: XYZ → sharpened cone space, scale each channel by the white-point ratio, transform back.” | https://www.color.org/ICC1V42old.pdf | Checker reread retained from N2-codex.md: C076 `cat-three-stages`. Manually read ICC.1:2004-04 Annex E, PDF pages 97-98. It recommends linear Bradford and gives the diagonal white-ratio scaling between the Bradford matrix and its inverse. The page coefficients and the three-stage procedure match the displayed equations.
- CONFIRMED | “Matrix from S. Bianco & R. Schettini / Lam (1985), as standardised in ICC profile connection and widely tabulated.” | https://acorn.stanford.edu/psych221/projects/2010/JasonSu/Papers/Two%20New%20von%20Kries%20Based%20Chromatic%20Adaptation.pdf | Checker reread retained from N2-codex.md: C077 `cite-bianco-lam`. Opened the Bianco/Schettini PDF after the initial TLS certificate failure. Page 1 names S. Bianco and R. Schettini; Table 1 reproduces the page Bradford matrix; reference [3] names Lam and the 1985 Bradford thesis. The date modifies Lam, not the Bianco/Schettini paper.
- CONFIRMED | “sRGB / Rec. ITU-R BT.709 primaries” | https://registry.color.org/rgb-registry/srgb; https://registry.color.org/rgb-registry/files/sRGB.pdf; git show f5e5f84f34:public/strata/there-is-no-white/index.html | Checker reread retained from N2-codex.md: C078 `cite-srgb709`. ICC sRGB registry gives R (0.64,0.33), G (0.30,0.60), B (0.15,0.06), D65, and IEC 61966-2-1:1999. The linked PDF page 1 gives the XYZ-to-linear-sRGB matrix; the page uses its familiar rounded coefficients. These are primary chromaticities, not pure spectral lines.
- CONFIRMED | “R (0.640, 0.330), G (0.300, 0.600), B (0.150, 0.060), white D65.” | https://registry.color.org/rgb-registry/srgb; https://registry.color.org/rgb-registry/files/sRGB.pdf; git show f5e5f84f34:public/strata/there-is-no-white/index.html | Checker reread retained from N2-codex.md: C079 `srgb-primaries`. ICC sRGB registry gives R (0.64,0.33), G (0.30,0.60), B (0.15,0.06), D65, and IEC 61966-2-1:1999. The linked PDF page 1 gives the XYZ-to-linear-sRGB matrix; the page uses its familiar rounded coefficients. These are primary chromaticities, not pure spectral lines.
- CONFIRMED | “IEC 61966-2-1:1999” | https://registry.color.org/rgb-registry/srgb; https://registry.color.org/rgb-registry/files/sRGB.pdf; git show f5e5f84f34:public/strata/there-is-no-white/index.html | Checker reread retained from N2-codex.md: C080 `cite-iec`. ICC sRGB registry gives R (0.64,0.33), G (0.30,0.60), B (0.15,0.06), D65, and IEC 61966-2-1:1999. The linked PDF page 1 gives the XYZ-to-linear-sRGB matrix; the page uses its familiar rounded coefficients. These are primary chromaticities, not pure spectral lines.
- CONFIRMED | “the linear-RGB↔XYZ matrices are the standard ones.” | https://registry.color.org/rgb-registry/srgb; https://registry.color.org/rgb-registry/files/sRGB.pdf; git show f5e5f84f34:public/strata/there-is-no-white/index.html | Checker reread retained from N2-codex.md: C081 `linear-rgb-standard`. ICC sRGB registry gives R (0.64,0.33), G (0.30,0.60), B (0.15,0.06), D65, and IEC 61966-2-1:1999. The linked PDF page 1 gives the XYZ-to-linear-sRGB matrix; the page uses its familiar rounded coefficients. These are primary chromaticities, not pure spectral lines.
- CONFIRMED | “The two-greys demo: a classic perceptual effect (a patch's apparent lightness depends on its surround).” | https://cie.co.at/eilvterm/17-22-040 | Checker reread retained from N2-codex.md: C082 `simultaneous-contrast`. CIE explicitly includes stimulus surround in perceived colour. The two target patches are identical RGB values; the unchanged-photon statement is about the target, while the surround changes. The claimed direction is the standard lightness-contrast demonstration, not a guarantee of identical reports by all viewers.
- CONFIRMED | “There is no white frequency.” | https://spaceplace.nasa.gov/review/magic-windows/vis.html | Checker reread retained from N2-codex.md: C093 `meta-frequency`. NASA says "A prism separates white light into its different wavelengths." The standard CIE spectrum contains no unique monochromatic white; the pinned 65-row computation independently confirms its separation from E.
- CONFIRMED | “It is not 'all the colors': two will do.” | https://cie.co.at/eilvterm/17-23-063 | Checker reread retained from N2-codex.md: C094 `meta-two`. The CIE complementary-wavelength definition and NASA report support suitable two-wavelength white mixtures. Manually read NASA PDF page 29: "by two monochromatic sources of proper wavelengths". This is an existence claim, not a claim that every arbitrary pair works.
- CONFIRMED | “It is your visual system's running estimate of neutral, reset by whatever light you are under.” | https://cie.co.at/eilv/140 | Checker reread retained from N2-codex.md: C095 `meta-estimate`. CIE chromatic adaptation explicitly describes "approximate compensation" for changes of illuminant. Combined with the verified complementary-pair example, this supports the general non-fixed-white account; it does not establish perfect constancy.
- CONFIRMED | “White has no wavelength, doesn't need all the colors (two will do), and isn't even fixed: it's your eye's running estimate of neutral, reset by the light you're under.” | https://cie.co.at/eilvterm/17-23-009 | Checker reread retained from N2-codex.md: C096 `og-description`. CIE chromatic adaptation explicitly describes "approximate compensation" for changes of illuminant. Combined with the verified complementary-pair example, this supports the general non-fixed-white account; it does not establish perfect constancy.
- CONFIRMED | “White has no wavelength, isn't 'all the colours' (two will do), and isn't even a fixed colour: it's the eye's running estimate of neutral.” | https://cie.co.at/eilvterm/17-23-009 | Checker reread retained from N2-codex.md: C097 `twitter-description`. CIE chromatic adaptation explicitly describes "approximate compensation" for changes of illuminant. Combined with the verified complementary-pair example, this supports the general non-fixed-white account; it does not establish perfect constancy.
- CONFIRMED | “White is not even 'all the colours': two complementary wavelengths, mixed in the right ratio, land dead on the white point, and the recipe is never unique.” | https://cie.co.at/eilvterm/17-23-063 | Checker reread retained from N2-codex.md: C104 `dek-4`. The CIE complementary-wavelength definition and NASA report support suitable two-wavelength white mixtures. Manually read NASA PDF page 29: "by two monochromatic sources of proper wavelengths". This is an existence claim, not a claim that every arbitrary pair works.
- CONFIRMED | “illuminant A (0.44758, 0.40745; tungsten, ~2856 K)” | https://cie.co.at/eilvterm/17-23-021 | Checker reread retained from N2-codex.md: C130 `illuminant-a-temperature`. CIE gives illuminant A as Planckian radiation at "about 2 855,5 K". Rounding this to approximately 2856 K is sound.
- CONFIRMED | “It is any balance that lands in the middle [...] and two lights can carry it there.” | https://ntrs.nasa.gov/api/citations/19700024467/downloads/19700024467.pdf | Checker reread retained from N2-codex.md: C133 `white-two-midpoint`. The CIE complementary-wavelength definition and NASA report support suitable two-wavelength white mixtures. Manually read NASA PDF page 29: "by two monochromatic sources of proper wavelengths". This is an existence claim, not a claim that every arbitrary pair works.
- CONFIRMED | “380 nm [...] violet 700 nm [...] red” | https://science.nasa.gov/ems/09_visiblelight/; git show f5e5f84f34:public/strata/there-is-no-white/index.html | Checker reread retained from N2-codex.md: C134 `spectral-ranges`. NASA Visible Light explicitly gives violet around 380 nm, red around 700 nm, and prism separation in wavelength order. The page slider and endpoint labels use those bounds.
- CONFIRMED | “three primaries (your screen [...] not pure wavelengths)” | https://registry.color.org/rgb-registry/srgb; https://registry.color.org/rgb-registry/files/sRGB.pdf; git show f5e5f84f34:public/strata/there-is-no-white/index.html | Checker reread retained from N2-codex.md: C136 `screen-not-pure`. ICC sRGB registry gives R (0.64,0.33), G (0.30,0.60), B (0.15,0.06), D65, and IEC 61966-2-1:1999. The linked PDF page 1 gives the XYZ-to-linear-sRGB matrix; the page uses its familiar rounded coefficients. These are primary chromaticities, not pure spectral lines.
- CONFIRMED | “The two halves of one fact: a colour is not in the light, it is in the eye.” | https://cie.co.at/eilvterm/17-22-040 | Checker reread retained from N2-codex.md: C147 `related-magenta-0`. CIE defines perceptual colour as a "characteristic of visual perception" and explicitly includes adaptation and surround. Read as the page discussion of perception, not a denial that light has measurable spectral properties.
- CONFIRMED | “This is its centre [...] the neutral the eye builds where no wavelength lives, and keeps re-deciding as the light changes.” | https://cie.co.at/eilvterm/17-22-040 | Checker reread retained from N2-codex.md: C149 `related-magenta-2`. CIE defines perceptual colour as a "characteristic of visual perception" and explicitly includes adaptation and surround. Read as the page discussion of perception, not a denial that light has measurable spectral properties.
- CONFIRMED | “Magenta has no wavelength because it is a chord across the spectrum's ends; white has no wavelength because it is the balance at the middle.” | https://cie.co.at/eilvterm/17-23-058 | Checker reread retained from N2-codex.md: C150 `related-magenta-3`. The CIE purple-stimulus and purple-boundary entries distinguish mixtures across the short- and long-wave ends from spectral stimuli. The 530 nm green example has no spectral complement to E. The description of the eye closing the spectrum is explanatory language, not a separate physiological mechanism demonstrated by that definition.
- CONFIRMED | “Read together they bracket the colour solid: the eye manufactures both its edge and its origin.” | https://cie.co.at/eilvterm/17-23-058 | Checker reread retained from N2-codex.md: C151 `related-magenta-4`. CIE defines perceptual colour as a "characteristic of visual perception" and explicitly includes adaptation and surround. Read as the page discussion of perception, not a denial that light has measurable spectral properties.
- CONFIRMED | “Both are about colour as a property of the observer, not the light.” | https://cie.co.at/eilvterm/17-22-040 | Checker reread retained from N2-codex.md: C154 `nearby-observer`. CIE defines perceptual colour as a "characteristic of visual perception" and explicitly includes adaptation and surround. Read as the page discussion of perception, not a denial that light has measurable spectral properties.
- CONFIRMED | (added) “The Berry the Supreme Court Called a Vegetable” | https://supreme.justia.com/cases/federal/us/149/304/; https://www.kew.org/read-and-watch/vegetable-fruit-seeds-up-close | Checker reread retained from N2-codex.md: C173 `added-tomato-card`. Omitted factual title in the nearby links. Nix v. Hedden, 149 U.S. 304, classifies tomatoes as vegetables under the 1883 tariff act; Kew says "The tomato fruit is technically a berry". The title is supported. Source text/output: “Tomatoes are "vegetables," and not "fruit," within the meaning of the Tariff Act”.

## Findings
- [fixed] X1: Corrected the affirmative paint claim in the page and catalogue, the OG endorsement, and the nearby card via this layer's own relates note. White paint is an achromatic colour, without hue. The dated paint correction is mirrored in the markdown.
- [fixed] X2: Corrected the affirmative paint claim in the page and catalogue, the OG endorsement, and the nearby card via this layer's own relates note. White paint is an achromatic colour, without hue. The dated paint correction is mirrored in the markdown.
- [declined] X3: I read the abstract and introduction of https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/, retrieved through the search index after the direct page returned a browser challenge. The abstract says "highest luminance surface appears nearly white" in the tested simple surface displays. The summary at f5e5f84f34:public/strata/there-is-no-white/index.html:176 turns that heuristic into a general definition. This is MINOR scope/precision wording. Tier C: list; its sentence is separate from the paint assertion being fixed in the same paragraph.
- [fixed] X4: Qualified the light recipe by spectral power balance and separated real white surfaces from the perfect-reflector ideal. The dated light-and-surface correction is mirrored in the markdown.
- [fixed] X5: Qualified the light recipe by spectral power balance and separated real white surfaces from the perfect-reflector ideal. The dated light-and-surface correction is mirrored in the markdown.
- [fixed] X6: Qualified the light recipe by spectral power balance and separated real white surfaces from the perfect-reflector ideal. The dated light-and-surface correction is mirrored in the markdown.
- [fixed] X7: The page, catalogue, screen text and live/offline checks distinguish mixed white light from a single wavelength reaching E. The shared narration at public/tv/lib/narration/batch-15.js:803 still says "No beam of light is white". NASA's prism explanation describes white mixed light; the single-wavelength calculation does not rule it out. This shared-file copy remains outside the repair scope and needs a board note; the directing instance must replace PENDING with its issued id. The directing instance corrected that narration copy on 2026-10-14.
- [declined] X8: I read https://cie.co.at/eilvterm/17-23-058, which defines the purple boundary using mixtures, and f5e5f84f34:public/strata/there-is-no-white/index.html:476-481, which draws the spectral curve and the faint dashed purple closure separately. The caption at line 195 uses rim loosely for a diagram whose closed boundary includes both. The intended spectral-curve statement is supported, so retain MINOR precision, not a false claim about the actual drawn curve. Tier C: list; no WRONG fix shares this sentence.
- [declined] X9: I read https://www.adobe.com/creativecloud/design/discover/is-black-a-color.html: its colour/shade section says "white light comprises all hues", not the exact schoolbook wording. I also read f5e5f84f34:public/strata/there-is-no-white/index.html:218, which presents an unattributed familiar saying rather than an Adobe quotation. The defect is N1 schoolbook-quotation being marked as a sourced verbatim quote. Tier C: list this MINOR record issue; classify or source the saying accurately in the record. Do not edit the words inside the quotation, and do not characterize it as a fabricated Adobe quotation.
- [fixed] X10: Kept the demonstrated reference whites and named them accurately: full spectrum and two-line recipes target E, screen primaries target D65. Metamerism is conditional on equal XYZ, including luminance. The offline check now compares unrounded XYZ at Y=1; the dated recipe correction is mirrored in the markdown.
- [fixed] X11: Replaced coarse near-match acceptance with a verified intersection of the interpolated spectral locus and a final residual <=1e-9 in xy, in the live and offline solvers. Both now reject 494 nm and 569 nm; their 0.5 nm scan gives a gap of 494 to 569.5 nm. The dated solver correction is mirrored in the markdown.
- [fixed] X13: Kept the demonstrated reference whites and named them accurately: full spectrum and two-line recipes target E, screen primaries target D65. Metamerism is conditional on equal XYZ, including luminance. The offline check now compares unrounded XYZ at Y=1; the dated recipe correction is mirrored in the markdown.
- [fixed] X14: Kept the demonstrated reference whites and named them accurately: full spectrum and two-line recipes target E, screen primaries target D65. Metamerism is conditional on equal XYZ, including luminance. The offline check now compares unrounded XYZ at Y=1; the dated recipe correction is mirrored in the markdown.
- [fixed] X15: The page and catalogue identify real chromatic adaptation as approximate and constant neutral paper as an ideal Bradford-model result. The shared narration at public/tv/lib/narration/batch-15.js:805 still says "Turn the lamp from daylight to fire, and the paper stays white anyway." CIE defines approximate compensation, not guaranteed constancy. This shared-file copy remains outside the repair scope and needs a board note; the directing instance must replace PENDING with its issued id. The directing instance corrected that narration copy on 2026-10-14.
- [declined] X16: I read the CIE definition at https://cie.co.at/eilvterm/17-22-013, which describes approximate chromatic compensation, and f5e5f84f34:public/strata/there-is-no-white/index.html:291 and 318, which discuss the simulated paper and ideal transform. Comparing that ideal complete compensation with the way the reader actually sees is a MINOR qualification issue. Tier C: list. This is a different sentence from the universal real-surface claim being fixed for X15.
- [declined] X17: I read the abstract of https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/: the evidence concerns anchoring lightness in simple surface displays and a near-white appearance, not a universal definition of all whites. At f5e5f84f34:public/strata/there-is-no-white/index.html:305 the brightest-neutral wording is a useful but overgeneralized heuristic. Retain MINOR and list under tier C; no WRONG finding requires editing this sentence.
- [declined] X18: I read runCheck at f5e5f84f34:public/strata/there-is-no-white/index.html:705-724 and the assertions at f5e5f84f34:research/there-is-no-white/verify.mjs:139-215. The browser has eight assertions; the standalone program has fifteen, including extra identity, rim and metamer assertions and different solver precision. The advertised 15/15 result need not be changed into a failure. The issue is the overstatement that the check sets are the same. Under the supplied rule for claims about the page's own checking this is MINOR. Tier C: list; no WRONG fix shares the sentence at page line 308.
- [declined] X19: I re-read f5e5f84f34:public/strata/there-is-no-white/index.html:318 and 422-438, and f5e5f84f34:research/there-is-no-white/verify.mjs:181-190. The standalone metamer assertion accepts "dist(cTwo,cFlat)<0.02" after rounding the complementary wavelength to a 5 nm bin, while the browser uses a different solver and leaves final acceptance residuals unchecked. The sentence promises recomputation to the digit offline and live. Relevel WRONG to MINOR under the user's explicit rule that overclaims about the page's own checking are MINOR: this item overstates the precision and equivalence of that checking. This does not downgrade the actual false metamer equality or false successful complements, which remain WRONG in X10/X11/X13/X14/X28. Tier C: list; those WRONG assertions are in other sentences.
- [declined] X20: I read https://cie.co.at/eilvterm/17-23-009: achromatic appearance is defined under the prevailing adaptation conditions. The page's reference-white locus calculation does not establish its unconditional scope wording at f5e5f84f34:public/strata/there-is-no-white/index.html:318. Keep this MINOR scope qualification and list under tier C. The quoted phrase can remain intact; any later clarification should qualify its scope outside the quotation. This is a separate sentence from the checking-precision claim.
- [fixed] X21: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- [declined] X22: I read https://cie.co.at/eilvterm/17-23-021 and 17-23-023, and checked the newer official standard at https://www.cie.co.at/publications/colorimetry-part-2-cie-standard-illuminants-0. ISO/CIE 11664-2:2022 explicitly defines "D50 as a new standard illuminant". The checker's implication that D50 is not standard is outdated and must not be adopted. E remains an equi-energy spectrum treated as an illuminant, so the collective heading at f5e5f84f34:public/strata/there-is-no-white/index.html:327 still has a MINOR terminology issue. List only that grouping issue under tier C; preserve the standard status of A, D65 and D50.
- [declined] X23: I read f5e5f84f34:public/strata/there-is-no-white/index.html:331 and 705-724 and f5e5f84f34:research/there-is-no-white/verify.mjs:139-215. They use the same CMF table and related calculations, but the eight live assertions do not reproduce all fifteen offline assertions, and the offline metamer check additionally bins the complementary wavelength. Treat the broad same-checking implication as MINOR, not as a failed numerical check. Tier C: list, sharing the label with X18 and X30; no WRONG fix shares this sentence.
- [fixed] X24: Replaced no-database wording with aggregate page-view and browser-pulse counting in a database. Read worker.js at f5e5f84f34 and at publication commit 9560d8e7f963623965457330cce236c2d3fbf1e2; SQLite page counters already existed then. The dated database correction is mirrored in the markdown; no claim of personal identification or cross-site tracking was inferred.
- [fixed] X26: Corrected the affirmative paint claim in the page and catalogue, the OG endorsement, and the nearby card via this layer's own relates note. White paint is an achromatic colour, without hue. The dated paint correction is mirrored in the markdown.
- [fixed] X27: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- [fixed] X28: Kept the demonstrated reference whites and named them accurately: full spectrum and two-line recipes target E, screen primaries target D65. Metamerism is conditional on equal XYZ, including luminance. The offline check now compares unrounded XYZ at Y=1; the dated recipe correction is mirrored in the markdown.
- [declined] X29: I read the abstract of https://pmc.ncbi.nlm.nih.gov/articles/PMC4245049/ and the relevant sentence in f5e5f84f34:src/content/strata/there-is-no-white.md:4. The anchoring account is supported for the tested surface displays, but it is a loose universal definition of white. Retain MINOR and list under tier C, sharing the label with X3/X17. This sentence is separate from the paint, beam and metamer sentences being corrected in the same dek.
- [declined] X30: I read the last sentence of the dek at f5e5f84f34:src/content/strata/there-is-no-white.md:4 and compared f5e5f84f34:public/strata/there-is-no-white/index.html:705-724 with f5e5f84f34:research/there-is-no-white/verify.mjs:139-215. Eight live assertions and fifteen related offline assertions are different sets. The 15/15 offline report is not disproved; the same-checks wording overstates equivalence. This is MINOR under the explicit checking-overclaim rule. Tier C: list, sharing the error with X18/X23; the WRONG claims occur in other dek sentences.
- [fixed] X31: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- [fixed] X32: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- [fixed] X33: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- [declined] X34: I read https://registry.color.org/rgb-registry/srgb, which specifies an sRGB reference white, and f5e5f84f34:public/strata/there-is-no-white/index.html:572. It cannot establish the actual calibration, viewing settings or white point of every reader's display. This is a MINOR qualification of the reference example. Tier C: list. The full-drive assertion is one sentence; the screen-never-emits-white assertion fixed for X32/X33 is the next sentence. Sharing one JavaScript string does not trigger the same-sentence exception.
- [fixed] X35: Restricted the assertion to a single wavelength reaching E for the stated CIE 1931 model. Corrected page prose, catalogue/body, screen text and live/offline transcript copies, and regenerated the published verifier. The dated mixed-light correction is mirrored in the markdown. Shared TV narration remains outside the permitted scope.
- [declined] X36: I read f5e5f84f34:public/strata/there-is-no-white/index.html:570, which explicitly says "not pure wavelengths", and lines 713-714, which sum sRGB primary tristimulus vectors while labeling them lines. ICC https://registry.color.org/rgb-registry/srgb supplies primary chromaticities, not monochromatic spectral lines. This is MINOR terminology inconsistent with the page's own qualification. Tier C: list. It is a different generated check sentence from the no-light-is-white error at line 710.
- [open note:56b327] X37: This layer's authored relationship note and Nearby card distinguish colour from hue. On 2026-10-18 the originating note at src/content/strata/is-a-tomato-a-fruit-or-vegetable.md:14 and its Nearby card on public/strata/is-a-tomato-a-fruit-or-vegetable/index.html were replaced with "Both questions need their terms named: white paint has color without hue; a tomato can be a botanical fruit and a culinary vegetable.", and that page and its markdown carry a dated Corrected 2026-10-18 line. CIE classifies white as achromatic colour. The only thing left is the preserved quotation in There, Here (public/strata/there-here/index.html:250 and research/lineage-edge-grammar/page.json:1, notes[644]), which still says both answers are right once you name the frame and needs an annotation.
- [dated] X38: Set JSON-LD dateModified to 2026-10-14. No separate Updated line under the explicit metadata exception: the seven Corrected lines already date this revision. datePublished remains 2026-06-29.
- [declined] X39: I read the 600 + 489 button at f5e5f84f34:public/strata/there-is-no-white/index.html:247, its 600 nm handler at lines 618-622, and complementOf at lines 422-438. Running the extracted pinned arithmetic gives B=490.00485583442924 nm, which the displayed readout rounds to 490. The static label is a small numerical precision/preset slip. Retain MINOR and list under tier C; no WRONG fix shares this button label.

## Scope and evidence

The numerical reread and regression results are in research/assay-line/cited-1014b/there-is-no-white/fix-evidence.json. The fixer's before/after report, exact dated lines, commands and limits are in research/assay-line/cited-1014b/there-is-no-white/fix-report.md. The repair review and board-note text are in research/assay-line/cited-1014b/there-is-no-white/repair-report.md. Its historical layer note also corrects the beam and metamer assertions. The remaining shared narration and tomato-layer source note are held open under X7, X15 and X37 above; the local repairs described in the claims do not close those copies. Published checks and their manifest, search, placards, Nearby boxes, plain cards and social cards are rebuilt at landing. The repair did not run the layer build script, edit shared files or another layer, or render or re-encode film, audio or images.