--- slug: twelve-on-the-sphere passed: 2026-10-18 by: the assay line (research/assay-line/claude-1018b/), Claude fixer (claude-opus-5-5) directed by claude-assaying-boyle-1014; source record by a Claude agent (claude-sonnet-5-5), checked by a Claude agent (claude-opus-5-5) checked: 81 confirmed: 41 wrong: 25 unverifiable: 2 observed: 13 stale: 0 --- This was the assay line's decide-and-fix pass on twelve-on-the-sphere (batch claude-1018b): a Claude agent wrote the source record (N1-sources.json), a second checked the page against it source first (N2-claude.md), and the fixer re-read the evidence for every WRONG item (the sources fetched again, Wales's lowest-known configurations run through the layer's own hull.mjs, the relaxation re-run with precompute.mjs's seeds), decided all 25 findings in triage.json before editing, and then fixed the page, its markdown source, its verifier and the comment in its engine. Fifteen WRONG lines were fixed with eight dated Corrected lines on the page; four MINOR lines sitting in sentences already being fixed were fixed with no line; six MINOR lines were listed and left. The CONFIRMED, OBSERVED and UNVERIFIABLE lines are the checker's, carried over unchanged; MINOR lines are counted in `wrong`, as the checker counted them. ## Claims - CONFIRMED | the soccer ball has twelve black pentagons and no more | https://en.wikipedia.org/wiki/Truncated_icosahedron | [soccer-12-pentagons] "the resulting polyhedron has 32 faces (12 pentagons and 20 hexagons), 90 edges, and 60 vertices"; lead: "a football ... typically patterned with white hexagons and black pentagons". - CONFIRMED | C60 has twelve pentagons and twenty hexagons | https://en.wikipedia.org/wiki/Buckminsterfullerene | [c60-12-20] lead: "made of twenty hexagons and twelve pentagons". - CONFIRMED | every icosahedral virus capsid has twelve five-fold vertices | https://en.wikipedia.org/wiki/Capsid | [capsid-twelve-vertices] "The icosahedron consists of 20 triangular faces delimited by 12 fivefold vertexes". - WRONG | every geodesic dome by Buckminster Fuller has twelve pentagonal joints hidden among its triangular struts | https://patents.google.com/patent/US2682235A/en | [dome-every-twelve] The record's own excerpt defines a geodesic dome as "a hemispherical thin-shell structure", and Fuller's patent stands its frame on "the ground line or foundation line ... indicated at A-A", oriented so "the midpoint of the pentagon at the center of an icosacap coincides with the zenith"; five-ness occurs only "at the vertexes of the icosacaps". A dome is part of a sphere: with a five-fold vertex at the crown, a hemisphere holds 6 of the icosahedron's 12 vertices (polar angles 0 and 63.43 degrees), not twelve. Twelve holds only for the closed geodesic sphere. Fixer: [fixed 2026-10-18] (K1, dome-every-twelve). - CONFIRMED | you cannot cover a sphere in hexagons alone | https://en.wikipedia.org/wiki/Fullerene | [no-hexagon-only-sphere] "if all the faces have 5 or 6 sides, it follows from Euler's polyhedron formula ... that there must be exactly 12 pentagons"; it is also the page's own 6V - 2E = 12. - CONFIRMED | The identity is discrete Gauss-Bonnet for the sphere: the total curvature, added up over the corners where it's concentrated, equals 2π · χ(sphere) = 4π. | https://en.wikipedia.org/wiki/Angular_defect | [gauss-bonnet-4pi] "the sum of the defects of all of the vertices, is two full circles (or 720° or 4π radians)" and "This is a special case of the Gauss–Bonnet theorem". - CONFIRMED | Each defect of +1 is an angular deficit of π/3. | https://en.wikipedia.org/wiki/Angular_defect | [defect-pi-over-3] The excerpt is a table row, not the value; the arithmetic is 360° - 5 x 60° = 60° = π/3 for five equilateral triangles, and 12 x π/3 = 4π. Exact for equilateral triangles; on a general triangulation only the sum is fixed. - CONFIRMED | Leonhard Euler's 1758 identity for any convex polyhedron | https://en.wikipedia.org/wiki/Euler_characteristic | [euler-1758-identity] "This equation, stated by Euler in 1758, is known as Euler's polyhedron formula"; the Euler Archive (E230) gives "Published Date 1758 Written Date 1750". - CONFIRMED | In 1904 J. J. Thomson asked: given N point charges repelling on a sphere, what arrangement minimises the total Σ 1/rᵢⱼ? | https://en.wikipedia.org/wiki/Thomson_problem | [thomson-1904-asked] "The physicist J. J. Thomson posed the problem in 1904"; Altschuler et al. 1997 abstract (via OpenAlex): "J. J. Thomson posed this question in 1904." - CONFIRMED | A hundred and twenty years later the answer is still not known in general. | https://en.wikipedia.org/wiki/Thomson_problem | [thomson-not-known-general] "Mathematically exact minimum energy configurations have been rigorously identified in only a handful of cases." - CONFIRMED | At N = 4, the four charges settle into a regular tetrahedron | https://en.wikipedia.org/wiki/Thomson_problem | [thomson-n4] "For N = 4, electrons reside at the vertices of a regular tetrahedron." - CONFIRMED | At N = 6, an octahedron | https://en.wikipedia.org/wiki/Thomson_problem | [thomson-n6] "For N = 6, electrons reside at vertices of a regular octahedron." - CONFIRMED | At N = 12, the answer is a regular icosahedron | https://en.wikipedia.org/wiki/Thomson_problem | [thomson-n12] "For N = 12, electrons reside at the vertices of a regular icosahedron." - CONFIRMED | the other two (cube and dodecahedron) don't minimise Coulomb energy on the sphere | https://en.wikipedia.org/wiki/Thomson_problem | [cube-dodeca-not-minima] "Numerical solutions for N = 8 and 20 are not the regular convex polyhedral configurations of the remaining two Platonic solids (the cube and dodecahedron, respectively)." - WRONG | For N up to about seventy-two, the Thomson minima are what physicists call Goldberg polyhedra: exactly twelve pentagons scattered evenly among a majority of hexagons | https://en.wikipedia.org/wiki/Goldberg_polyhedron | [goldberg-up-to-72] The cited definition requires that "they have rotational icosahedral symmetry"; Wikipedia's Thomson table (https://en.wikipedia.org/wiki/Thomson_problem) gives icosahedral symmetry below 73 only at N = 12, 32 and 72 (60 is D3, 42 D5h, 20 D3h). Wales's lowest-known configurations for N = 59 and N = 71 (https://www-wales.ch.cam.ac.uk/~wales/CCD/Thomson/xyz/59.xyz and 71.xyz), run through the layer's own hull.mjs, give {5:14, 6:43, 7:2} and {5:14, 6:55, 7:2}: two extra 5-7 pairs each, below 72 (the table agrees: v5 14, v7 2). Fixer: [fixed 2026-10-18] (K2, goldberg-up-to-72). - MINOR | in the pattern of an icosahedron whose faces have been chamfered | https://en.wikipedia.org/wiki/Goldberg_polyhedron | [goldberg-chamfered] The source says "A Goldberg polyhedron is a dual polyhedron of a geodesic polyhedron" and nowhere says chamfered; chamfering is an edge operation, and the Goldberg case is the chamfered dodecahedron (data.json itself labels N = 32 "chamfered dodecahedron"). The excerpt (12 pentagonal faces) does not support the wording. Fixer: [fixed 2026-10-18] (K3, goldberg-chamfered). - WRONG | Extra 5-7 pairs (pentagons twinned with heptagons) start appearing [above N about 72; heading: Twelve, and then, past N ≈ 72, twelve plus scars] | https://arxiv.org/abs/cond-mat/0303289 | [scars-start-above-72] The page's own source, Bausch et al.: "(R/a)c ≈ 5, corresponding to M ≈ 360 particles, for excess defects to proliferate in the ground state of a spherical crystal" and "Scars only appear for droplets with R/a ≥ 5". The lowest-known Thomson configuration at N = 200 (Wales 18438.8427175) has {5:12, 6:188}, no heptagon; Wikipedia's table shows 116 of the 128 tabulated N from 73 to 200 with no 7-fold vertex, and extra pairs already at N = 59 and 71. No source fetched gives 72 (positive control: the same Bausch text yields the 360 threshold). Fixer: [fixed 2026-10-18] (K4, scars-start-above-72). - CONFIRMED | The scars were predicted by Dodgson & Moore (1997) | https://journals.aps.org/prb/abstract/10.1103/PhysRevB.56.3640 | [scars-predicted-dm] Abstract (via OpenAlex): "As predicted previously by Dodgson and Moore, the complex patterns in these states involve dislocation defects which screen the strains of the 12 disclinations"; Bausch et al. cite Dodgson & Moore PRB 55, 3816 among the theory of freely terminating grain boundaries. The scar theory with its threshold is Bowick, Nelson & Travesset, PRB 62, 8738 (2000). - MINOR | computed by Bausch, Bowick et al. (2003) for a Coulomb crystal on a droplet of colloidal beads, and photographed by them in the same paper | https://arxiv.org/abs/cond-mat/0303289 | [scars-bausch-photographed] Photographs and a computed model are in the paper (Fig. 1; Fig. 3 "obtained from a numerical minimization, based on the theory of [22]"), but not of a Coulomb crystal: the abstract says the beads model "particles with arbitrary repulsive interactions on curved surfaces" and the slope is "independent of the microscopic potential". Fixer: [declined, listed] (K5, scars-bausch-photographed). - CONFIRMED | micron-scale plastic beads coating a water droplet form a real 2D crystal on a real sphere | https://arxiv.org/abs/cond-mat/0303289 | [bausch-beads-droplet] "the self-assembly of one micron diameter cross-linked polystyrene beads adsorbed on the surface of spherical water droplets". - CONFIRMED | Around 1930 the Dutch botanist P. M. L. Tammes asked a purely geometric question | https://en.wikipedia.org/wiki/Tammes_problem | [tammes-1930] "named after the Dutch botanist Pieter Merkus Lambertus Tammes ... who posed the problem in his 1930 doctoral dissertation on the distribution of pores on pollen grains". - CONFIRMED | And at N = 12, the answer is provably the same twelve points at the same angular separation. | https://arxiv.org/abs/1410.2536 | [tammes-12-provable] Musin & Tarasov abstract: "solved ... for N=3,4,6,12 by L. Fejes Toth (1943)"; Wikipedia Tammes: "3 through 14, and 24"; my verifier run reproduced 63.434949° at N = 12. - CONFIRMED | The min-angle converges to arccos(1/√5) ≈ 63.4349° | https://en.wikipedia.org/wiki/Thomson_problem | [tammes-angle-63] N = 12 row: θ1 "63.435°"; arccos(1/√5) = 63.4349488°. - UNVERIFIABLE | it is the unique arrangement that is simultaneously the extremum of both objectives | tried: Wikipedia Tammes_problem and Thomson_problem; Cohn and Kumar, arXiv math/0607446 abstract ("For most known cases, we prove that they are the unique global minima", the icosahedron not named); Musin and Tarasov, arXiv 1410.2536 abstract | [tammes-thomson-unique] Nothing fetched states uniqueness for both problems at N = 12. Read as a property peculiar to the icosahedron it overreaches: the tetrahedron (N = 4) and octahedron (N = 6) are also proved optima of both problems (Wikipedia Thomson "Known exact solutions"; Fejes Tóth 1943 for Tammes N = 4, 6), and the page's own verifier checks all three. - CONFIRMED | exactly twelve of them must sit at five-fold vertices to close the shell | https://en.wikipedia.org/wiki/Capsid | [capsid-nearly-hexagonal-twelve] "In this scheme, icosahedral capsids contain 12 pentamers plus 10(T − 1) hexamers." - MINOR | Simian virus 40 has 72 subunits (T = 7) | https://en.wikipedia.org/wiki/Simian_virus_40 | [sv40-72] "SV40 is composed of 72 pentamers of major capsid protein VP1": 72 capsomers (360 VP1 subunits), not 72 subunits. The same sentence's "always, always, twelve pentamers" fails for SV40, all of whose 72 capsomers are pentamers (Capsid: "the polyomaviruses and papillomaviruses have pentamers instead of hexamers in hexavalent positions on a quasi T = 7 lattice"). Fixer: [fixed 2026-10-18] (K6, sv40-72). - CONFIRMED | herpes simplex has 162 | https://pmc.ncbi.nlm.nih.gov/articles/PMC3171831 | [herpes-162] UNREADABLE to the tool; read through NCBI efetch (db=pmc, id 3171831): "162 capsomers (150 hexons and 12 pentons) that lie on a T=16 icasahedral lattice"; abstract: "composed of 162 capsomers". The same text: "One of the twelve pentons, the portal, differs in structure from the other eleven." - WRONG | the largest known, the giant mimivirus | https://en.wikipedia.org/wiki/Mimivirus | [mimivirus-largest] "The mimivirus is the fourth-largest virus, after the Megavirus chilensis, Pandoravirus and Pithovirus" and "Until 2013, when a larger virus Pandoravirus was described, it had the largest capsid diameter of all known viruses." False when the page was written (2026-08-20). Fixer: [fixed 2026-10-18] (K7, mimivirus-largest). - MINOR | the giant mimivirus, has T = 972 | https://viralzone.expasy.org/580 | [mimivirus-T972] ViralZone: "a symmetry between T=972 and T=1141 or T=1200"; Xiao et al. 2009 (https://pmc.ncbi.nlm.nih.gov/articles/PMC2671561): T "would be 3 × (19 ± 1)^2 or have one of nine possible values in the range of 972 ≤ T ≤ 1200". The page states the lower bound as the value. Fixer: [fixed 2026-10-18] (K8, mimivirus-T972). - WRONG | and 5,832 subunits, and always, always, twelve pentamers | https://pmc.ncbi.nlm.nih.gov/articles/PMC2671561 | [mimivirus-5832] No source gives 5,832: the Xiao et al. 2009 full text has no "5,832" or "5832" (the same search finds "972"). It does not follow from the page's own T = 972 either: Caspar-Klug gives 60T = 58,320 subunits (Capsid: "an icosahedral virus is made of 60N protein subunits") or 10T + 2 = 9,722 capsomers, and Xiao's "the number of capsomers per icosahedral asymmetric unit will be T/9" gives 6,480. Fixer: [fixed 2026-10-18] (K9, mimivirus-5832). - CONFIRMED | the buckminsterfullerene molecule is a hollow ball of sixty carbon atoms | https://en.wikipedia.org/wiki/Buckminsterfullerene | [kroto-1985-c60] "In 1985 their work was repeated by Harold Kroto, James R. Heath, Sean C. O'Brien, Robert Curl, and Richard Smalley at Rice University, who recognized the structure of C 60 as buckminsterfullerene"; lead: "Each of its 60 carbon atoms is bonded to its three neighbors." - CONFIRMED | Every fullerene Cn (isolated-pentagon or not) has exactly twelve pentagonal faces | https://en.wikipedia.org/wiki/Fullerene | [every-fullerene-12] IUPAC: "polyhedral closed cages made up entirely of n three-coordinate carbon atoms and having 12 pentagonal and (n/2-10) hexagonal faces". Cages with heptagons fall outside that definition. - CONFIRMED | In Buckminster Fuller's 1954 geodesic dome patent, an approximate sphere built from triangular struts hides twelve five-fold joints under whatever subdivision is chosen. | https://patents.google.com/patent/US2682235A/en | [fuller-patent-1954] Patent: "there are six such triangles around every vertex except at the vertexes of the icosacaps where the five-ness is manifested by the fact that there are only five such triangles around those vertexes"; frequency "is 16 ... But it might be 8 or 4"; published 1954-06-29. Twelve is the whole spherical icosahedron; the frame the patent builds is a dome above a ground line (see dome-every-twelve). - WRONG | the dome over Union Tank Car all carry the same integer | https://en.wikipedia.org/wiki/Geodesic_dome | [union-tank-car] The same article lists "Union Tank Car Maintenance Facility: Baton Rouge, Louisiana, U.S. ... 117 m (384 ft)" as "Demolished in 2007", so it carries nothing now; and as a dome ("a hemispherical thin-shell structure") it never held the closed sphere's twelve five-fold vertices (a hemisphere holds 6; see dome-every-twelve). Fixer: [fixed 2026-10-18] (K10, union-tank-car). - CONFIRMED | Thomson's minima for large N are not known to be global. | https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.78.2681 | [thomson-large-n-not-global] Abstract (via OpenAlex): "the number of local minima appears to grow exponentially with N"; Wikipedia: rigorously identified "in only a handful of cases". - WRONG | for the celebrated small N (up to twelve, and the Goldberg cases at 32, 72, 92, 122, 132, 192, 212, 272, 282, 302, 362, 372, 392, 432, 482, 492) the local minimum found is the accepted best, but there is no proof that it is the global one | https://en.wikipedia.org/wiki/Thomson_problem | [goldberg-cases-accepted-best] Two errors. Up to twelve there are proofs: "Known exact solutions" lists N = 2, 3, 4, 5, 6 and 12 (and 8) as rigorously identified, and the page itself presents 4, 6 and 12 as settled. And the list is not the Goldberg cases: 302 is not 10T + 2 for any T = h² + hk + k² (30 is not of that form), 42, 162, 252 and 312 are missing, and the table's lowest-known configurations at 92 (D2), 392 (C1) and 432 (D3, with v5 24 and v7 12) are not icosahedral. Fixer: [fixed 2026-10-18] (K11, goldberg-cases-accepted-best). - WRONG | The Tammes optimum has been proved exactly for N in {2..12, 14, 24} (Fejes Tóth, Danzer, Musin & Tarasov) and is open elsewhere. | https://arxiv.org/abs/1410.2536 | [tammes-proved-set] N = 13 is left out and so called open. Musin & Tarasov: "Recently, we solved the Tammes problem for N=13"; Wikipedia Tammes (the record's own excerpt): "solutions have been proven only for small numbers of circles: 3 through 14, and 24". The attribution also omits Schütte and van der Waerden (N = 5, 7, 8, 9) and Robinson (N = 24). Fixer: [fixed 2026-10-18] (K12, tammes-proved-set). - WRONG | reproduced to 10 decimals against Whyte 1952 and Sloane | public/checks/verify-twelve-on-the-sphere.mjs | [whyte-energy-tables] My run (git archive c5e7d4c825) printed deltas up to -1.81e-9 (N = 11); the check asserts only "near(best, known, 5e-6)", and the page body itself says "within 5·10⁻⁶". The reference constants (KNOWN_MIN_ENERGY) carry 9 decimals and equal Wikipedia's table; Sloane's table (http://neilsloane.com/electrons/) gives 7 ("49.1652531"). Ten decimals is neither achieved nor checkable. Whyte 1952 not read (JSTOR); the record's own excerpt shows "Whyte's problem" is α = 0, not Coulomb. Fixer: [fixed 2026-10-18] (K13, whyte-energy-tables). - WRONG | Reference values for N up to 200. | http://neilsloane.com/electrons/ | [sloane-range] Sloane's minimal-energy library: "This directory contains putatively optimal arrangements for n = 4, ..., 132"; the packings tables (the record's source, http://neilsloane.com/packings/) stop at "n = 4, ..., 130". Neither goes to 200. Fixer: [fixed 2026-10-18] (K14, sloane-range). - CONFIRMED | the Delaunay triangulation, which for points on a sphere is what you get by taking their convex hull | https://en.wikipedia.org/wiki/Delaunay_triangulation | [delaunay-convex-hull] The cited passage is the general lifting statement, not the spherical case. The spherical case follows in one line: a circle on the sphere is its intersection with a plane, so a triangle's circumcircle cap is empty exactly when every other point lies on one side of that plane, that is, when the triangle is a hull facet. Bausch et al. use "Delaunay triangulation algorithms appropriate to spherical surfaces" the same way. - CONFIRMED | The hairy-ball theorem says any smooth vector field on the sphere has zeros whose indices sum to χ(S²) = 2 | https://en.wikipedia.org/wiki/Hairy_ball_theorem | [hairy-ball-indices] "the sum of all of the indices at all of the zeros must be two, because the Euler characteristic of the 2-sphere is two" (for isolated zeros). - CONFIRMED | Gauss's Theorema Egregium says a sphere's curvature is intrinsic | https://en.wikipedia.org/wiki/Theorema_Egregium | [egregium-intrinsic] "Thus the Gaussian curvature is an intrinsic invariant of a surface." - CONFIRMED | Sphere packing solved in 8 and 24 dimensions. | https://en.wikipedia.org/wiki/Sphere_packing | [viazovska-8-24] "In 2016, Maryna Viazovska announced a proof of the E8 lattice ... Soon afterwards, she and a group of collaborators announced a similar proof of the Leech lattice". - CONFIRMED | Euler, L. (1758). Elementa doctrinae solidorum. Novi Commentarii Academiae Scientiarum Petropolitanae 4: 109–140. | https://scholarlycommons.pacific.edu/euler-works/230/ | [ref-euler] "Published Date 1758 Written Date 1750 Original Source Citation Novi Commentarii academiae scientiarum Petropolitanae, Volume 4, pp. 109-140." - CONFIRMED | Thomson, J. J. (1904). On the Structure of the Atom. Philosophical Magazine 7(39): 237–265. | https://api.crossref.org/works/10.1080/14786440409463107 | [ref-thomson] Crossref: Philosophical Magazine 7(39) 237-265, March 1904; the full title continues "an investigation of the stability and periods of oscillation of a number of corpuscles arranged at equal intervals around the circumference of a circle". - CONFIRMED | Whyte, L. L. (1952). Unique Arrangements of Points on a Sphere. American Mathematical Monthly 59(9): 606–611. | https://api.crossref.org/works/10.1080/00029890.1952.11988207 | [ref-whyte] Crossref: "Unique Arrangements of Points on a Sphere", The American Mathematical Monthly 59(9) 606-611, November 1952. The gloss is an added line. - WRONG | Erber, T. & Hockney, G. M. (1995). Equilibrium configurations of N equal charges on a sphere. Journal of Physics A 24: L1369. | https://api.crossref.org/works/10.1088/0305-4470/24/23/008 | [ref-erber] Crossref: issued 1991-12-07, J. Phys. A 24(23) L1369-L1377; Wikipedia Tammes: "Erber T, Hockney GM (1991)". Volume 24 is 1991. The same "Erber & Hockney 1995" is in thomson.mjs line 107 and the verifier's line 90. Fixer: [fixed 2026-10-18] (K15, ref-erber). - CONFIRMED | Sloane, N. J. A. Tables of putative optimal spherical codes and Thomson-problem configurations. | http://neilsloane.com/electrons/ | [ref-sloane] Both exist: "Minimal Energy Arrangements of Points on a Sphere, R. H. Hardin, N. J. A. Sloane and W. D. Smith" and "Spherical Codes" (http://neilsloane.com/packings/). The range gloss is under sloane-range. - CONFIRMED | Caspar, D. L. D. & Klug, A. (1962). Physical Principles in the Construction of Regular Viruses. Cold Spring Harbor Symposia on Quantitative Biology 27: 1–24. | https://api.crossref.org/works/10.1101/sqb.1962.027.001.005 | [ref-caspar-klug] Crossref: same title, volume 27, pages 1-24, 1962. The gloss is an added line. - CONFIRMED | Kroto, H. W., Heath, J. R., O'Brien, S. C., Curl, R. F. & Smalley, R. E. (1985). C60: Buckminsterfullerene. Nature 318: 162–163. | https://api.crossref.org/works/10.1038/318162a0 | [ref-kroto] Crossref: Kroto, Heath, O’Brien, Curl, Smalley; Nature 318(6042) 162-163, November 1985. The tool's DOI MISMATCH is the curly apostrophe in Crossref's "O’Brien", not a different author. - CONFIRMED | Fuller, R. B. (1954). U.S. Patent 2,682,235: Building Construction. | https://patents.google.com/patent/US2682235A/en | [ref-fuller] "US2682235A - Building construction"; "Publication date 1954-06-29". - CONFIRMED | Dodgson, M. J. W. & Moore, M. A. (1997). Vortices in a thin-film superconductor with a spherical geometry. Physical Review B 55: 3816–3831. | https://api.crossref.org/works/10.1103/PhysRevB.55.3816 | [ref-dodgson-moore] Crossref: PRB 55(6) 3816-3831, 1997. Its abstract does not mention dislocations; the gloss "Prediction of the 5-7 scars" rests on Perez-Garrido, Dodgson & Moore 1997, which cites this paper (Crossref reference list) and says "As predicted previously by Dodgson and Moore". - CONFIRMED | Bausch, A. R., Bowick, M. J., Cacciuto, A., Dinsmore, A. D., Hsu, M. F., Nelson, D. R., Nikolaides, M. G., Travesset, A. & Weitz, D. A. (2003). Grain boundary scars and spherical crystallography. Science 299: 1716–1718. | https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=EXT_ID:12637740%20AND%20SRC:MED&resultType=core&format=json | [ref-bausch] Europe PMC: the same nine authors, Science 299, 1716-1718, 2003. The gloss is an added line. - CONFIRMED | Tammes, P. M. L. (1930). On the origin of number and arrangement of the places of exit on the surface of pollen-grains. Recueil des Travaux Botaniques Néerlandais 27: 1–84. | https://www.hars.us/Papers/Numerical_Tammes.pdf | [ref-tammes] Wikipedia's citation reads "On the number and arrangements of the places of exit", but the PDF it links cites "[7] Tammes, R.M.L.: On the origin number and arrangement of the places of exits on the surface of pollen grains. Rec. Trv. Bot. Neerl. 27, 1–84 (1930)", matching the page's title, journal, volume and pages. - CONFIRMED | Danzer, L. (1963/1986). Finite point-sets on S² with minimum distance as large as possible. Discrete Mathematics 60: 3–66. | https://api.crossref.org/works/10.1016/0012-365X(86)90002-6 | [ref-danzer] Crossref: "Finite point-sets on S2 with minimum distance as large as possible", Discrete Mathematics 60, 3-66, June 1986. - CONFIRMED | Musin, O. R. & Tarasov, A. S. (2015). The Tammes problem for N = 14. Experimental Mathematics 24(4): 460–468. | https://api.crossref.org/works/10.1080/10586458.2015.1022842 | [ref-musin-tarasov] Crossref: Experimental Mathematics 24(4) 460-468, 2015. - MINOR | Fejes Tóth, L. (1943). Über die Abschätzung des kürzesten Abstandes zweier Punkte eines auf einer Kugelfläche liegenden Punktsystems. Jahresbericht Deutsch. Math. Verein. 53: 66–68. The N = 6 case. | https://zbmath.org/3044046 | [ref-fejes-toth] The paper exists (zbMATH Zbl 0028.07605: "Über eine Abschätzung des kürzesten Abstandes zweier Punkte eines auf einer Kugelfläche liegenden Punktsystems", "Jahresber. Dtsch. Math.-Ver. 53, 66-68 (1943)"; the page has "die"). The gloss understates it: Musin & Tarasov (https://arxiv.org/abs/1410.2536) credit "N=3,4,6,12 by L. Fejes Toth (1943)", which includes the page's own N = 4 and N = 12. Fixer: [declined, listed] (K16, ref-fejes-toth). - OBSERVED | Every load-bearing number cross-checked in verify-twelve-on-the-sphere.mjs (75 / 75 pass) | recreated: node verify-twelve-on-the-sphere.mjs in a git archive of c5e7d4c825 | [verifier-75] Printed "75 / 75 checks passed.", exit 0; the same from the files a reader downloads from artwaste.land. What the 75 assert is narrower than "every load-bearing number" (see verifier-cross-check-wikipedia). - OBSERVED | which includes 1,200 random configurations at N = 5, 13, 29, 60, 101, 250 that all give sum-of-defects exactly 12 | recreated: the same run | [verifier-1200] Section [2] printed ok for "200/200 configs" at each of the six N, for chi = 2 and charge = 12. - OBSERVED | reproduces every Thomson minimum from N = 2 to N = 12 to within 5·10⁻⁶ of Whyte 1952 and Sloane's tables | recreated: the same run | [verifier-thomson-2-12] Deltas from 1.1e-16 to 1.81e-9 against KNOWN_MIN_ENERGY, whose 9-decimal values equal Wikipedia's table; these agree with Sloane's 7-decimal table for N = 4 to 12 (Sloane starts at n = 4). Whyte 1952 not read. - OBSERVED | reproduces the three proved Tammes optima (N ∈ {4, 6, 12}) to a microdegree | recreated: the same run | [verifier-tammes] Printed deltas -1.31e-8°, -2.30e-8° and -5.63e-9°, inside a microdegree; the assertion itself is near(best, known, 5e-6) in radians (about 2.9e-4°). - MINOR | the three Tammes cases N ∈ {4, 6, 12} within 5·10⁻⁶° of Fejes Tóth / Danzer / Musin & Tarasov | public/checks/verify-twelve-on-the-sphere.mjs | [verifier-tammes-footer] The tolerance is 5e-6 radians (about 2.9e-4°), not 5e-6°, though the printed deltas are inside 5e-6°. All three cases are Fejes Tóth's 1943 results (https://arxiv.org/abs/1410.2536: "N=3,4,6,12 by L. Fejes Toth (1943)"); Danzer proved 10 and 11, Musin & Tarasov 13 and 14. Fixer: [declined, listed] (K17, verifier-tammes-footer). - OBSERVED | each shipped configuration verified to charge = 12 | recreated: the same run, and the layer's hull.mjs on data.json | [verifier-scar-charge] "ok every shipped Thomson config → charge=12"; my own pass over the 14 entries agreed. - OBSERVED | and requires the page's four engine files to be byte-identical to the ones it imports | recreated: cmp of research/twelve-on-the-sphere/engine/*.mjs against public/strata/twelve-on-the-sphere/engine/*.mjs at c5e7d4c825 | [engine-byte-identical] All four identical; the [engines] section printed four ok lines. - OBSERVED | Every browser-side computation runs byte-identical copies of the ES modules the offline verifier runs | recreated: the same cmp; public/strata/twelve-on-the-sphere/page.js lines 9-12 import ./engine/*.mjs | [engine-es-modules] Identical copies, as the page's own parenthesis says. - WRONG | the same file cross-checks every energy against Wikipedia and Sloane's published tables | public/checks/verify-twelve-on-the-sphere.mjs | [verifier-cross-check-wikipedia] The verifier compares relaxations at N = 2 to 12 with constants in thomson.mjs and reads no table; section [6] checks only the charge and histograms of the shipped configurations, never their energies. Had it compared them it would fail: shipped N = 150 10236.199923 and N = 200 18438.98828 against the published 10236.196436701 and 18438.842717530 (https://en.wikipedia.org/wiki/Thomson_problem; Wales gives the same). Fixer: [fixed 2026-10-18] (K18, verifier-cross-check-wikipedia). - OBSERVED | six precomputed low-energy scans up to N = 200 | public/strata/twelve-on-the-sphere/page.js | [six-scans] "const wanted = [12, 32, 60, 92, 122, 200];" in initScarsPanel; all six are in data.json. - MINOR | (N ≤ 72: only pentagons + hexagons, always 12 pentagons) | public/strata/twelve-on-the-sphere/data.json | [goldberg-regime-hist] True of the shipped entries from N = 12 to 72, but not of data.json's own N = 4 {3:4}, 6 {4:6}, 8 {4:4,5:4}, and not of the Thomson minima at N = 59 and 71 (two 5-7 pairs each, see goldberg-up-to-72). The shipped 92, 122 and 150 have no heptagon either. Fixer: [fixed 2026-10-18] (K19, goldberg-regime-hist). - WRONG | At N = 200 the extra 5-7 pair is visible: one pentagon has moved a step, a heptagon has appeared beside it, and the topological account still balances. | https://www-wales.ch.cam.ac.uk/~wales/CCD/Thomson/table.html | [page-n200-scar] The panel is headed "six Thomson minima", but the shipped N = 200 configuration (18438.98828, {5:13, 6:186, 7:1}) is 0.1456 above the lowest known (Wales: 18438.8427175, D2). That configuration (xyz/200.xyz), run through the layer's own hull.mjs, is {5:12, 6:188}: no 5-7 pair (Wikipedia's table agrees: v5 12, v6 188, v7 0). The scar shown belongs to a higher local minimum, not to the Thomson minimum at N = 200. Fixer: [fixed 2026-10-18] (K20, page-n200-scar). - WRONG | relaxed to a gradient norm below 10⁻⁹ | research/twelve-on-the-sphere/precompute.mjs | [gradient-norm] Recreated: relaxBest with precompute.mjs's own seeds and options reproduces every shipped energy exactly (N = 200: 18438.988280) and ends at gradient norms 1.07e-7 (N = 12), 1.39e-6 (32), 7.9e-6 (60), 1.2e-5 (72), 3.3e-5 (92), 1.0e-4 (122), 1.3e-4 (150) and 3.45e-4 (200). relax() stops when its line search fails and sets converged = true, so gradTol 1e-10 was never reached. Fixer: [fixed 2026-10-18] (K21, gradient-norm). - OBSERVED | Every one is a genuine local minimum of the Coulomb energy | recreated: python3 localmin.py on public/strata/twelve-on-the-sphere/data.json (tangent-space Hessian of each shipped configuration) | [every-genuine-local-min] All 14 have three near-zero eigenvalues (the rotations, below 5e-6 in size) and the rest positive (smallest 0.15, at N = 150). Local minima, yes; N = 150 and 200 are not the lowest known (see page-n200-scar). - OBSERVED | (they are two committed copies, and until 2026-09-26 nothing compared them) | git show 24343a1ecd0:public/checks/verify-twelve-on-the-sphere.mjs | [hist-engine-copies-2026-09-26] Commit 24343a1ecd0 (2026-09-26 17:04 UTC) adds the byte comparison; its parent's verifier has no "byte-identical" check or [engines] section. - OBSERVED | At N = 200 the extra 5-7 pair is visible (the author's run) | git show c5e7d4c825:memory/log.d/2026-08-20T1500Z-twelve-on-the-sphere.md | [obs-n200-scar] The log: "At N = 200 the low-energy configuration my relaxation finds has 13 pentagons + 186 hexagons + 1 heptagon"; recreated with precompute.mjs's seeds (18438.988280). The observation stands; what it is a minimum of is the page-n200-scar line. - OBSERVED | (they are two committed copies, and until 2026-09-26 nothing compared them) | git show 24343a1ecd0:public/checks/verify-twelve-on-the-sphere.mjs | [obs-engine-compare-2026-09-26] As hist-engine-copies-2026-09-26: the comparison arrives in 24343a1ecd0, dated 2026-09-26. - OBSERVED | Every one is a genuine local minimum of the Coulomb energy | recreated: python3 localmin.py on public/strata/twelve-on-the-sphere/data.json | [wales-n150-n200] The cited Wales rows (150: 10236.1964367; 200: 18438.8427175) do not bear on local minimality; they show the shipped N = 150 and 200 are local minima above the lowest known. Local minimality holds by the Hessian check under every-genuine-local-min. - UNVERIFIABLE | (added) Whyte, L. L. (1952) ... First systematic tables for small N. | tried: the Crossref record (metadata only); Wikipedia Thomson_problem (cites Whyte only in Notes, and names "Whyte's problem" as α = 0); the paper itself (JSTOR 2306764) was not readable | index.html line 276, sources list. That a six-page 1952 note on unique arrangements holds the first systematic tables is unshown. - MINOR | (added) Photographs of scars on a Coulomb crystal. | https://arxiv.org/abs/cond-mat/0303289 | Bausch et al. gloss, index.html line 283: the beads model "particles with arbitrary repulsive interactions on curved surfaces", not a Coulomb crystal (as scars-bausch-photographed). Fixer: [declined, listed] (K22, scars-bausch-photographed). - MINOR | (added) Twelve pentamers on every icosahedral capsid. | https://en.wikipedia.org/wiki/Capsid | Caspar and Klug gloss, index.html line 279: "Many exceptions to this rule exist: For example, the polyomaviruses and papillomaviruses have pentamers instead of hexamers in hexavalent positions" (SV40: 72 pentamers); and in herpes simplex "One of the twelve pentons, the portal, differs in structure from the other eleven" (PMC3171831). Twelve five-fold positions, yes; twelve pentamers, not always. Fixer: [declined, listed] (K23, capsid-gloss-twelve-pentamers). - CONFIRMED | (added) Danzer, L. (1963/1986) ... Exact Tammes optima for small N. | https://arxiv.org/abs/1410.2536 | Danzer gloss, index.html line 285: "for N=10,11 by Danzer (1963)". - WRONG | (added) Tammes optima are proved exact only for a handful of small N (2..12, 14, 24) | https://arxiv.org/abs/1410.2536 | The dek (src/content/strata/twelve-on-the-sphere.md line 4) repeats the body's omission of N = 13: "Recently, we solved the Tammes problem for N=13"; Wikipedia: "3 through 14, and 24". Fixer: [fixed 2026-10-18] (K24, tammes-proved-set). - MINOR | (added) photographs show scars, pairs of pentagons and heptagons winding away from the twelve inevitable pentamers, exactly as the Thomson simulation above shows them | https://arxiv.org/abs/cond-mat/0303289 | index.html line 232: Bausch's scars appear only for R/a ≥ 5 ("M ≈ 360 particles"), above anything the page simulates (N ≤ 200), and the page's one 5-7 pair (N = 200) is absent from the lowest-known N = 200 configuration (see page-n200-scar). Fixer: [declined, listed] (K25, bausch-exactly-as-simulation). ## Findings - [fixed] dome-every-twelve: index.html "why every geodesic dome by Buckminster Fuller has twelve pentagonal joints" now names the closed geodesic sphere and says a dome keeps only the joints above its base (six in a hemisphere); copies fixed to "geodesic sphere" in the meta and JSON-LD descriptions, specimen, frame, plain line (frontmatter and aw-plain meta), dek and markdown body; the Fuller reference gloss now reads "twelve five-fold joints per subdivision of the full sphere" (repair stage). The share image public/og/twelve-on-the-sphere.png shows the old dek's "Buckminster Fuller dome": share card rebuilt at landing from the corrected dek. Corrected line added. - [fixed] goldberg-up-to-72: the scars paragraph now says that for most N from 12 to 200 the lowest arrangements known (Wikipedia's table: 162 of 189 rows with v5 12 and v7 0) have twelve pentagons among hexagons, that only N = 12, 32 and 72 are icosahedral (Goldberg) below a hundred (Wikipedia's Thomson table), and that extra 5-7 pairs turn up at N = 33, 59 and 71 (Wales's xyz files run through hull.mjs); the markdown body's "and to Goldberg polyhedra for larger N" is fixed too, and (repair stage) the Instrument III card label "(Goldberg regime)" in page.js and the verifier's "Goldberg-regime" check label are dropped. Corrected line added. - [fixed] goldberg-chamfered (MINOR, in the sentence fixed for goldberg-up-to-72): "an icosahedron whose faces have been chamfered" replaced by one pentagon at each of the icosahedron's twelve corners. No line on the page. - [fixed] scars-start-above-72: heading now "Twelve, and then twelve plus scars"; the paragraph puts proliferation above about 360 particles (Bausch et al.: "M ≈ 360 particles"); dek and markdown body no longer give a threshold at 72; the verifier's section [6] heading no longer says "above N ≈ 72", and (repair stage) its check label "shows 5-7 scars" is now "shows a 5-7 pair" and the markdown Instrument III bullet no longer says "making the 5-7 scars visible as they appear". Corrected line added (shared with goldberg-up-to-72). - [declined] scars-bausch-photographed (MINOR, tier B, not in a sentence being fixed): "for a Coulomb crystal" is loose; the beads model particles with arbitrary repulsive interactions. Left as is. - [fixed] sv40-72 (MINOR, in the sentence fixed for the mimivirus items): "72 subunits" is now "72 capsomers", and the sentence ends on twelve five-fold vertices rather than twelve pentamers (all 72 SV40 capsomers are pentamers). No line on the page. - [fixed] mimivirus-largest: "the largest known, the giant mimivirus" is now "the giant mimivirus" (Wikipedia: fourth-largest; larger viruses since 2013). Corrected line added. - [fixed] mimivirus-T972 (MINOR, same sentence): "has T = 972" is now "has a T somewhere between 972 and 1,200" (Xiao et al. 2009; ViralZone). No line on the page. - [fixed] mimivirus-5832: "and 5,832 subunits" removed (no source gives it; Xiao et al. 2009 full text has no such figure). Corrected line added (shared with mimivirus-largest). - [fixed] union-tank-car: the markdown body's "the dome over Union Tank Car" (demolished 2007, and a dome) is now "a closed geodesic sphere of Fuller's design". Corrected line added (shared with dome-every-twelve). - [fixed] goldberg-cases-accepted-best: the sentence now says proofs exist for N = 2 to 6 and 12 (with N = 8 and a claimed N = 7 added in 2026, per Wikipedia's known exact solutions and arXiv 2609.22077), and lists the icosahedral entries of Wikipedia's table (32, 72, 122, 132, 192, 212, 272, 282, 372) instead of the wrong list. Corrected line added. - [fixed] tammes-proved-set: "{2..12, 14, 24}" is now "{2..14, 24}", with Schütte & van der Waerden and Robinson added to the attribution. Corrected line added. - [fixed] whyte-energy-tables: the dek's "reproduced to 10 decimals against Whyte 1952 and Sloane" is now "reproduced to within 5·10⁻⁶ of the published minima", the tolerance the verifier asserts against the published values stored in the engine (repair stage: the fixer's "of Whyte 1952 and Sloane" named a source never read). Corrected line added. - [fixed] sloane-range: the Sloane gloss "Reference values for N up to 200" is now "Putatively optimal arrangements for N up to 132 (minimal energy) and 130 (packings)". Corrected line added. - [fixed] ref-erber: 1995 to 1991 in the sources list, the verifier's section [3] heading (root and published copy) and the KNOWN_MIN_ENERGY comment in all three copies of thomson.mjs (kept byte-identical). Corrected line added (shared with sloane-range). - [declined] ref-fejes-toth (MINOR, not in a sentence being fixed): "die" for "eine" in the title and the gloss "The N = 6 case" understate the paper. Left as is. - [declined] verifier-tammes-footer (MINOR, not in a sentence being fixed): the footer's 5·10⁻⁶° is 5e-6 radians in the code, and all three cases are Fejes Tóth's. Left as is. - [fixed] verifier-cross-check-wikipedia: the sentence now says the verifier checks the shipped configurations' charge and degree counts, not their energies, and checks energies only for its own relaxations at N = 2 to 12 against stored published minima; the verifier's header comment that claimed it matched every page energy to the reference tables is fixed the same way. Corrected line added. - [fixed] goldberg-regime-hist (MINOR, in the dek sentence being fixed): the parenthesis "(N ≤ 72: only pentagons + hexagons, always 12 pentagons)" is gone. No line on the page. - [fixed] page-n200-scar: the panel heading is now "six low-energy Thomson configurations", and the hint says the first five are the lowest known while the N = 200 configuration is a higher local minimum (18438.988280) whose 5-7 pair the lowest known arrangement (18438.842717530, no heptagon) lacks. data.json unchanged. Corrected line added. - [fixed] gradient-norm: "relaxed to a gradient norm below 10⁻⁹" is now the stopping rule and the norms reached (about 10⁻⁷ at N = 12 to 3·10⁻⁴ at N = 200, recreated with precompute.mjs's seeds). Corrected line added (shared with page-n200-scar). - [declined] Bausch gloss "Photographs of scars on a Coulomb crystal" (MINOR, not in a sentence being fixed): loose in the same way as scars-bausch-photographed. Left as is. - [declined] Caspar-Klug gloss "Twelve pentamers on every icosahedral capsid" (MINOR, not in a sentence being fixed): twelve five-fold positions holds, twelve pentamers has exceptions (polyomaviruses). Left as is. - [fixed] tammes-proved-set, dek copy: "(2..12, 14, 24)" is now "(2..14, 24)". Covered by the same Corrected line. - [declined] "exactly as the Thomson simulation above shows them" (MINOR, not in a sentence being fixed): loose, since Bausch's scars are chains in crystals above about 360 particles. Left as is.