# Does light charge the quons? Doug's model of 12 November 2026 and its relay form, tested By Doug Westmoreland (project lead: direction and review) and Claude, Anthropic (research and analysis) · Draft 0.4, 12 November 2026 · Sections 1 to 6 checked by an independent Fable reviewer at maximum effort (sim/charge_wave/check_fable/REPORT.md); its 22 findings are folded in (change log at the end) · Section 7, Doug's relay form, added in draft 0.3 and checked by a second independent Fable reviewer at maximum effort (sim/charge_wave/check_relay_fable/REPORT.md); its 12 findings are folded in · The history in section 5 is from memory until the Qwen seat's F-61 quotes it at source. Labels: [S] standard physics; [D] derived here; [A] Aspden's own words, with the source; [M] a modelling choice; [P] proposed. ## 1. The idea Doug's model variation (12 November 2026) [P]: - Light passing a quon increases its charge, and possibly its size, but passes on at c. - The extra charge travels in the direction of the Poynting vector S, so it looks like a current along the beam. - At the quon, the extra charge makes an E-field, which makes a B-field perpendicular to S. - This sounds like the displacement current of the old aether texts. **What is right about it.** - In a light wave, B is perpendicular to S, and a moving electric field makes a magnetic field. A plane wave is exactly an electric and magnetic pattern moving at c, with B = (ẑ × E)/c for a wave moving along ẑ [S]. - One standard picture matches Doug's words exactly. The field of a charge moving at nearly c is a thin pulse with E and B both perpendicular to the motion and |B| = |E|/c, light's own ratio (Heaviside 1888–89) [S]. - The old aether physicists pictured the moving field this way (section 5). What fails is specific, and sections 2 to 4 say what. ## 2. Each step, as physics | Step | What it would have to be | Result | |---|---|---| | Light raises the quon's charge | New charge, or charge moved onto the quon | A light wave's E has no sources (∇·E = 0). At first order in the field and at long wavelengths (kd ≪ 1), it moves charge within each cell and adds none to it [D]. New charge would break charge conservation [S]. Aspden's quons carry exactly e (Part 1A §2; POC ch. 7) [A]. | | …and its size | Absorbed energy changing the radius | Aspden's rule for a charge is the Thomson formula: energy inversely proportional to radius, so a charge that takes up energy contracts (*Physics of Creation* App. VI, p. 257, said of the electron, read at source; the same rule gives the quon's radius in Part 1A §3 line 9) [A]. For the quons he goes further: they "retain their basic rest-mass" (Tutorial Note 8; Part 1 U24) [A]. And a change of size would not show outside the quon, because the field of a spherical charge does not depend on its radius [S]. | | The extra charge travels along S | A pattern passed from quon to quon, or a real flow | The pattern moves along S, but the charges move across the beam, as in a stadium wave [S]. | | So a current flows along the beam | Ampère–Maxwell: ∇×B = μ₀(J + ε₀ ∂E/∂t) | For a plane wave, the total current that sustains B lies across the beam. A current along S in a wide beam gives no B, because its displacement current cancels it. In a narrow beam it gives B circling the beam, as round a wire. For charge moving at c, that B is exactly \|E\|/c and perpendicular to S (Heaviside 1889), which is Doug's picture. But light carries no net charge (§4, test 3); a quon of mass m₀ cannot move at c; and the field of a moving line of charge is radial, not a plane polarization [S, D]. (This row answers the first wording, in which the charge itself travels along S. In Doug's relay form, section 7, nothing travels along the beam but the pattern.) | | Its E makes B ⊥ S | A changing E makes B | Correct [S], but the change of E that makes the wave's B lies across the beam, not along S. | ## 2a. What Aspden says about charge and energy Doug asked (12 November) how Aspden has charge change with energy. His texts, about 20 passages from 1972 to the 2000s, hold three rules. Except where marked as read at source, the quotes are from the site's own copies of his pages, not yet his original scans. 1. **The amount of charge never changes.** Every charge is the unit e, "given that e is a universal unit of electric charge" (Tutorial Note 17) [A]. 2. **Energy changes a charge's radius, inversely.** He uses J. J. Thomson's formula for a charged sphere: - "the Thomson formula for the electron requires energy to be inversely proportional to charge radius" (*Physics of Creation* App. VI, p. 257, read at source) [A]; - "As charge radius expands, so the energy and mass decreases and vice versa" (*Physics of Creation* ch. 2) [A]. 3. **The space occupied by charge is conserved.** - "the volume of space occupied by fundamental electric charge forms is always conserved" (essay "Cosmic mud or cosmic muddle?", of his 1972 account) [A]. - So a charge that gains energy and contracts frees space that new charges must fill: "this contraction will make space available for occupancy by electron-positron pairs created by absorbing some of that energy" (App. VI, p. 257, read at source) [A]. What follows is that **energy changes the number of charges, always in neutral pairs, never the net charge.** For the quon specifically: - "transmutations whereby energy injected into the quon can convert it into an electron plus a number of electron-positron pairs" (Tutorial Note 8) [A]. The quon holds the volume of 1843 electrons, so this "demand[s] the energy quantum of the order of that of the proton mass" (Lecture 1, appendix) [A]. - Two charges can "exchange energy by very slight volume fluctuations where one expands in radius slightly as the other contracts slightly" (*Physics of Creation* ch. 2, of gravitons) [A]. - He classes such oscillations as electrostatic, not as light: "retardation and speed-of-light propagation are phenomena we associate with electromagnetic energy transfer, rather than electrostatic action" ("The Fifth Question") [A]. **What this means for the model, by Aspden's own rules [D]:** - A quon taking up light's energy would contract, not grow, and its net charge would not change. - The space it frees would go to pairs. But a real electron–positron pair costs at least 1.022 MeV, and turning a quon over costs about a GeV, while visible light carries a few eV. - A change of size makes no field outside the charge (Gauss), so breathing charges cannot make light's E or B. Aspden himself calls their action electrostatic. - What his framework does allow is charges trading energy through linked changes of size. Passed from quon to quon, that runs into the lattice limits of section 4, test 1, and it still makes no outside field. ## 3. The three ways to read "the quon's charge increases" In the record's model (Part 1A §2: point quons q = e, inertia m₀, simple cubic lattice of spacing d, in a rigid neutralising continuum [M]): - **Q1, the cell's net charge changes.** That happens only where the quons' displacement has a divergence, δρ = −nq ∇·u [D]: the longitudinal wave, whose displacement lies along the wave vector. This is the reading in which charge travels along the direction of travel (which would be S if the wave had a B). Trial 1 computes it, below. - **Q2, the quon's own charge grows while the continuum around it gives up the same charge.** The cell stays neutral and keeps no dipole moment, so it has no field at distances beyond its own size. The radial charging current is cancelled by its own displacement current, so it makes no B either. It cannot make the wave's fields [D]. - **Q3, the continuum gives up the charge on one side.** A dipole across the beam is the transverse polarization of Aspden's dual displacement, already tested in Part 1B. Carried by the quon lattice it is a plasma, which blocks light below 511 keV with both frames moving (Part 1 U34). A dipole along the beam is Q1. **Trial 1** (lock: papers/Lock_charge_wave_2026-11-12.md, committed at 00:43 EST before the run; the script was committed with its output at 00:44:57, after the run, so strictly only the lock was pre-registered; its specification is complete, and the independent reviewer's rerun matched byte for byte). The longitudinal branch of the record's unchanged Ewald dynamical matrix, along three symmetry directions: | | [100] | [110] | [111] | |---|---|---|---| | Band of ħω (keV) | 361.3 to 379.7 | 267.1 to 361.3 | 208.6 to 361.3 | | Fastest group velocity | 0.030c | 0.111c | 0.147c | - Along the three directions computed, the charge-carrying branch spans 209 to 380 keV, and its group velocity (the speed of its energy and of any signal) is at most 0.147c. - Over the whole zone (checked after the run by the independent reviewer, so post hoc), the band is the same, with its edges at R and X. The group velocity reaches about 0.20c away from the symmetry directions, where the branch is mixed (|ê·k̂| down to 0.71) and the energy moves at up to 63° from k. - Its phase speed exceeds c for kd < 1.2, but a phase speed carries no energy or signal; P1 is about the group velocity, as the lock says. - It carries no magnetic field: for a longitudinal wave ∇×E = 0, so ∂B/∂t = 0 [S, D]. - It fails all three criteria fixed in the lock: P1 (travel at c at every energy), P2 (B ⊥ S with |B| = |E|/c) and P3 (two transverse polarizations). Notes on the trial: - **Two expectations in the lock were wrong.** The band reaches down to 209 keV along [111], not only 361 to 380 keV; and the fastest speed along the three directions is 0.147c, not "a few per cent". Neither changes a criterion. The record already had the numbers: Part 1's U28 output gives the top branch at 0.546 at M and 1/3 at R, so the expectation was an error of reading, not a surprise of the physics. - **The band starts at 361 keV,** the quons' frequency with the continuum held still. 511 keV is the frequency with both frames moving (Part 1 U34). - **At M and R** (post hoc remark): at M the longitudinal branch is degenerate with one transverse branch (both at λ = 0.546), and at R with both (all three at λ = 1/3, by the trace sum rule). So the eigenvector solver returns an arbitrary mixture there, and |ê·k̂| is not meaningful at those two points. The eigenvalues, and the band edges, are unaffected. - **Derivatives:** centred differences inside the grid, one-sided at its two ends, so the value printed at t = 0 is not the limit, which is 0. - **The model's limits:** the matrix is electrostatic, with instantaneous Coulomb forces and no retardation. That does not touch the longitudinal branch's B = 0, which holds with retardation too (E ∥ k). The lattice it describes is formally unstable (U28: 18.4% of modes), so "the branch" is a branch of an unstable lattice; the record knows this, and the conclusion does not depend on it. ## 4. Tests any version would face 1. **If the quons carry the light from site to site.** - A lattice of spacing d carries no wave shorter than 2d = 1.27 × 10⁻¹² m along a cube axis, or 2d/√3 along a cube diagonal: photons above 0.97 MeV along the axis and above 1.68 MeV along the diagonal [D]. - A lattice at rest would pass cobalt-60's 1.33 MeV gamma rays only within cones about its cube diagonals. They cross every laboratory in every direction. - Well below the cutoff, the lattice also disperses what it does carry: the Crab pulsar's X-ray timing limits the spacing of any carrier lattice to 2.4 × 10⁻¹⁹ m (Part 1 U16). - And 6 GeV photons were timed at c (Alväger et al. 1964, Part 1 U13), and 31 GeV photons, 3 × 10⁴ times shorter than 2d, arrived with the keV light of GRB 090510 (Part 1 U14). 2. **If the light passes through on its own and the quons only respond.** Their response adds to the vacuum's, and changes light's speed with frequency. The record bounds that share below about 10⁻⁶ of ε₀ at optical frequencies and 10⁻¹³ at X-ray (handover summary of 4 November, §2, under its assumptions A1 to A5 on the Crab pulsar's emission). So, under those assumptions, the quons cannot be the displacement current itself: that would be a share of 1. 3. **If the extra charge were real and not compensated nearby,** light would carry charge, and the limits on a photon's charge are below about 10⁻³⁵ e (Particle Data Group; from memory, to be read at source). If the continuum gives the charge up nearby, the pulse is neutral and this test says nothing; that is reading Q2 or Q3. **A real effect that looks like the picture.** An intense laser pulse in a plasma pushes electrons out of its path, forward at its front and back at its tail, and a wave of charge follows the pulse at its group velocity, near c along S: the plasma wakefield (Tajima and Dawson, Phys. Rev. Lett. 43, 267, 1979). But it grows with the square of the field, its fields lie along and around the beam, and light itself adds linearly. It is something light does to a medium, not how light travels. ## 5. The history (from memory until F-61) - **Maxwell, "On Physical Lines of Force", Part III (Phil. Mag., January 1862).** Displacement was an elastic shift of electricity within the molecules of the medium, along the electromotive force, so across the ray. From the medium's elasticity he found a speed for transverse waves equal to the measured ratio of units, and concluded that light "consists in the transverse undulations of the same medium which is the cause of electric and magnetic phenomena". - **Maxwell 1865 and the Treatise (1873).** He dropped the mechanism but kept the displacement. In light, the displacement and the magnetic disturbance both lie across the ray. - **J. J. Thomson (1881) and FitzGerald (1881).** The first calculation of the magnetic field of a moving charge (Thomson, with a wrong factor that FitzGerald corrected). - **Poynting (1884), and Heaviside independently (1884–85).** Energy flows along E × H. This came after Maxwell's death in 1879. - **Heaviside (1888–89).** The complete fields of a moving charge: a moving electric field carries a magnetic field, B = v × E/c². - **J. J. Thomson, *Notes on Recent Researches in Electricity and Magnetism* (1893), ch. 1.** He took the displacement current to be the motion of "tubes" of electric force, and light to be such tubes moving at c along the ray, perpendicular to themselves. This is closest to Doug's picture, with tubes of force where he has charge. - **Lorentz (1892, 1895).** The aether's displacement carries no charge; only the charges of matter are displaced. ## 6. What survives, and what is open - **What survives.** What travels along the Poynting vector is a pattern of displacement (Thomson's tubes), moving at c, while in each cell the charge shifts across the beam. That is Maxwell's displacement, and in Aspden's terms his dual displacement. - **Where it can be carried.** - Carried by the quon lattice, it fails (Part 1B, and sections 3 and 4 above). - Carried by a continuum whose free-field equations are Maxwell's (level 1 of the consistent model of 5 November, a postulate), the surviving part of the picture is Maxwell's own wave: a displacement pattern moving at c with B across it. Nothing of the charging of quons is needed or allowed there, since the quons' share of the response is bounded below about 10⁻¹⁴ of ε₀ (M2). What makes the quons transparent to that wave is still open (T4). - **The relay form (section 7).** Read as a separation of charge across the beam, handed from quon to quon, Doug's model is Maxwell's two curl laws applied quon by quon. - Its free waves travel at c with E and B across the beam, if the quons' stiffness against separation supplies ε₀ (μ₀ being the strength of the magnetic action between currents). - A quon that answers at once is its best case. - It is not yet a theory. That stiffness is a new property of the quon, and on the record's numbers the natural choice gives √2 c. The same ε₀ must set Coulomb's law. The carrying level must be finer than 2.3 × 10⁻²⁶ m. And the coarser levels, Aspden's included, must stay out of light's way. - **Open.** - A continuum that compresses. It would move Q1's band toward 511 keV, but would not give it a B-field or a transverse polarization. - Second-order charging, as in a wakefield, bounded by limits on light-by-light scattering in vacuum (not computed). - The history, read at source (F-61). - The same direction-dependent cutoff qualifier belongs in Part 1 U16 step 1 and Part 1A §8, which say "no wave shorter than 2d" (for a later revision, with Doug's approval). ## 7. The relay form (Doug, 12 November 2026, morning) Doug's clarification, after reading draft 0.2 [P]: - The quon does not move, and no real current flows. - The quon is a particle made of charge. Energy arriving makes it grow, perhaps at once, and raises its charge for a moment. - That local rise in charge makes an E-field and a B-field, which reach the next quon and hand the energy on. The first quon returns to normal. The next has its own brief rise, lasting about the time light takes to cross between quons, and passes it on, and so on. - E and B both lie across the direction of travel, which is the direction the energy goes. - To work, this needs several levels of quons. These are the nested lattices of 1 November (U44, a draft register entry of 1 November 2026, papers/Register_draft_nesting_leptons.md, not yet in the register): smaller copies of Aspden's lattice inside it, 3 × 3 × 3 cubes within 3 × 3 × 3 cubes, with level k having spacing d/sᵏ and energy sᵏ × 511 keV. U44 found them no use as a ladder of particle masses and ended: "A nested lattice, if it exists, can only be a ladder for light." This removes the objection in the fourth row of section 2, which answered the first wording (charge travelling along S). In the relay nothing travels along the beam except the pattern and its energy. That is how every wave travels. ### 7.1 Each step, as physics | Step | What it has to be | Result | |---|---|---| | The quon does not move; no current along the beam | Energy handed on, not carried | Right for every wave [S]. | | Its charge rises for a moment | (a) a net rise; or (b) a separation: one side of the quon up by δq, the other down by δq | (a) fails three ways. Charge cannot change unless a current brings it (charge conservation, ∂ρ/∂t + ∇·J = 0) [S]. Aspden's charge is always e (§2a) [A]. And the field of a charge that grows evenly points straight out from it: at the next quon along the beam it points along the beam, and its change makes no B at all [S]. (b) works. The net charge stays e, and the change is a dipole across the beam [D]. | | …and it grows | A swelling, or a stretching across the beam | A swelling changes nothing outside (§2) [S]. A stretching across the beam is (b) again. | | Its E and B reach the next quon | The fields of a dipole across the beam, seen along the beam | E lies along the separation and B across both: both perpendicular to the travel, as in light [S]. | | The energy goes on, not back | A rule for the direction | The energy flows along E × B (the Poynting vector). A pulse started with E and B in that relation travels one way only [S]. | | The first quon returns to normal | Each quon's separation rises and falls as the pulse passes | As in any wave [S]. | | Each step takes the time light takes | Step time d/c, the same at every quon, every level and every frequency | The time comes from the coupling between the quons, not from the quon. It is set by the stiffness of the separation and the magnetic inertia of the current that the separation makes (§7.2, items 4 and 5). The quon itself may respond at once, as Doug said. The values that give exactly c are the open part (§7.3, item 2). | Reading (b) has one consequence. A separation that changes is charge moving across the quon, so there is a current. It runs across the beam, back and forth inside each quon, never along the beam. It is Maxwell's displacement current, and it is what makes B. Outside the quon, a shifted charge looks the same as a moved quon. The difference is only the quon's own inertia, which sets the highest frequency the relay can carry (§7.2, item 5). ### 7.2 What works 1. **Nothing travels along the beam but the pattern.** Maxwell's 1862 picture of light was the same: cells of the medium storing energy and handing it on (§5, from memory until F-61). 2. **Light's geometry.** With the separation, E and B at the next quon are both across the beam and across each other, and the energy goes along E × B [S]. 3. **Two polarizations.** The separation can lie along either of the two directions across the beam. A separation that turns gives circular polarization [S]. 4. **The speed c, in principle.** Suppose each cell stores two kinds of energy [S]: - electric energy in its separation, like a capacitor (capacitance ε₀d); - magnetic energy in the field of the current that its changing separation makes, like an inductor (inductance μ₀d). It then hands a pulse on in a time √(LC) = d√(ε₀μ₀) = d/c, whatever its size. - So such a relay travels at c, its long waves obey Maxwell's wave equation, and every level built this way has the same speed. - A capacitor takes its charge at once; the delay comes from the inductance. So Doug's "growing at once" is what the relay needs. The time light takes is set by the coupling between the quons, not by the quon. - The hand-on is Doug's chain. A changing separation, which is a current across the beam, makes B (Ampère–Maxwell). A changing B makes E at the next quon (Faraday), and that E separates the next quon's charge. - On a cubic lattice this relay is the method computers use to solve Maxwell's equations, with E at the cells' edges and B at their faces, each updated from its neighbours (K. S. Yee, *IEEE Trans. Antennas Propag.* 14(3), 302–307, 1966, doi:10.1109/TAP.1966.1138693; checked by the reviewer). 5. **It works with instantaneous forces** [D, under the assumptions marked]. Aspden held electrostatic action to be instantaneous, and retardation and speed-of-light propagation to belong to "electromagnetic energy transfer" ("The Fifth Question", §2a) [A]. A relay that uses only instantaneous forces still gives a wave of finite speed. - **The model.** Each quon's charge separates by a distance x across the beam, against a stiffness K (which includes the pull of its neighbours) and with an inertia m. The stiffness is a new property of the quon [P]. Aspden's quons have none: nothing pulls a sideways shift of the quons back at long wavelengths (Part 1A, W3 line 6). Without it they are the plasma of section 3 and U34, and here they would carry no sideways wave at all. Between the quons there is no vacuum of its own; the quons are the whole medium [M]. - **The steps,** for a wave of wavenumber k, with n quons per unit volume: 1. the separations make a polarization P = nex and a current J = ∂P/∂t across the beam; 2. the current's magnetic field, taken as instantaneous (Ampère's law without retardation), has the vector potential A = μ₀J/k²; 3. as A changes, it pushes on every quon with the field E = −∂A/∂t (self-induction); 4. each quon obeys mẍ = −Kx + eE; 5. together these give ω² = Kk²/(μ₀ne² + mk²) = c_q²k²/(1 + c_q²k²/ω_r²), with ω_r² = K/m and c_q² = 1/(μ₀ε_q). Here ε_q = ne²/K is the dipole moment (charge times separation) the quons make per unit volume, per unit field. - **What it gives.** - At long wavelengths the speed is 1/√(μ₀ε_q), which is c when ε_q = ε₀. - E and B lie across the beam, |B| = |E|/c_q, and the electric and magnetic energies are equal, as in light. - The quon's own inertia matters only at short wavelengths: the wave slows near ω_r, and nothing travels above it. With m → 0, a quon that separates at once, the relay has no dispersion at any wavelength. - **Two limits.** - If Maxwell's vacuum were kept as well, and these quons added to it with ε_q = ε₀, light would travel at c/√2. The relay replaces the vacuum; it does not sit in one. That is why the "share" of section 4, test 2, and "the quons supply ε₀" here are different pictures. - Forces that act at once let a disturbance be felt at once, out to about c_q/ω_r: about 0.6 of the spacing if ω_r is the level's rhythm. Maxwell's vacuum has no such reach. Here it is far below anything measurable, but not zero. - **Aspden's two halves.** The relay realises Aspden's electrostatic half: the forces act at once while the energy moves at a finite speed. It contradicts his magnetic half. - Half of the relay's energy is in the magnetic field of the current. Aspden wrote that the wave's energy is the quons' kinetic energy, and that "there is no such thing as a magnetic field out there in free space" (Research Note 4/98; Part 1B, section 3). - In the relay, the quons' kinetic energy is only (ω/ω_r)² of the electric energy. Part 1B §8.2 found that his kinetic version has the wrong size and the wrong timing. 6. **The levels share one speed, but on the record's numbers it is √2 c, not c** [D; the reviewer's finding F2]. Under the nesting rule (U44: quon mass m₀sᵏ, rhythm Ωsᵏ, spacing d/sᵏ) the circling speed is c/2 at every level. - If a quon's stiffness against separation is its mass times its rhythm squared, K_k = m_kΩ_k², then ε_q is the same at every level (relay_requirements.py, item 6). - Its value follows from the record: ε_q = ε₀(ω_p/Ω)², with ħω_p = 361.331 keV for the quons alone and ħΩ = 511 keV (Part 1A W4). So ε_q = ε₀/2, and the relay runs at √2 c on every level, or 1.29c counting the neighbours' pull with Coulomb's constant ε₀. - Part 1B §8.2's non-circular test, that a model's own wave speed divided by 2Ωr must come out as 1, gives √2. - Getting c needs K_k = m_kω_p,k², half that stiffness. The continuum's pull on a displaced quon has that size, but it acts only along the wave, not across it (Part 1A W3). So the relay's stiffness must be a new property of the quon, with its size still to be derived (the reviewer's check_relay.py, section G). ### 7.3 What still needs work 1. **"Its charge rises" must become "its charge separates".** A net rise needs a current flowing in, which the model excludes. It would break Aspden's universal e, and its field points along the beam with no B. With the separation the net charge stays e, which is closer to Aspden. 2. **The two numbers that set the speed are not derived.** The quon's own response may be as fast as you like, and it must be prompt (item 4). What makes the step take d/c is two things: - **the stiffness of the separation, which gives ε₀.** It is a new property of the quon (§7.2, item 5). Its size is open: under the nesting rule, on the record's numbers, the natural choice gives √2 c, not c (§7.2, item 6). - **the magnetic inertia of the current it makes, which gives μ₀.** μ₀ enters through Ampère's law, as the measured strength of the magnetic action between currents, not as a property of the quon. The 5 November model has the same gap: level 1's stiffness and inertia are "identified, not derived" (Consistent model §0). - **The two pictures are duals.** In the relay the stiffness gives ε₀ and the inertia gives μ₀, as in a capacitor and an inductor. In level 1, written as MacCullagh's medium, the stiffness is 1/μ₀ and the inertia ε₀. For free waves the equations are the same. - **Part 1B asks for this.** Its §8.2 asks for "a restoring force that does not presuppose a bare vacuum term". The relay is one form of it, with the restoring force postulated. 3. **One ε₀ for both jobs.** In Maxwell's theory the same ε₀ sets Coulomb's force between charges and light's speed. Maxwell's 1862 discovery turned on it: a speed computed from electrical measurements matched light's (§5, from memory). In the relay, the ε₀ of the hand-on is the quons' separation. So the force between charges as we measure it must also be set by that separation [D, P]: - The bare force between charges must be far stronger than the measured one, and the quons' separation must screen it down to the measured value. On a cubic lattice the same separation answers a steady field and a long light wave, so one ε_q does both jobs. - If the screening is incomplete, light's speed and the speed computed from electrical measurements differ by half the unscreened fraction. - This is the classical counterpart of charge screening by the vacuum in quantum electrodynamics. The balance is a fine one. The neighbours' fields help each separation along, and if they help too much the lattice polarizes by itself (the instability of the Clausius–Mossotti relation). To screen that strongly, the quon's own stiffness must exceed that point by only about three times the unscreened fraction. 4. **The carrying level must be very fine and very prompt** [D; relay_requirements.py, from the record's U15 and U16]: - **A relay slows short waves.** With hand-on between nearest neighbours, a relay of spacing d slows them by (kd)²/8 of c along a cube axis and (kd)²/24 along a cube diagonal (group speed). So the speed also depends on direction, a lattice signature. Longer-range coupling changes these numbers but not the ceiling (U16). Away from the cube's symmetry directions it can also make the two polarizations travel at slightly different speeds. - **It has a ceiling.** It carries nothing above 2ħc/d along a cube axis (0.62 MeV for Aspden's spacing; U16 step 2), or 2√3 ħc/d along a diagonal (1.07 MeV). - **GRB 090510.** Its 31 GeV photon, taken at 28 GeV (the low end of its error, as Abdo et al. take it), arrived with the keV–MeV light: |Δt| < 0.859 s. This is limit (a) of Abdo et al.'s Table 2 (arXiv:0908.1832, read at source on 12 November). With the (1 + z)² weighting a quadratic delay needs, as in U15, it puts the carrying level's spacing below 2.3 × 10⁻²⁶ m along a cube axis, or 3.9 × 10⁻²⁶ m along a diagonal. Aspden's spacing, 6.37 × 10⁻¹³ m, is 2.8 × 10¹³ times too coarse. - **Promptness.** In the version with instantaneous forces (§7.2, item 5), the quon's internal frequency ω_r must lie above 3.0 × 10¹⁹ eV, U15's bound for any response of that form. - **The two add.** The lattice's slowing and the quon's add. Bounding each alone, as here, is lenient by at most √2. 5. **The levels it needs** (U44's rule) [D]: | Factor s | First level fine and prompt enough | Its spacing | Its energy | |---|---|---|---| | 2 | 46 | 9.1 × 10⁻²⁷ m | 3.6 × 10¹⁹ eV | | 3 | 29 | 9.3 × 10⁻²⁷ m | 3.5 × 10¹⁹ eV | | 4 | 23 | 9.1 × 10⁻²⁷ m | 3.6 × 10¹⁹ eV | For s = 3, light would ride 29 levels below Aspden's lattice. The record found nearly the same on 2 November (Filter_check §8, "Light on the ladder"): - A photon carried by the level of its own energy sits at kd = 1.65 on every level. There the simplest lattice wave is 32% slower (group) than its long-wave speed, so "each level serving a higher energy" does not work for light. - The 31 GeV photon needs level 28 or finer for s = 3. Only the ladder's continuum limit works, which is the record's route R1. - Here the cube-axis coefficient with 28 GeV still gives 28. The (1 + z)² weighting makes it 29, and the promptness condition gives 29 on its own. 6. **Only the finest levels can carry light; the coarser ones must stay out of its way** [D]: - **A responding level changes light's speed at its own energy** [M]. Suppose a level's quons respond up to their own energy and not above. The level then holds a share f of light's ε₀ below that energy and none above, so light above it is faster by f/2. (A resonant response instead of this step changes the bounds below by factors of 0.78 to 1.41, the reviewer's check.) - **GRB 090510 bounds that share** (Abdo et al., arXiv:0908.1832, Table 2). A constant difference in speed builds up over the comoving path, 10.3 billion light-years. Both limits used here rest on matching single photons to spikes in the low-energy light curve, which the authors rank as their least conservative: - **Limit (f).** One photon of 0.75 GeV (0.694 GeV at the low end of its error) arrived no more than 19 ms ahead of the 0.1 MeV light. That is the sign that applies here. So f < 1.2 × 10⁻¹⁹ for any level between about 0.2 MeV and 0.7 GeV: levels 0 to 6 for s = 3. - **Limit (e).** The 31 GeV photon arrived within 10 ms of the 0.1 MeV light, either sign. So f < 6.2 × 10⁻²⁰ for any level between about 0.2 MeV and 28 GeV: levels 0 to 9 for s = 3. - **But under the nesting rule every level would take the same share** (§7.2, item 6). So one of two things must hold: - **The coarser levels do not respond to light at all.** This includes Aspden's lattice. It is the 5 November model's open law T4, why matter's charges feel light and the quons do not, and the relay does not solve it. - **Or each level responds in the balanced way the vacuum does in quantum electrodynamics** [P]. There, virtual pairs screen charge but leave freely travelling light at c (Part 1, section 10, item 1). Light's speed is then unchanged; what changes slowly with energy is the strength of charge, as α does, from 1/137.036 at low energy to about 1/128.95 at 91 GeV. That test is in effect already made [D, the reviewer's estimate; the numbers are from memory, to be read in the Particle Data Group's review]: - six levels of s = 3 between 1.5 and 373 MeV, each screening like a particle of unit charge, would change 1/α at 91 GeV by about −10.5; - that is more than the whole measured change, −8.1, which the known leptons and quarks already account for to about ±0.02; - so such levels could screen at no more than about 2 × 10⁻³ of a unit charge's strength. The quons are not particles of that kind, so this is a guide, not a bound. 7. **Aspden's own lattice is still the problem.** The record bounds its response to light below 1.5 × 10⁻¹⁴ of ε₀ (R1_checks, for a stiff orbit resonating above 3 TeV). Under item 6's assumption, that it stops responding above its own 511 keV, the bound is 6 × 10⁻²⁰ to 1.2 × 10⁻¹⁹. A relay of finer quons does not say why Aspden's quons, which are charges, would not respond while matter's charges do. 8. **The photon.** The relay is a classical wave. It does not say why a photon's energy hν arrives whole at one detector; neither does Maxwell's theory. Aspden's photon unit is tested in Part 1 (U27–U31). ### 7.4 Where this leaves the idea - **What it is.** Read as a separation of charge across the beam, Doug's relay is Maxwell's two curl laws applied quon by quon. Its free waves at long wavelengths obey Maxwell's wave equation, so they pass the tests of speed, polarization and dispersion that Maxwell's vacuum passes. - **Not yet shown:** - that matter bound by these fields contracts and keeps its clocks as matter does in Maxwell's vacuum (the 5 November model's M5); - the Coulomb ε₀ (§7.3, item 3); - strict causality (§7.2, item 5); - T4. - **A picture of E and B.** It gives one: the separated charge in each cell, and the magnetic field of its current. - **How it fits the record.** On the finest levels it is the record's route R1 (light carried by level 1, the lattice out of its way), with a mechanism proposed for level 1. It is also one form of Part 1B's P1, the aether as the source of ε₀ and μ₀. - **A correction to draft 0.3, in Doug's favour.** Draft 0.3 said a quon that "grows at once" could not work. It is the relay's best case. The delay that makes the speed c comes from the magnetic coupling between the quons, and a quon that answers at once makes the relay free of dispersion. - **What it is not yet.** It is not yet a theory. It needs: - a stiffness of the quon against separation, a new property, of the size that gives c (on the record's numbers the natural choice gives √2 c); - the same ε₀ in Coulomb's law; - a reason the coarser levels, Aspden's included, stay out of light's way. - **Its own signature** is the lattice: a speed that falls with energy and depends on direction, at order (kd)², with the spacing now pushed below 2.3 × 10⁻²⁶ m. - **No trial was run for this section.** It applies the record's bounds to the relay (sim/charge_wave/relay_requirements.py and its output, relay_requirements_out.txt). The derivations in §7.2, items 5 and 6, and §7.3, items 3 and 6, are new and labelled [D]. All of it was checked by a second independent reviewer (sim/charge_wave/check_relay_fable/: REPORT.md, check_relay.py and its output, and an excerpt of Abdo et al.'s Table 2). ## Change log - **0.1 (12 November 2026, about 00:45 EST):** first draft. - **0.2 (12 November 2026):** the independent reviewer's findings folded in: - F1: the group velocity, 0.147c along the three directions and 0.20c over the whole zone, post hoc; the phase speed named. - F2: the degeneracies at M and R. - F3: the direction-dependent cutoff and the dispersion bound. - F4, F5: qualifiers. - F6, F7: the App. VI citation, p. 257 only and said of the electron, and Tutorial Note 8. - F8, F9: which reading each test bounds, and credit to the moving-charge picture. - F10 to F21: wording, labels, the trial's notes, history additions, the wakefield. - F22: the tracker. - Section 2a added, on what Aspden says about charge and energy, at Doug's request. - **0.3 (12 November 2026, from 05:55 EST):** section 7, Doug's relay form (his clarification after reading 0.2), evaluated: each step as physics, what works, what still needs work, and the levels it needs (sim/charge_wave/relay_requirements.py); pointers in section 2 and section 6; the title. Not yet independently checked. - **0.4 (12 November 2026, from 06:30 EST):** the second reviewer's 12 findings folded into section 7 (sim/charge_wave/check_relay_fable/REPORT.md): - F1 (major): a quon that answers at once is the relay's best case, not a fault. The delay comes from the magnetic coupling, and μ₀ enters through Ampère's law; draft 0.3 had this backwards. - F2 (major): under the nesting rule with K_k = m_kΩ_k², the relay runs at √2 c on the record's numbers. - F3: the running of α already tests the balanced option. - F4: limits (e) and (f) read at source (arXiv:0908.1832, saved in sources/fable_fetches/abdo2009/), with their caveats; the step labelled [M]. - F5: the derivation's steps; no vacuum of its own; the stiffness a new property; the small reach of instantaneous forces. - F6: Aspden's magnetic half. - F7: the overclaim in 7.4. - F8: half the unscreened fraction; the Clausius–Mossotti balance. - F9: qualifiers; the arXiv Table 2 citation; Yee verified. - F10: 28 against 29 levels. - F11: item 7 against item 6. - F12: the duality with level 1. - Also: the header and the section 6 bullet.