Science Contrarian review: Bryan G. Wallace, "The Farce of Physics" (1993) and the c+v radar claim

Reviewer: Science Contrarian (Step 2). Date: 2026-10-04. Standard applied: evidence over consensus, the same rigour for the inventor as for his critics. Tags: [read] from the project documents, [read: page image] from a page I rendered, [web] with URL, [memory] with confidence.

0. Source gap and what I read

• Project folder holds one document, the 1993 manuscript farce.txt. The equations of "mass dynamics" and the atomic models are not in it [read: deep/farce-d10910.md section 0]. Claims 3, 4 and 5 therefore rest on the book's prose and the deep read, not on Wallace's papers (Found. Phys. 3, 381 (1973); J. Clas. Phys. 1(2), 17 (1982); Speculations Sci. Technol. 9, 9 (1986)), none of which I could obtain.
• New in this step: I downloaded and read (page images) Wallace's primary radar paper, "Radar Testing of the Relative Velocity of Light in Space", Spectroscopy Letters 2, 361-367 (1969), from http://sciliterature.50webs.com/SpecLetters1969-p361-367.pdf [web; read: page images pp.361-367]. The Investigator had only located it. This closes the claims.md gap for the 1961 fit. His 1971 paper (Spectroscopy Letters 4, 79) at gsjournal.net I did not read, so the "3 ms vs 10 microseconds" figure (from the 1993 book, ch.4 [read]) rests on the book alone.
• I read book ch.4 (Physics Today letters, PRL story) and ch.5 (Shapiro correspondence), the Ch.7 deep read, and all Investigator notes.

Corrections to Step 1

1. claims.md item 2 says "Shapiro's side is not given". The book does give it: Shapiro's letter of 17 Jan 1969 says the radar data "are not consistent with the 'ballistic theory' of light, but are consistent with general relativity", citing the Oct 1966 PRL second-order Doppler paper [read: farce lines ~3535-3600]. Wallace's reply: "the article does not support your argument and you know it". The Investigator's lead was right and should be folded into claim 2.
2. Claim 1 says residuals were analysed "from 1961-1966". The only quantitative fit in Wallace's own paper uses eight published 1961 observations compared with Newcomb's ephemeris (Fig. 1, p.366). The 1962-66 material is argued in prose from the Lincoln Lab papers (goodness-of-fit 1.57, 0.97, 7.10; a quotation about the USSR residuals) [read: p.362]. Claim 1 should be split: (a) the 1961 eight-point fit; (b) the argument from later published residuals.
3. claims.md calls the "3 ms vs 10 us" a residual from the c analysis. In the book the reference is "the best fit the Lincoln Lab could generate" [read: ch.4]. In the 1969 paper the comparison is instead against Newcomb's tables (which Wallace calls "N"), not a Lincoln fit [read: p.365-366]. These are two different yardsticks; the book's wording blurs them.
4. Hafele-Keating is cited in claim 3 context; the Investigator's "claim 11" label refers to a numbering that does not exist in claims.md (only 6 claims). Bears on claim 5 / absolute-frame ideas only.
5. Deep reader item M12 (frame of the internal speed c) and the "104 vs 10^4" pion ambiguity stand. The extraction dropped superscripts, so any numeric statement in the extracted text must be checked against the original (already flagged by the deep reader).

Claim 1: radar residuals show c+v

a. Is the mathematics internally consistent?

Wallace's two equations [read: page image p.364]: c model D_E = tc/2 - tv/2 (eq. 1); c+v model D_G = t(c+v)/2 - tv/2 = tc/2 (eq. 2), with t the round-trip time and v the relative radial velocity from the Doppler shift (v = dc/2f, his notation). Reasoning behind eq. 2: outbound speed c, returning speed c+2v after reflection from a moving "mirror" in the emission picture, so average speed c+v.

My own check: - In the standard c treatment, the average of the two legs is already the distance at the moment of reflection to second order, so subtracting a first-order tv/2 is a statement about which epoch the distance refers to (the book says "the time the beam left the transmitter"). Wallace's eq. (1) and eq. (2) therefore differ by exactly tv/2, a first-order quantity: about 128 km for t = 276 s and a relative speed of 0.93 km/s (two stations on opposite sides of the Earth), matching his 260 km "difference between two stations" (his number is t times the speed difference; [read: p.362]) up to a factor that depends on whether tv/2 or tv is meant. The size is internally consistent. - The risk I see is bookkeeping: the two curves E and G are plotted against the same epoch labels (Julian date, Fig. 1). If one expression is the distance at transmission and the other at reflection, a shift of v*t/2 appears with no physics behind it. I cannot tell from the paper whether he applied the same epoch convention to both; the paper is silent. This is a question for the Mathematician, not a finding. - A genuine ballistic round trip gives a first-order delay change (D/c)(v_target - v_source)/c. With D = 41.4 million km and a relative speed of 10 km/s this is 4.6 ms; for Earth rotation alone 0.2-0.4 ms [my calculation]. The size of Wallace's "3 ms" is thus the natural scale of a first-order ballistic effect. That is the strongest honest observation in his favour: the effect he reports has the right order of magnitude for his hypothesis. The same fact makes it testable, see (c). - Equation (2) uses the mirror-doubling rule (c+2v). That is a specific ballistic assumption (reflected light travels at c relative to the reflector). A pure emission theory is ambiguous about reflection (Fox's extinction objection, which Wallace cites as saying nothing disproves the remaining theories [read: p.361]).

b. Is it consistent with established experiments?

For the ballistic hypothesis itself, no direct support remains in the mainstream record I know: - Alvager et al. (CERN, 1964): photons from 6 GeV neutral-pion decays, source speed above 0.99975c, measured at 2.9977 +/- 0.0004 x 10^8 m/s, bounding any source-velocity dependence to about 10^-4 [web: https://arxiv.org/pdf/gr-qc/0504085 and search summary of the experiment; [memory, high] for the paper Phys. Lett. 12, 260]. - Binary stars (de Sitter 1913; Brecher 1977 on X-ray binaries) [memory, medium] constrain emission theories; Wallace's own 1970 paper claimed to explain spectroscopic binaries with c+v [read: deep, section 1]. Critics (Fox 1965, cited by Wallace) concede extinction makes some of these tests ambiguous [memory, medium]; I did not verify the 1977 Brecher numbers. - Planetary and spacecraft ranging after 1969, with several stations at different times and places, reached metre-level accuracy under the c-based (GR) light-time model [memory, medium]. If a c+v first-order term of hundreds of km existed, it would show as station-dependent, diurnal residuals. Wallace's answer is that Lincoln Lab's later fits avoided it by using a constant observing time (12:00 UT) and empirical corrections [read: ch.5, letter of 3 April]. This is a reasonable thing to ask for, but spacecraft ranging from three DSN stations is the data that would settle it, and I have no source in hand: I did not verify the claim, so the evidence against rests on [memory].

The decisive point against: Wallace's statement of the c+v prediction is also a statement of what the dedicated test should find, and he did not have the data to run it. His own conclusion says so: "I don't think it can be considered reasonably conclusive until a full c+v investigation is made" [read: p.367].

c. Has it been tested? By whom? Replicated?

Classification: partly never properly tested by neutral parties, partly dismissed without a stated analysis, partly answered. - Wallace's own test: eight 1961 points, one analyst, compared with Newcomb tables [read: p.364-366]. Not replicated by anyone I can find. ADS lists 12 citing papers; a few are c+v-friendly (Rapier, SpecL 4, 303 (1971), VLBI of 3C 279; Wegener and Ohrstrom 1975) and one is a rebuttal (Horedt and Ruck, SpecL 5, 455 (1972)) [web: Investigator, ADS/tandfonline]. Their text I could not read. - Shapiro's response, as relayed by Wallace (an interested correspondent, quoted verbatim in the book, no copy of Shapiro's letters available): the data "are not consistent with the 'ballistic theory'" and the PRL Doppler article supports this [read: farce ch.5]. Wallace says the article "does not support your argument". I could not read the PRL paper (paywalled: https://link.aps.org/doi/10.1103/PhysRevLett.17.933). This is a published rebuttal from the data-holder, though one-sided and very short in the telling. - Data access: Shapiro stated the data "did not exist in the form" Wallace asked for; Hollingsworth (AFCRL) supplied an IBM card deck that Wallace judged useless since the stations were close together and observing time constant [read: ch.5]. Hollingsworth reportedly said, in person (Wallace's account), that the individual AU values "ranged over thousands of kilometers" with variations related to relative velocity, and that the wanted data existed but could not be released without Shapiro's permission [read: ch.5]. This is one source (Wallace's own recollection of a hotel-room meeting, with no independent confirmation) and Hollingsworth's letter said it was impossible to persuade Shapiro to "prove the velocity of light can be measured as anything else than c" [read: ch.5]. I weigh it as testimony, not proof. - The key Lincoln Lab quote that Wallace uses is real in the sense that it sits in a peer-reviewed paper: USSR residuals "systematically negative relative to the Arecibo and Lincoln Laboratory residuals... when all three groups were observing Venus. This incompatibility cannot be removed by assuming simply that different units of time were used" and "discrepancy of up to five times the quoted measurement error" [read: p.362, quoting Ash, Shapiro and Smith AJ 72, 338 (1967)]. A discrepancy between three stations is evidence of something unmodelled; it is not specific to c+v, and clock/timing differences and propagation (plasma) are alternatives the authors themselves considered. Wallace himself concedes that the c theory predicts small plasma effects and that his own plasma argument is based on a 440 vs 1295 Mc/s amplitude comparison [read: p.363]. - A neutral test would be a reanalysis of the multi-station Venus and later Mars/spacecraft ranging with a free ballistic parameter. Nobody has done this to my knowledge [memory, medium].

Important weaknesses I call out in Wallace's own analysis

• The reference "N" (Newcomb's perturbed radius vectors) is not an independent standard. Newcomb's Venus tables were optical, with errors of several thousand km at this epoch [memory, medium]; the 1961 radar contact was precisely what exposed these errors. Fig. 1's axis is in units of 10^-6 AU (1 unit is about 150 km), and the curves differ by about 5 units, i.e. 750 km. Matching N to the level of the ephemeris' own uncertainty is not a discriminating result.
• Eight points, each from a different day, with the AU value fixed at 149,597,850 km "the same as used by the Lincoln Lab" [read: p.365]. Since the Lincoln value was itself derived from the same radar data under the c assumption, the comparison is partly circular.
• Wallace argues that least-squares fitting "covers up inadequacies of the theory" by quoting Shapiro's own sentence [read: p.363]. That is fair as a methodological point, but it works both ways: the c+v curve G is a single integration with no stated fitting of free parameters, so it is not a fit with the same freedom. Either the comparison is fair (then say how the AU, orbital elements and masses were treated for each model) or it is not. The paper does not say.
• "3 ms vs 10 microseconds" ([read: ch.4]) compares an early residual with a later-achieved accuracy. In 1961 the Venus ephemeris was uncertain by about 3 ms scale (450 km one way = 3.0e-6 AU [my calculation]), and no one expected 10 us accuracy from the 1961 ephemeris. The "30,000%" phrasing is therefore misleading as stated; it measures the 1961 ephemeris error against a later radar-era error budget.
• The book says Shapiro "admitted" that only c was tested. Shapiro's own recorded reply is that the data are inconsistent with the ballistic theory. These are not incompatible (c was assumed, then the model was tested through its fit) but the book's phrasing treats the assumption as the weakness without addressing the reported Doppler test.

d. If true, what would have to change?

Special relativity's second postulate in its source-independence form; the whole GR light-time and Shapiro-delay modelling used for planetary and spacecraft navigation; Doppler tracking; the interpretation of binary-star and pion-decay experiments (Alvager) would need a different explanation. The Ritz tradition (see Investigator: Tolchelnikova, ritz-btr.narod.ru) accepts this and argues for alternative readings of those experiments.

Verdict claim 1

Split verdict: - Wallace's 1961 eight-point fit: Never properly tested (no independent replication or reanalysis found; the yardstick is weak, so also not a convincing test as given). - The strong form (light speed is c+v relative to the observer in general): Tested, not supported on the strength of source-independence experiments (Alvager) and ranging practice, with the planetary-ranging part resting on my [memory, medium]. The Doppler test by Shapiro's group is a published answer, which I could not read. - The radar-specific assertion (the c fit leaves unexplained station-dependent terms): Never properly tested by a neutral party. The data-holder said the data were consistent with GR; the requested form of the data was not provided; no one has published a c+v reanalysis. Dismissed with a short rebuttal, not ignored without comment.

Claim 2: Einstein's second postulate is the linchpin; Shapiro admitted c+v was never tried

• a. Maths: not applicable; this is a historical and methodological statement.
• b. The "linchpin" argument is partly true: Marilyn vos Savant is quoted as authority [read: ch.4], which is a weak source for physics.
• c. Whether Shapiro said this at the 4th Texas Symposium: single source (Wallace's account of a comment session; "I have quoted your remarks as I remembered them", Wallace's own letter, [read: ch.5]). No transcript found; the Investigator also found a date conflict (Wallace says Dallas 1968; ADS lists a report on the Fourth Texas Symposium in Soviet Astronomy dated Dec 1967) [web via Investigator]. Unresolved. Shapiro's reported letter (17 Jan 1969) does not deny that the analysis assumed c; it argues the data are inconsistent with the ballistic theory.
• d. If true: the radar practice of fitting everything with c would not test c. That is correct as a logical point: a fit model that assumes c can absorb a deviation into the fitted parameters (Shapiro: "the values of the parameters will be selected from the data in a manner that tends to cover up the inadequacies of the theory" [read: p.363]). Tolchelnikova-Murri (Pulkovo, independent professional) makes a related but distinct argument: the 1961-62 reductions used classical kinematics, and the round trip cancels the solar-system velocity, so radar neither confirms nor refutes SR [web: https://ritz-btr.narod.ru/tolchel.html, via Investigator; I read the Investigator's summary only]. Clemence (USNO), quoted by her (secondary), said second-order terms can be interpreted outside relativity.
• Verdict: Never properly tested in the narrow sense that the radar reductions were not designed as c-versus-c+v tests (a point both Wallace and Tolchelnikova make); the statement that this was publicly admitted is a single-source account (Wallace, with no transcript). The conclusion that nothing of this kind was later tested is not established.

Claim 3: inertial mass = photon virtual mass; internal circulation at c; one compressible "dynamic ether"

• a. Maths: no equations in the book [read: deep, section 0]. Deep-reader flags stand. My additional check from the deep read: the circulation radius implied by 1.2e-22 s per rotation at c is at most about 5.7e-15 m, giving angular momentum m_e c r of about 1.6e-36 J s, which is about 34 times smaller than hbar/2 [deep reader M3, I accept the arithmetic; hbar/2 = 5.27e-35 J s]. So the model as described does not yield electron spin without extra assumptions. Also, the ether frame in which the internal speed is c is unspecified, in tension with Wallace's own c+v light (M12).
• b. Experiments: e+e- annihilation into two 0.511 MeV photons is consistent with it, but is equally standard physics (E = pc for photons). Wallace's complaints about "motion created" and "momentum created" misdescribe the standard account (M2). The zitterbewegung literature (Dirac, de Broglie; Wallace cites the "small clock" of de Broglie) is a mainstream relative with the same circulating-at-c idea; it is not a vindication of Wallace's particular model [memory, medium].
• c. Tested: no experiment tests the model's distinctive prediction. The only candidate, force-law deviation below 1.1e-14 m, I discuss under claim 4.
• d. Change: the interpretation of the photon and electron as structureless; no measured outcome changes unless the model makes distinct predictions.
• Verdict: Not testable as stated in the book (no equations); the internal-circulation idea is Plausible, untested as an interpretation, with the spin mismatch as an open numerical problem for the model as described.

Claim 4: the "mass-in-cell" computer model reproduces photon emission, 1/r electron mass distribution, forces, spin, gravity and nuclear forces; predicts deviation below about 1.1e-14 m; pion made of electrons and positrons

• a. Maths: cannot check. Observations from the deep read that matter: (i) gravity is matched "to 3 s.f." with a surface-tension term, but the coupling G m^2/(e^2/4 pi eps0) is about 2.4e-43 and no source says whether it is derived or fitted (M5); (ii) cells of 2.8e-15 m are essentially the classical electron radius, an input, not an output (M4); (iii) a neutral pion of 104 or 10^4 electrons and positrons has the wrong mass defect in both readings (M7: 104 x 0.511 MeV = 53 MeV vs 135 MeV; 10^4 gives 5110 MeV).
• b. Experiments: the deviation prediction (Coulomb force law departing below 1.1e-14 m for electrons) contradicts electron point-likeness: high-energy e+e- scattering and g-2 constrain electron structure far below 10^-17 m [memory, high; deep reader M6]. Wallace's own crossed-beam experiment (Eckerd College, 300 eV, J. Chem. Phys. 66, 2108 (1977)) confirmed Moller scattering and did not test the model [read: deep section 6]. I credit the experiment as real laboratory work with a peer-reviewed publication; it supports standard QM.
• c. Tested: the model's own authors are the only checkers. Published in a journal of classical physics (J. Clas. Phys.) and Speculations in Science and Technology [read]. No independent replication found.
• d. If true, QED's electron would have to have a 5.5e-15 m mass radius, contradicting scattering data.
• Verdict: split. The e+e- composite pion: Refuted by calculation (mass budget in either reading, using accepted pion mass 134.98 MeV). Electron extended-structure prediction: Tested, not supported by existing scattering data [memory, high; no URL given, flagged for Physicist]. The simulations as a whole: Not testable as stated (equations not shown; resolution admitted to be low and "qualitative").

Claim 5: infinite eternal universe, non-Doppler redshift, matter recycling, pressure-induced annihilation powering stars and quasars

• a. Maths: none in the book; the arguments are citations (Arp, Pecker-Roberts-Vigier, Sandage, Oort, Holcomb).
• b. Experiments: Arp's anomalous-redshift claims are contested; the mainstream explains them as chance projections and selection effects [memory, medium]. The Jupiter figure is misquoted ("2 times" in extraction, 2.5 in the original; present measurements give about 1.7 [deep M13, memory medium]). Cosmic microwave background, light-element abundances and the Hubble relation favour an expanding, finite-age universe; Wallace does not address the CMB in the chapter [read: deep section 5, absence noted].
• c. Tests: Arp's work had a documented resistance history (telescope time denied, Arp left Mount Wilson/Palomar for Germany [read: ch.3 mentions Arp; details not reviewed by me]); the dispute is real but the objective status is unresolved by the sources I hold.
• d. A new redshift mechanism or steady-state cosmology.
• Verdict: Plausible, untested for tired-light/photon-photon energy loss only as a stated hypothesis; the cosmology as a whole, Historically superseded but instructive (steady-state-family ideas), with the Arp redshift anomalies Never properly tested by a neutral campaign as far as the record I hold shows. I did not review the Arp literature for this step.

Claim 6: obstruction by journals, referees, funders and meetings; QCD/quarks/relativity as pathological science

Evidence for (stated fairly)

• PRL 1969: referee report "suitable... if revised, and deserves immediate publication if the radar data can be compared directly..." (15 Aug); revised and resubmitted 21 Aug; editor Goudsmit 11 Sep says it went to another referee and was rejected; 23 Sep says he had "made a mistake" and it went back to the first referee [read: ch.4]. Documents quoted from his correspondence; letters not reproduced, single source (Wallace), but specific with dates and names. Weighed as testimony from an interested party.
• Spectroscopy Letters editor J. W. Robinson, letters of 4 Apr 1979 and 23 Jul 1990, said he could no longer publish Wallace's papers "because the very negative reader reaction" and had received "completely unsavory and unobjective anonymous letters and phone calls" [read: deep section 1]. This is a documented claim by one person with no stake in Wallace's own side, and Wallace's report is consistent with an independent published rebuttal: Horedt and Ruck, SpecL 5(12), 455 (1972), answering "a series of papers 1-15" that attempted to discredit the constancy of light speed [web: https://www.tandfonline.com/doi/abs/10.1080/00387017208065415]. That the journal ran such papers for several years and then ran a published reply is evidence of an editorial process, not a ban; the later closure is documented only by Robinson's letters as Wallace relays them.
• Walter Sullivan (NYT) deferring to Thornton Page, 13 Mar 1969 [read: ch.5]. NSF rejection 1978 [read: ch.4, claims.md]. The Santilli allegations are Santilli's own, relayed by Wallace [deep section 8].
• Standish (JPL) at the Dec 1974 AAS meeting reporting significant unexplained systematic variations in all the interplanetary data, requiring empirical correction factors [read: ch.4, Wallace's report; I did not verify]. If correct, this is the type of evidence that deserves an independent look: unexplained systematics are not a c+v finding, but they were also not published as a proper c+v test.

Evidence against, and countervailing points

• Wallace's papers did appear: four Spectroscopy Letters papers (1969-71), Found. Phys. (1973), J. Chem. Phys. (1977), Rev. Sci. Instrum. (1975), letters in Physics Today, Physics Essays (1990), Galilean Electrodynamics (1990) [read; web: NPA wiki]. The "first radar paper" was rejected by "a large selection of journals" [read: deep], with one referee's positive report; the other referee's strongly negative report was not shown to the reader. The outcome is unresolved without those documents.
• The PRL story involves a negative second report whose content Wallace does not give; the rejection may have been on substance. The dispute on process vs substance cannot be settled from the book.
• Wallace's own text contains rhetoric that weakens his case: "Shapiro has pulled a fast one", "little more than misleading fabrications", "heresy" analogies, and a speculation that the "real reason" for the data not being released was military secrecy tied to "Star Wars" [read: ch.5]. These are inferences with no evidence in the book. He also suggests a joint paper to "repair any damage I may have done to your reputation" [read: ch.5]: that tone plausibly made collaboration harder.
• Wallace's pathological science (Langmuir) argument, aimed at QCD, quarks and relativity, is an opinion; I note that Langmuir's criteria (effect near the limit of detectability, ad hoc excuses, disagreement with the critics) could be turned on the c+v result itself, whose size is comparable to the ephemeris uncertainty.

The absence-of-evidence anomaly

After 1971 Wallace's radar case receives little follow-up in the record I hold (Investigator: ADS shows only 12 citing papers, no replication). Competing explanations: (1) the argument was judged wrong on the merits by specialists (the Horedt and Ruck rebuttal, Shapiro's letter); (2) hostile reception through informal channels as Robinson wrote; (3) lack of access to the data, the real bottleneck as Wallace himself says; (4) classification of tracking data (Wallace's speculation, no evidence). Evidence for (3) is Shapiro's and Hollingsworth's letters as relayed. A record that would decide between them: the Horedt and Ruck text, the Fourth Texas Symposium discussion record, Lincoln Lab technical reports, and independent reanalysis of early multi-station radar. All of them are unfound.

Verdict claim 6

Split: the claim that Wallace faced editorial hostility is supported by documents quoted from correspondence and one published rebuttal, not independently verified; the claim of organised suppression is not established. Classification for the underlying physics: dismissed with a brief answer, not properly tested.

What I would keep in the record (Investigator findings weighed)

• Tolchelnikova-Murri (Pulkovo) independent analysis of the 1961-62 radar reductions, reaching a related but different conclusion (round trip cancels; classical Doppler formula used): keep, as the strongest independent support that the radar reductions are not a clean test of SR. Her paper is not a c+v result.
• Clemence (USNO) quote on second-order Doppler: keep as her rendering; primary (IAU Symposium 21) not read.
• Rapier 1971 (3C 279 VLBI), Wegener and Ohrstrom 1975, Malykin 2010 and 2012, Cao 2011, Tran 2022: keep as the citing literature; none read.
• Horedt and Ruck 1972: keep as the nearest rebuttal; text not read.
• NPA wiki biography (born 8 Feb 1933, died 19 Mar 1997; Leningrad conferences 1989 and 1991): keep, as a sympathetic society's biography. The viXra "American radio astronomer" description conflicts with other sources (he was not a professional astronomer by the book's own account); not used.
• A. G. Kelly on Hafele-Keating (FOIA reanalysis) and Essen: related work on absolute-frame readings; I note Hafele and Keating computed in a non-rotating geocentric frame, not a galactic one [deep M8, memory high]; did not weigh Kelly's reanalysis.

Summary table

What would settle it

A fair reanalysis of multi-station early radar and later spacecraft range data with a free ballistic (c+v) parameter, using the same epoch bookkeeping for both models; obtain and read the 1971 Wallace paper, the Shapiro-Ash-Tausner PRL text, the Horedt and Ruck rebuttal, and the Fourth Texas Symposium discussion.