# Stellar Parallax Timeline A strict chronological walk-through of every published stellar parallax measurement that was negative, reversed in direction, or mutually incompatible with its peers, from Ptolemy to Gaia EDR3. Each entry is paired with the exact screenshot from the primary source so the quote can be read in situ. The narrative thread: the heliocentric model predicts positive-only parallax in a single rotational sense, and the catalogue record has never stopped producing roughly a quarter negative parallaxes across 2000 years, 100 years of photographic plates, and two space missions. Each entry is sized to one presentation slide. --- ## 0. The geometric prediction being tested ### Heliocentric parallax must be positive only, in one rotational sense From [[Shack_Chapter25_Negative_Parallax_Demystified]] p.3: ![[Shack_Chapter25-03.png]] > "Imagine yourself travelling in a car orbiting around the Sun (as Earth supposedly does according to Copernican theory). As you look out from your lefthand window, you will see the Sun at all times. [...] In order to SEE any stars, you will have to look out of your righthand window at all times. Therefore, if you were to measure any stellar parallax (of relatively nearby stars against more distant ones), only 'positive' parallax could possibly be observed at all times." > "Yet, about 1/2 (or 50%) of all the stellar parallaxes listed in ESA's Tycho catalogue have negative values, essentially meaning that they were observed to move from right to left (or from west to east) in relation to the more distant background stars. How can this possibly be?" From the same chapter p.21 closing argument: ![[Shack_Chapter25-21.png]] > "This is why 'negative' stellar parallax constitutes a physical impossibility within the heliocentric paradigm and an utterly insurmountable problem." > [!critical] The prediction to be falsified > Heliocentric parallax has a fixed sign convention tied to Earth's orbital sense. A negative published parallax is a sky motion 180° out of phase with Earth's orbit. If the catalogue record contains a structured population of negatives, the heliocentric prediction fails. --- # Era I: Before Bradley (150 AD to 1727) ## 1. Ptolemy, *Almagest* Book I Chapter 7, c. 150 AD Ptolemy's geocentric argument was parallax-based: if Earth moved, relative stellar positions would shift annually. No such shift was detected in the next 1,700 years of pre-telescopic observation. Ptolemy treated the absence of parallax as positive evidence for a fixed Earth. The Copernican counter-argument, that the stars are "too distant for parallax to be detected," is the same move modern astronomers make when they treat the negative Gaia parallaxes as Gaussian noise. Context in [[Stellar_Parallax]] §"Ptolemy, *Almagest* Book I Chapter 7 (c. 150 AD)". ## 2. Hooke 1674, γ Draconis at 30 arcseconds in the wrong direction Robert Hooke, *An Attempt to Prove the Motion of the Earth from Observations* (London 1674). Zenith-tube at Gresham College, four observations of γ Draconis in 1669. Computed annual parallax approximately 30″. Modern value: 0.02″. Hooke was high by factor 1,500, and the direction was wrong for heliocentric γ Draconis. Source context is in the Williams 1981 PhD thesis, Chapter 1. Thesis title page: ![[Williams1981-001.png]] Chapter list showing Ch. 1 "Hooke and Flamsteed's early attempts": ![[Williams1981-003.png]] ![[Williams1981-004.png]] Summary in [[Stellar_Parallax]] §"Hooke 1674": > "Robert Hooke's *An Attempt to Prove the Motion of the Earth from Observations* (London 1674) reports four observations of γ Draconis from a Gresham College zenith tube during 1669. From these he computed an annual parallax of approximately 30 arcseconds. The modern value for γ Draconis is 0.02 arcseconds. Hooke was high by a factor of 1,500. The direction of the apparent shift, 'more northerly in July than in October,' does not match what heliocentric parallax of γ Draconis should produce." ## 3. Flamsteed 1698, Polaris "parallax" of 40 arcseconds was aberration plus nutation John Flamsteed, first Astronomer Royal. Reported an annual variation in Polaris's declination of approximately 40″ between July minimum and September maximum. Published as parallax in 1698. Cassini flagged the phase as wrong. Bradley's later work showed the signal was aberration (20″) + nutation (9″) + instrumental residuals. True Polaris parallax: zero at Flamsteed's precision. Summary in [[Stellar_Parallax]] §"Flamsteed 1698": > "John Flamsteed, the first Astronomer Royal, reported an annual variation in Polaris's declination of approximately 40 arcseconds between July minimum and September maximum. He published this in 1698 as parallax. Cassini noted that the direction and phase were not consistent with heliocentric parallax of Polaris. Bradley's work 30 years later showed Flamsteed was measuring a combination of aberration (20 arcseconds) and nutation (9 arcseconds) with instrumental residuals filling the remainder. The parallax signal was zero." See Williams 1981 Chapter 1 (screenshots above). --- # Era II: Bradley himself (1725 to 1728) ## 4. Bradley 1728, discovered aberration because γ Draconis was 90° out of phase James Bradley, *A Letter from the Reverend Mr. James Bradley ... to Dr. Edmond Halley ... giving an Account of a New Discovered Motion of the Fix'd Stars*, *Phil. Trans.* 35, 637–661 (1728). Zenith-tube observations from Kew and Wanstead of γ Draconis. A 20.5″ annual modulation peaked in March and September. Heliocentric parallax would peak in June and December. The 90° phase mismatch told Bradley the signal was velocity-driven, not geometric. Primary source, [[1728_Bradley_New_Motion_Fixed_Stars]] page 1: ![[page-01 2.png]] Page 3, the key discovery moment when the star moves in the wrong direction for parallax: ![[page-03 2.png]] Bradley expected parallax. He found γ Draconis moving south when parallax would have moved it north. That was the discovery of aberration. Page 4, the full annual cycle of approximately ±20″: ![[page-04 2.png]] Bessel's 1838 paraphrase of Bradley's conclusion, [[1838_Bessel_61_Cygni_Parallax]] p.3: ![[Bessel1838_61Cygni-003.png]] > "When Bradley had succeeded in explaining his observations at Kew and Wanstead, which led to the discovery of aberration and nutation, by these effects alone, without needing to assume an annual parallax for the observed fixed stars, he did not fail to note that a value of more than one second of arc for the parallax could not have escaped his observations of γ Draconis and η Ursae majoris." ## 5. Bradley's Polaris data: NEGATIVE parallax (surfaced later by Bessel) Bradley's Greenwich Polaris observations, when reduced by Bessel around 1815, yielded a negative parallax. Bradley never published this as such; Bessel flagged it in his 15 February 1816 letter to Olbers (see entry 11 below). This entry foreshadows the post-Bradley pattern: the archival record contained negatives from the start, and the noise-framework was imposed on top of that archive. Source: [[1852_Bessel_Olbers_Briefwechsel_Polaris_Quote]] (see entry 11 for the full primary-source screenshot). --- # Era III: After Bradley, before Bessel 1838 (1761 to 1837) ## 6. Lalande (Paris 1761–1762), Sirius 2.5″ in the wrong direction Jérôme Lalande's Paris observations of Sirius, 1761–1762, showed an annual displacement of approximately 2.5 arcseconds, direction wrong for heliocentric Sirius parallax. Source: historical table in [[Shack_Chapter25_Negative_Parallax_Demystified]] p.19: ![[Shack_Chapter25-19.png]] ## 7. Maskelyne (from Lacaille), Sirius 8″ in opposite direction to Lalande Nevil Maskelyne derived π(Sirius) ≈ 8″ from Lacaille's Cape of Good Hope observations, in a direction opposite to Lalande's Paris result. Same star, same era, two opposite-sign answers. Source: [[Shack_Chapter25_Negative_Parallax_Demystified]] p.19 (screenshot above). > [!note] Shack's 3:1 ratio cross-check > Shack flags that 2.5 / 8 ≈ 1/3, which matches the ratio his Tychos quadrant model predicts for observations made in different time windows of the year. Whether one accepts Shack's interpretation or not, the raw fact is that two independent 18th-century observers reported Sirius parallax in opposite directions. ## 8. Zach vs Maskelyne, same ten stars, opposite signs (reported by Baily 1845) Francis Baily, four-time President of the Royal Astronomical Society, documents in *The Catalogue of Stars of the British Association for the Advancement of Science* (1845) that Baron von Zach and Nevil Maskelyne reported parallaxes for the same ten stars with opposite signs. Zach: ten positives. Maskelyne: ten negatives. Source: [[Shack_Chapter25_Negative_Parallax_Demystified]] p.7: ![[Shack_Chapter25-07.png]] Baily's own words, as reproduced by Shack: > "For, in many cases, some of the greatest names have differed even as to the direction of the motion of particular stars: one making it positive whilst in the same star another considers it as negative." > [!critical] The cleanest pre-Bessel falsifier > Same ten stars, same era, opposite signs. This cannot be explained by saying one observer had a better instrument than the other. It is the historical fingerprint of a method whose sign is not determined by the sky. ## 9. Brinkley vs Pond 1810–1824, a 14-year negative-vs-null fight John Brinkley at Dunsink (8-foot circle) reported 1 to 2.7 arcsecond parallaxes for Altair, Deneb, and Vega. John Pond, Astronomer Royal at Greenwich, found no such parallaxes with superior mural instruments. The dispute ran 14 years in *Philosophical Transactions* and the *Transactions of the Royal Irish Academy*. Brinkley never retracted; Pond never conceded. Primary secondary-source treatment in Williams 1981 thesis, §4.4. The Brinkley-Pond chapter opening (PDF p.159): ![[Williams1981-159.png]] ![[Williams1981-160.png]] Williams p.159 direct quote: > "the celebrated controversy over whether or not Brinkley had detected [parallax]." Pond's 1822 doctrine, as summarised in [[Stellar_Parallax]] §"Brinkley vs Pond 1810–1824": > "In proportion as instruments have been imperfect in their construction, they have misled observers into the belief of the existence of sensible parallax." > [!note] Pond's 1822 doctrine is the ancestor of the modern Gaussian-tail defence > Pond's claim that imperfect instruments generate spurious parallaxes is structurally the same argument Luri 2018 and Dworetsky 2016 make for Gaia negatives. The statistical-noise explanation for unwanted parallaxes is 200 years old. ## 10. Bessel 1815, first negative reduction of 61 Cygni (Königsberg Vol. II) Bessel's first published reduction of 61 Cygni, using the right-ascension difference with six comparison stars, came out negative. Published in *Königsberger Beobachtungen* Volume II, 1815. Primary secondary-source quote from Williams 1981 p.140 ([[1981_Williams_Hooke_to_Bessel_PhD_Thesis]]): ![[Williams1981-140.png]] > "But Bessel was to be disappointed again: when he had finished the reduction of the position of 61 Cygni relative to the six different stars he was forced to the conclusion that its parallax was negative! The paper in which this result was announced took the form of a report only, with no explanation of why a negative answer might have been obtained." ## 11. Bessel 1815, second negative reduction: μ Cassiopeiae (Königsberg Vol. II) Same volume, Bessel applied the same method to μ Cassiopeiae using θ Cassiopeiae as comparison. The result was negative, smaller in magnitude than the 61 Cygni negative. Williams 1981 p.140 (same screenshot above): > "Bessel gave tables of observations, and results of the application of the method of least squares to these observations for each comparison in turn; he followed this with exactly the same information for μ Cassiopeiae which he had compared with θ Cassiopeiae. For this star also he had a negative, though numerically smaller result." ## 12. Bessel to Olbers, 15 February 1816: "observational errors" framing established Bessel writes to Olbers privately that Bradley's Polaris data, reduced, came out negative, and that negative parallax in general is an artefact of observational error. This is the primary document of the noise framework. Primary source, [[1852_Bessel_Olbers_Briefwechsel_Polaris_Quote]] p.17 (PDF p.29 of the combined-volume scan): ![[BesselOlbers1852-029.png]] Tight crop of the quote: ![[BesselOlbers1852-029_QuoteCrop.png]] Direct quote in Bessel's own German: > "Die negative Parallaxe, die man wohl hin und wieder finden wird, und die ich bei dem Polarsterne wirklich aus Bradley's Beobachtungen gefunden habe, ist allerdings eine Wirkung von Beobachtungsfehlern; allein ich glaube sie wird gewöhnlich in den Grenzen der wahrscheinlichen Fehler liegen, oder doch nicht weit ausser denselben, wie es auch bei dem angeführten Falle stattfand." English translation (following Williams 1981): > "The negative parallax, which one will certainly find here and there, and which I have actually found for the Pole Star from Bradley's observations, is indeed an effect of observational errors; but I believe it will usually lie within the limits of the probable errors, or at least not far outside them, as was also the case in the instance cited." Williams 1981 English translation of the same letter at p.141: ![[Williams1981-141.png]] > "The negative parallax which one found here and there and which he had in fact found for the Pole Star from Bradley's observations was of course the result of observational errors." Williams footnote 58 pinning the date and source: ![[Williams1981-248.png]] > "Footnote 58: Bessel to Olbers, 15 February 1816, *Bessel-Olbers Correspondence*, ii, 14–20, 17: 'Die negative Parallaxe, die man wohl hin und wieder finden wird, und die ich bei dem Polarsterne wirklich aus Bradley's Beobachtungen gefunden habe, ist allerdings eine Wirkung von Beobachtungsfehlern.'" > [!critical] The noise framework was declared in a private letter, not proved > Bessel inherited Bradley's negative Polaris result, and had found two of his own at Königsberg. His response was to call them "observational errors" in a letter to Olbers. He carried that framework for 22 years. The interpretation of negative parallax that is still dominant in 2026 starts here, in correspondence, not in a proof. ## 13. Bessel 1816–1817, THIRD negative reduction of 61 Cygni: T = −1.32″ (Königsberg Vol. III) Bessel's 1816 observations of 61 Cygni using α Cygni as the comparison star, reduced by least squares on 43 measurements from 9 July to 8 December 1816, gave T = −0.088 = −1.32 arcseconds. Published in *Königsberger Beobachtungen* Volume III, 1817, section IX. Primary-source screenshot from [[1817_Bessel_Koenigsberger_Beobachtungen_Vol3]] PDF p.11: ![[BesselKoenigsberg1817v3-011.png]] Title page of Vol. III: ![[BesselKoenigsberg1817v3-001.png]] Direct quote from section IX (verbatim German): > "Nach der Methode der kleinsten Quadrate findet man hieraus > Unterschied der AR. für 1816 .... = 32'29,"7214 > T = −0,08800 = −1,32 in Bogentheilen > Die erste Bestimmung hat das Gewicht von 28,67, die andere von 15,04 directen Beobachtungen." English: > "By the method of least squares one finds from this: Difference of right ascension for 1816 = 32'29.7214, **T = −0.088 = −1.32 in arcseconds**." Williams 1981 p.140 summary of the third reduction: > "In volume III of the Konigsberg observations Bessel gave another set of observations, this time of the difference of right ascension between α and 61 Cygni from which he deduced an even larger negative result for the parallax of 61 Cygni." > [!critical] Primary-source accounting > Bessel published a negative parallax value of −1.32″ for 61 Cygni in 1817. That is not a secondary-source narrative; that is the number on the page of his own book. The three successive negatives (61 Cygni 1815, μ Cas 1815, 61 Cygni 1816–1817) are the archival record. The "first stellar parallax" textbook story starts in 1838, after three published negatives and one private retraction. ## 14. Struve 1837, first positive stellar parallax (Vega, 0.125″) F. G. W. Struve at Dorpat published π(Vega) = 0.125″ ± 0.055″ in *Mensurae Micrometricae* (1837). The modern Hipparcos value for Vega is 0.129″. Struve's 1837 figure was within 4% of modern. Primary modern reassessment in [[2020_Reid_Menten_First_Parallaxes_Revisited]] PDF pp.2–3: ![[ReidMenten2020_FirstParallaxes-002.png]] ![[ReidMenten2020_FirstParallaxes-003.png]] > "Struve first used all of his measurements, but in the end lost confidence in some and discarded those." --- # Era IV: The textbook "first parallax" and its immediate companions (1838 to 1840) ## 15. Bessel 1838, textbook "first stellar parallax": π(61 Cygni) = 0.314″ (fourth reduction) Bessel publishes $\pi = 0.3136'' \pm 0.0202''$ for 61 Cygni in *Astronomische Nachrichten* 16, cols. 65–96. Modern value: ~0.286″. This is the fourth reduction of 61 Cygni by the same observer. Three earlier reductions had given negative values; all were privately attributed to observational error. Primary source: [[1838_Bessel_61_Cygni_Parallax]] title page: ![[Bessel1838_61Cygni-001.png]] ![[Bessel1838_61Cygni-002.png]] Opening of the paper, p.3, framing the parallax problem: ![[Bessel1838_61Cygni-003.png]] > "From this arose the later commonly accepted assumption that the annual parallax of the fixed stars is in general very small. While this assumption is undoubtedly true for the great majority of these innumerable stars, it is equally undoubtable that some of them are much closer to us than the great mass of the others." The result: $\pi_{61\text{ Cyg}} = 0.3136'' \pm 0.0202''$ > [!warning] The textbook story omits three negatives and the 1816 letter > Every astronomy textbook prints "Bessel 1838" as the first successful stellar parallax. It is actually Bessel's fourth reduction of the same star. The three earlier negatives (entries 10, 11, 13) and their 1816 private-letter retraction (entry 12) are not mentioned. ## 16. Henderson 1839, α Centauri at 1.16″ (55% too large) Thomas Henderson, Cape of Good Hope observations of α Centauri reduced in 1839: π = 1.16″. Modern value: 0.747″. Henderson was wrong by more than half, but he underestimated his probable error so the result was accepted as a clean detection. [[2020_Reid_Menten_First_Parallaxes_Revisited]] direct quote (screenshot above at entry 14): > "Henderson underestimated some of their measurement uncertainties, which made their parallaxes appear somewhat more significant than they actually were." ## 17. Struve 1840, Vega revision: π doubled to 0.261″ by discarding measurements Three years after Struve's correct 1837 value (0.125″), he republished π(Vega) = 0.261″ ± 0.025″. Double the truth, with a tighter error bar. The revision was obtained by trimming measurements Struve "lost confidence in." Reid & Menten 2020 (screenshot above at entry 14) summary: | Year | π(Vega) | Relation to modern 0.129″ | |---|---|---| | 1837 (all data) | 0.125″ ± 0.055″ | Within 4% | | 1840 (trimmed) | 0.261″ ± 0.025″ | Twice too large | > [!critical] Data-trimming produced a worse answer with a tighter error bar > Struve 1837 (all data) was correct to 4%. Struve 1840 (trimmed) was wrong by a factor of 2 with half the reported error. The "clean" dataset was the wrong dataset. This is the earliest clearly-documented case of selection bias in the parallax programme. ## 18. Baily 1845 reporting Zach vs Maskelyne (canonical footnote) Francis Baily's *Catalogue of Stars of the British Association for the Advancement of Science* (1845) footnote on Zach and Maskelyne reporting opposite signs for the same ten stars. Already covered above at entry 8; the Baily attribution is locked here because 1845 is the publication date of the catalogue. [[Shack_Chapter25_Negative_Parallax_Demystified]] p.7 (screenshot above at entry 8). --- # Era V: The photographic-parallax era (1878 to 1966) ## 19. Newcomb 1878, "a paradoxical result can arise only from errors of observation" Simon Newcomb's general comment on the parallax programme, quoted in the historical table of [[Shack_Chapter25_Negative_Parallax_Demystified]] p.19 (screenshot at entry 6): > "Such a paradoxical result can arise only from errors of observation." > [!note] The 1878 Newcomb sentence is the 2026 Dworetsky sentence > Newcomb's one-line disposal of negative parallax has not been reopened in 148 years. Dworetsky 2016 (quoted in [[Stellar_Notes]] §3) says the same thing with statistical-distribution language. Same claim, same authority-by-assertion, different vocabulary. ## 20. Eichelberger 1911 (*Science*), roughly 1:1 positive-to-negative at ~6 stars W. S. Eichelberger, "The Distances of the Heavenly Bodies," *Science*, April 7, 1911. Six stars: somewhat more than half had measurable parallax; of those, roughly three were negative. Ratio positive:negative approximately 1:1. Source: [[Shack_Chapter25_Negative_Parallax_Demystified]] p.9: ![[Shack_Chapter25-09.png]] ## 21. Washburn meridian transits 1915, ~1:1 ratio again "Results for parallax from meridian transits at the Washburn Observatory" (1915) reported the same 1:1 positive-to-negative ratio in its published table. Source: [[Shack_Chapter25_Negative_Parallax_Demystified]] p.9 (screenshot above). ## 22. Aitken 1921, 26% of 1013 parallaxes negative or below probable error Robert G. Aitken, future director of Lick Observatory, in *Publications of the Astronomical Society of the Pacific* 33, no. 191 (February 1921), p. 41. Direct count of every published parallax from six programmes: McCormick, Allegheny, Yerkes, Sproul, Mount Wilson, Greenwich. Primary source: [[1921_Aitken_Recent_Parallax_Review]] title page (PDF p.1): ![[Aitken1921_PASP-001.png]] The key page (PDF p.5, book p.45): ![[Aitken1921_PASP-005.png]] Verbatim quote: > "A single (unchecked) count of the six series of measures (including the Greenwich results) shows that of the 1013 separately listed parallaxes 152 have the value 0".000 or else negative values; 111 have positive values less than or just equal to their probable error; and 141 have positive values greater than, but not more than double the size of their probable error." > "Stated as percentages **26 per cent of these parallaxes are negative or have positive values which are not greater than their probable errors; 14 per cent more have positive values which are not larger than twice their probable errors**." Aitken's own interpretation of the negatives (same page): > "The parallaxes of the first class merely indicate that the stars in question are not nearer to us, or are farther away from us, than the comparison stars employed. They give us no definite information as to the actual distances." > [!critical] The 26% figure is 1921 mainstream, not heterodox > *PASP* is one of the top three astronomical journals of the era. Aitken was about to become director of Lick Observatory. By his own direct count, 26% of the catalogue was negative or below probable error, and 40% was below 2× probable error. This is the baseline against which the modern Tycho 25% and Gaia 24% fractions must be compared. ## 23. Lee 1943, structural causes of negative parallax (Dearborn Observatory) O. J. Lee, "On a Reason for the Appearance of Negative Parallaxes in the Determination of the Distances of Stars," *Annals of the Dearborn Observatory of Northwestern University* IV, Part 1 (1943). Primary source: [[1943_Lee_Negative_Parallax]] p.1: ![[Lee1943_NegParallax-01.png]] Direct opening quote: > "Everyone who has made any considerable number of determinations of stellar parallaxes by trigonometric methods has occasionally found that the usual procedure gave him a negative parallax for the star under investigation." Lee's three accepted causes, same page: 1. Chance error on a small real parallax. 2. Close visual double stars mis-blending under seeing variations. 3. The comparison stars in the target's field are not at infinity. Cause 3 verbatim: > "Obviously, trigonometric conditions are just as valid for a distant star with comparison stars that are near to us as they are in the usual inverse situation. A large negative parallax may be just as real as an equal positive parallax. It must merely be recognized that it is the positive parallax of the comparison stars with respect to the distant star which has been gotten." Page 2, the empirical test ruling out seasonal errors: ![[Lee1943_NegParallax-02.png]] > "Obviously, seasonal errors in astrometry cannot be invoked to account for the distribution of negative parallaxes as shown by these curves, for, with all parallax observing confined to small hour angles, the plates of fields in all declinations are taken in the same season and should be affected by the same seasonal errors if such exist." Page 3, conclusion: ![[Lee1943_NegParallax-03.png]] > "Whatever reasons of chance may exist for the appearance of negative parallaxes, as they must exist, it seems inevitable to conclude that their distribution of size in right ascension is related to the parallaxes of the comparison stars." > [!note] Lee 1943 matches Aitken 1921, not Dworetsky 2016 > Lee and Aitken (22 years apart) both read negative parallax as structural geometry, not as noise. Negatives happen because the "far" comparison field is actually near. The modern noise framework comes later, after the transition to space-based astrometry. ## 24. Vasilevskis 1966, "the reality of some parallaxes were in doubt" Stan Vasilevskis, "The Accuracy of Trigonometric Parallaxes of Stars" (Lick Observatory review, 1966). Reviewed parallaxes from the four major American observatories of the mid-20th century. Source: [[Shack_Chapter25_Negative_Parallax_Demystified]] p.12: ![[Shack_Chapter25-12.png]] Direct quote: > "Parallaxes of the same stars determined by different observers and instruments often disagreed to such an extent that the reality of some parallaxes were in doubt." --- # Era VI: The space-astrometry era (1997 to 2026) ## 25. Hipparcos (1997), Pleiades 1 mas systematic survived 20 years Original Hipparcos (1997): Pleiades at 120.2 ± 1.5 pc. Ground-based and VLBI (Melis et al. 2014): 135.2 pc. The 1 mas systematic on a thoroughly studied reference cluster was documented by Makarov 2002, not fixed by the 2007 van Leeuwen re-reduction, and only resolved by Gaia DR2 in 2018 at 135.15 pc. Summary in [[Stellar_Notes]] §25: > "Hipparcos was the state-of-the-art for 20 years and had a 1 mas systematic in the Pleiades that nobody caught. Gaia could carry a similar systematic on some other reference population that only CHES / Theia / GaiaNIR will find." ## 26. Tycho catalogue (ESA, ~2 million stars), 25% negative / 29% positive / 46% zero ESA's Tycho Main Catalogue documentation on the distribution of its published parallaxes. Source: [[Shack_Chapter25_Negative_Parallax_Demystified]] p.17: ![[Shack_Chapter25-17.png]] Direct quote from the Tycho Main Catalogue as reproduced by Shack: > "The trigonometric parallax is expressed in units of milliarcsec. The estimated parallax is given for every star, even if it appears to be insignificant or negative (which may arise when the true parallax is smaller than its error). 25% have negative parallax, 29% positive parallax and 46% assumed zero parallax." | Tycho breakdown | % | |---|---| | Positive | 29% | | Negative | 25% | | Assumed zero | 46% | > [!critical] Only 29% of the Tycho catalogue is usable by the mainstream framework > 25% negative plus 46% assumed-zero equals 71% with no mainstream-usable distance. The 25% negative fraction is the central modern-era anchor for this timeline and for the argument that the ~25% number is a century-long constant (compare Aitken 1921's 26%). ## 27. Goretti 2013, "the parallax angle is positive by definition" Vittorio Goretti, Italian astronomer, demanded clarification from ESA on the Tycho negative parallaxes. His public note is summarised in [[Shack_Chapter25_Negative_Parallax_Demystified]] p.15 and in [[Goretti2013_RedStars]]: ![[Shack_Chapter25-15.png]] First page of Goretti's own 2013 paper: ![[Goretti2013_RedStars-001.png]] Direct quote from Goretti via Shack p.15: > "As a matter of fact, about half the average values of the parallax angles in the Tycho Catalogue turn out to be negative! The parallax angle, which is one of the angles of a triangle, is positive by definition." Goretti's RMS objection, same page: > "When averaging many parallax angles of a star, the measurement error of the average (root-mean-square error) cannot be smaller than the average of the errors (absolute values) of the single angles." ## 28. Gaia DR2 (2018), "the source going the wrong way around on the sky" The Gaia consortium's own manual for using DR2 parallaxes. Luri, Brown, Sarro, Arenou, Bailer-Jones, Castro-Ginard, de Bruijne, Prusti, Babusiaux, Delgado, *A&A* 616, A9 (2018). Abstract, [[2018_Luri_Gaia_DR2_Parallaxes]] p.1: ![[Luri2018_GaiaDR2-01.png]] > "The second Gaia data release (Gaia DR2) provides precise five-parameter astrometric data (positions, proper motions, and parallaxes) for an unprecedented number of sources (more than 1.3 billion, mostly stars)." > "In particular we also show that negative parallaxes, or parallaxes with relatively large uncertainties still contain valuable information." Page 2, the global zero-point offset from 556,849 quasars: ![[Luri2018_GaiaDR2-02.png]] > "The centring adopted in this plot reflects a global parallax zero-point shift of −0.029 mas." Page 4, the pipeline's own definition of negative parallax: ![[Luri2018_GaiaDR2-04.png]] > "In the presence of large measurement noise (comparable to the size of the parallax) it is entirely possible that the parallax value estimated for the source model vanishes or becomes negative. This case can be interpreted as the measurement being consistent with the source going 'the wrong way around' on the sky, as shown in Fig. 2." > "In the fitted solution the negative parallax effect is equivalent to a yearly motion of the star in the opposite direction of the true parallactic motion (which gives a phase-shift of π in the sinusoidal curves in the right panels)." Page 8, the delete-the-negatives bias: ![[Luri2018_GaiaDR2-08.png]] > "However, if we remove the negative parallaxes from this sample, deeming them 'unphysical', the mean of the observed values would be significantly positive, about 0.8 mas. This is completely unrealistic for quasars; in removing the negative parallaxes we have significantly biased the observed parallax set for these objects." > [!critical] The 180° phase flip is the direct contradiction of the heliocentric prediction > Luri et al. define negative parallax as a stellar sky motion 180° out of phase with the motion Earth's orbit should produce. This is exactly the signal the Shack car-and-windows argument says cannot happen under heliocentrism. The consortium itself uses the phrase "the wrong way around." ## 29. Gaia EDR3 Bailer-Jones 2021, 24% of 1.47 billion sources negative C. A. L. Bailer-Jones, J. Rybizki, M. Fouesneau, M. Demleitner, R. Andrae, "Estimating Distances from Parallaxes V: Geometric and Photogeometric Distances to 1.47 Billion Stars in Gaia Early Data Release 3," *AJ* (2021). Primary source: [[2021_BailerJones_EDR3_Distances]] p.1: ![[BailerJones2021_EDR3-01.png]] Direct quote from the introduction: > "While parallaxes ($\varpi$) are the basis for a distance determination, they are not themselves distances ($r$). [...] while parallaxes can be negative, distances cannot be. Thus for anything but the most precise parallaxes, the inverse parallax is a poor distance estimate." Page 2, the fractions: ![[BailerJones2021_EDR3-02.png]] > "In EDR3 43% of the sources have parallax uncertainties greater than 50% (63% greater than 20%), and a further **24% have negative parallaxes**." > [!critical] 24% negative in 1.47 billion EDR3 sources > This is the number to put on the "modern catalogues" slide. Not Shack's percentage, not Goretti's. The Gaia distance-catalogue paper itself says 24% negative. Cross this against Aitken 1921's 26% and Tycho's 25% and the constancy across a century and a factor-1000 precision improvement becomes the central argument. ## 30. Lindegren 2021, EDR3 parallax bias correction is a polynomial and its own uncertainty is of order the measurement L. Lindegren et al., "Gaia Early Data Release 3: The parallax zero-point correction," *A&A* (2021). The Gaia team's own polynomial recipe for the parallax zero-point offset. Primary source: [[2021_Lindegren_EDR3_Parallax_Bias]] p.1: ![[Lindegren2021_ParallaxBias-01.png]] ![[Lindegren2021_ParallaxBias-02.png]] Key facts from [[Stellar_Notes]] §5: - Correction spans −150 to +130 µas across the sky (total span 0.28 mas). - The $Z_5$ bias function is a polynomial in magnitude, colour, and ecliptic latitude. - Five fitted coefficients whose error bars are themselves 5–50 µas. - Lindegren's own verdict: "it is not possible to derive a definitive recipe for the parallax correction." - Independent samples (Cepheids, RR Lyrae, asteroseismology) show the official correction over-corrects by +14 to +25 µas. ## 31. Bobylev 2026, VLBI and Gaia disagree on the zero point by a factor of 2 From [[2026_Bobylev_VLBI_Gaia_Zero_Point]] p.1: ![[Bobylev2026_VLBI-001.png]] - 151 radio stars and masers with VLBI parallaxes common to Gaia EDR3/DR3. - Systematic shift: Δπ = −0.038 ± 0.011 mas relative to VLBI/ICRF inertial frame. - Lindegren's published global correction: −17 µas. - VLBI says: −38 µas. - Factor-of-2 disagreement between radio and optical astrometry on the correction that removes negative parallaxes. --- # The 100-year constant: a closing table for the presentation Cross-catalogue negative-parallax fraction, compiled from [[Stellar_Parallax]] §"Modern catalogues" and [[Stellar_Notes]] §1, §15: | Catalogue / Source | Year | Negative-parallax fraction | |---|---|---| | Aitken *PASP* review | 1921 | **26%** (negative or below probable error) | | Schlesinger Catalogue | 1924 | ~10% | | Hipparcos main | 1997 | 3.5% (but 1 mas Pleiades systematic survived 20 years) | | Hipparcos 2 (re-reduction) | 2007 | ~3% | | Tycho-1 (ESA, ~2M stars) | 2000 | **25%** | | Gaia DR2 | 2018 | ~20% | | Gaia EDR3 | 2021 | **24%** | | Gaia DR3 post-L21 | 2022 | 17% | > [!critical] The negative fraction has not converged to zero across 100 years and ~1000× precision improvement > Aitken 1921's 26% and Gaia EDR3's 24% bracket the range. Ground-based photographic plates with 20–80 mas individual errors and space-based Gaia with 25 µas faint-end precision give the same order-of-magnitude negative fraction. If negatives were Gaussian-tail noise, the fraction would scale inversely with improving precision. It does not. The signal is not behaving like measurement noise. --- # The direct contradiction with heliocentric geometry From entry 0 (Shack): heliocentric geometry predicts positive-only parallax, in a single rotational sense, always. From entry 28 (Luri DR2 p.4): the Gaia pipeline defines a negative parallax as the measurement showing the source going "the wrong way around on the sky," i.e., a sky motion 180° out of phase with the motion Earth's orbit is supposed to produce. The Aitken 1921 26% and the Gaia EDR3 24% fractions are not a few outliers. They are approximately a quarter of every major catalogue ever published. The mainstream response has taken three mutually incompatible forms: 1. **Dismissal as observational error** (Bessel to Olbers 1816; Newcomb 1878; Dworetsky 2016). 2. **Geometric inversion reading** (Aitken 1921; Lee 1943): negatives are real measurements, the target-vs-comparison assignment is reversed. 3. **Gaussian-tail reading** (Luri 2018): negatives are pure noise with no physical content, only the ensemble mean matters. Readings 2 and 3 cannot both be correct for the same catalogue. Reading 3 fails its own internal consistency test: the Luri 2018 p.6 delete-the-negatives result shows that removing the negatives from the quasar sample shifts the mean from ~0 mas to 0.8 mas, contradicting the cosmological-distance assumption. The negatives are therefore not pure noise even on the mainstream pipeline's own terms. --- # Slide-by-slide running order (one entry per slide) | # | Slide | Entry number | |---|---|---| | 1 | The heliocentric prediction (Shack car-and-windows) | 0 | | 2 | Ptolemy c. 150 AD: no parallax = fixed Earth | 1 | | 3 | Hooke 1674: γ Draconis 30″ wrong direction | 2 | | 4 | Flamsteed 1698: Polaris "parallax" was aberration + nutation | 3 | | 5 | Bradley 1728: aberration discovered because γ Dra was 90° out of phase | 4 | | 6 | Bradley's Polaris: negative parallax surfaced later by Bessel | 5 | | 7 | Lalande 1761–62: Sirius 2.5″ wrong direction | 6 | | 8 | Maskelyne from Lacaille: Sirius 8″ opposite direction | 7 | | 9 | Zach vs Maskelyne: same ten stars, opposite signs (Baily 1845 quote) | 8 | | 10 | Brinkley vs Pond 1810–1824: 14-year dispute, Pond 1822 doctrine | 9 | | 11 | Bessel 1815: first negative reduction of 61 Cygni | 10 | | 12 | Bessel 1815: second negative (μ Cas) | 11 | | 13 | Bessel to Olbers 15 Feb 1816: "observational errors" | 12 | | 14 | Bessel 1816–17: third negative T = −1.32″ (Königsberg Vol. III) | 13 | | 15 | Struve 1837: first correct Vega (0.125″) | 14 | | 16 | Bessel 1838: textbook "first parallax" (fourth reduction) | 15 | | 17 | Henderson 1839: α Cen at 1.16″ (55% off) | 16 | | 18 | Struve 1840: trimmed data doubled the Vega value | 17 | | 19 | Baily 1845: canonical Zach vs Maskelyne footnote | 18 | | 20 | Newcomb 1878: "such a paradoxical result can arise only from errors of observation" | 19 | | 21 | Eichelberger 1911: ~1:1 positive:negative | 20 | | 22 | Washburn 1915: ~1:1 ratio again | 21 | | 23 | Aitken 1921: 26% negative or below probable error in 1013 parallaxes | 22 | | 24 | Lee 1943: three structural causes, seasonal errors ruled out | 23 | | 25 | Vasilevskis 1966: "the reality of some parallaxes were in doubt" | 24 | | 26 | Hipparcos 1997: Pleiades 1 mas systematic lived 20 years | 25 | | 27 | Tycho catalogue: 25% / 29% / 46% breakdown | 26 | | 28 | Goretti 2013: "positive by definition" + RMS objection | 27 | | 29 | Gaia DR2 2018: "going the wrong way around on the sky" | 28 | | 30 | Gaia EDR3 2021: 24% of 1.47 billion sources negative | 29 | | 31 | Lindegren 2021: polynomial correction, own verdict "not possible to derive a definitive recipe" | 30 | | 32 | Bobylev 2026: VLBI vs Gaia factor-of-2 zero-point disagreement | 31 | | 33 | The 100-year constant table: 1921 26% → 2021 24% | Closing table | | 34 | The three incompatible mainstream readings; reading 3 fails on its own terms | Contradiction section | | 35 | Closing: Newcomb 1878 = Dworetsky 2016. Same sentence. 148 years. | — | --- # Source note directory for verification Every primary-source note below contains per-page PNG screenshots at 150 DPI in `Attachments`. ### Pre-Bradley era - [[1981_Williams_Hooke_to_Bessel_PhD_Thesis]] (Hooke, Flamsteed, Bradley, Brinkley-Pond, Bessel chapters) - [[Shack_Chapter25_Negative_Parallax_Demystified]] (compiles Hooke, Flamsteed, Lalande, Maskelyne, Zach, Baily) ### Bradley - [[1728_Bradley_New_Motion_Fixed_Stars]] (the 1728 *Phil. Trans.* paper page-by-page) - [[1798_Bradley_Greenwich_Observations]] (Greenwich archival data behind Bessel's negative Polaris reduction) - [[1838_Bessel_61_Cygni_Parallax]] (Bessel's framing of Bradley's programme) ### Between Bradley and Bessel 1838 - [[Stellar_Parallax]] §"Brinkley vs Pond 1810–1824" (Brinkley-Pond) - [[1817_Bessel_Koenigsberger_Beobachtungen_Vol3]] (third negative reduction of 61 Cygni, primary archival) - [[1852_Bessel_Olbers_Briefwechsel_Polaris_Quote]] (15 February 1816 letter to Olbers, with cropped quote screenshot) - [[1838_Bessel_61_Cygni_Parallax]] (the fourth, positive reduction) - [[2020_Reid_Menten_First_Parallaxes_Revisited]] (Struve trimming, Henderson underestimation) ### Photographic-parallax era - [[1921_Aitken_Recent_Parallax_Review]] (26% negative, inverted-geometry reading) - [[1943_Lee_Negative_Parallax]] (three structural causes) - [[Shack_Chapter25_Negative_Parallax_Demystified]] (Eichelberger 1911, Washburn 1915, Newcomb 1878, Vasilevskis 1966) ### Space-astrometry era - [[2018_Luri_Gaia_DR2_Parallaxes]] (negative as "wrong way around," quasar-deletion bias) - [[2021_BailerJones_EDR3_Distances]] (24% negative in 1.47 billion sources, direct quote) - [[2021_Lindegren_EDR3_Parallax_Bias]] (polynomial bias correction, −150 to +130 µas span) - [[2026_Bobylev_VLBI_Gaia_Zero_Point]] (VLBI vs Gaia factor-of-2 zero-point disagreement) - [[Shack_Chapter25_Negative_Parallax_Demystified]] (Tycho 25/29/46 breakdown) - [[2013_Goretti_Red_Stars_Hipparcos]] (Goretti's triangle argument and RMS objection) - [[2007_vanLeeuwen_Hipparcos_Rereduction]] (Pleiades systematic survived re-reduction) ### Master context notes - [[Stellar_Notes]] full topic-organised notes (42 sources indexed) - [[Stellar_Parallax]] null-hypothesis writeup --- ## See also - [[Stellar_Notes]] master topic-organised notes - [[Stellar_Parallax]] null-hypothesis writeup - [[Stellar_Graphs_Summaries]] plot index for presentation figures