What this page is. The globe side of the ledger, written by this site rather than by the book, and kept structural on purpose: each item is something the model stands or falls on, and nothing here is a fact about one object that no model rests on. It is a supporting page, not one of the twenty-four claims in the catalogue; the claim pages linked from each row are where the tests are run.
1 · Why fourteen, and why these
A model is its structural commitments. For the globe those are the shape and size of the Earth, its two motions, the air it carries and what that air does to light, the arrangement of the Sun and the bodies around it, the distance of the stars, and the two laws — gravitation and momentum — that hold the whole thing together. Take any one away and the model is a different model. Add Pluto’s glaciers or the smell of the space station and nothing changes; those are facts the model is consistent with, not facts it rests on. The book’s hundred and three mixes the two kinds freely, which is why the map sorts them, and why this page carries fourteen.
Each row states the claim with its number, names the observation that tests it and the observation that would falsify it, and says what kit the test takes. from the ground means a garden, a phone, a stopwatch, a tide table or a small telescope. tools, not agencies means the kind of equipment amateur astronomers own and use every week — a camera on a telescope, a season of patience — but not a space agency. Twelve of the fourteen are the first kind; two are the second. None requires taking anyone’s word.
2 · The fourteen
| # | The claim, with its number | The test | What would falsify it | Reader-check |
|---|---|---|---|---|
| 1 | Shape. The Earth is an oblate spheroid: mean radius 6,371 km, equatorial 6,378 km, polar 6,357 km, flattening 1/298. Moritz, “Geodetic Reference System 1980,” Journal of Geodesy 74 (2000) 128 for the ellipsoid constants; the hiding arithmetic on the Reverse Refraction Solver. | Horizon dip grows as √h and hidden height as D²/2R; Polaris’s altitude equals latitude; a degree of latitude is about 111 km everywhere. On this site: Bottom Up Observations, Chicago & Pontchartrain, Celestial Globes Exposed | A horizon that never hides objects bottom-up; Polaris not tracking latitude; a degree of latitude that changes length with position. | from the ground A camera at two known heights over calm water, or a sextant on Polaris and a road atlas. |
| 2 | Spin. The Earth turns once on its axis in 23 h 56 m 4 s (one solar day of 24 h), west to east. IERS Conventions (2010), Technical Note 36, ch. 1, for the rotation rate; the Coriolis drifter data on this site. | Star trails at 15.04° per hour about the celestial pole; the Foucault pendulum precessing at 15° × sin(latitude) per hour; Coriolis on drifting buoys and in every weather model. On this site: The Sky Turns at One Rate, Does the Earth Spin?, Coriolis Drifters | Stars and planets turning at different rates; a pendulum that does not precess; no Coriolis. | from the ground A fixed camera on the night sky for an hour, or a long pendulum watched for one. |
| 3 | An atmosphere held by gravity. Air pressure falls with height by the hydrostatic (barometric) law: about 12 hPa per 100 m at sea level, halving by 5.5 km. U.S. Standard Atmosphere 1976, Table 1; the University of Wyoming sounding archive for any station and date. | Any radiosonde sounding; a phone barometer carried up a hill; the same law predicting the altimeter setting every aircraft uses. On this site: Hydrostatics | A pressure profile that is not hydrostatic; pressure not falling with height. | from the ground A phone barometer and a 100 m hill. |
| 4 | Refraction, computable from the weather. Light through air bends toward denser air. The bend follows the temperature gradient, k = 503·(P/T²)·(0.0342 + dT/dh), is greatest along the horizon, lifts the setting Sun by about 34′ in standard air (k ≈ 0.13–0.17), and cannot exceed what a gradient the air can hold allows: past −34.2 °C/km a layer overturns. Bomford, Geodesy (4th ed., OUP 1980), §3.19 on the refraction coefficient; the Reverse Refraction Solver, which prices any photograph on both surfaces; Ives (1968) on instrumented looming, cited on the Rampion page. | Sunset and moonrise times against the almanac; hidden heights across water against the day’s sounding; the mirage classes matching the measured inversion. On this site: Chicago & Pontchartrain, The Black Swan (Rampion), Mirrored Reflections, Sunlight & Shadows | Refraction that does not follow the measured gradient; the gradients a flat plane needs for the same pictures actually occurring in the air. | from the ground A timed sunset against the almanac, or a photograph across water with the nearest sounding. |
| 5 | Orbit. The Earth orbits the Sun in 365.256 days at a mean distance of 149.6 million km. Bessel’s letter to Herschel on the parallax of 61 Cygni, MNRAS 4 (1838) 152; Bradley, “An account of a new discovered motion of the fix’d stars,” Phil. Trans. 35 (1728) 637; the astronomical unit from Venus radar, Muhleman, AJ 67 (1962) 277. | Stellar aberration of 20.5″, the same for every star, in phase with the orbit; annual parallax of the nearest stars (Proxima 0.77″, 61 Cygni 0.29″); the night sky shifting by 1° a day. On this site: The Sun Does Not Shrink, Sunlight & Shadows | Zero aberration; zero parallax; a sky that does not advance through the year. | tools, not agencies Bessel’s 1838 parallax of 61 Cygni has been repeated by amateurs with a CCD over a season; aberration needs a year of careful astrometry. Tools, not agencies. |
| 6 | Everything orbits the Sun; moons orbit planets. Kepler’s laws hold with one mass at the centre: T² ∝ a³ for every planet about the Sun and for every moon about its planet, with the constant fixed by the central mass. Kepler, Astronomia Nova (1609) and Harmonices Mundi (1619); the CLEA “Revolution of the Moons of Jupiter” exercise for the amateur measurement. | A fortnight of Galilean-moon timings in a small telescope gives Jupiter’s GM to a few per cent; Venus runs a full crescent-to-gibbous phase cycle; retrograde loops fall at opposition. On this site: Jupiter's Shadows, The Eight-Planet Photograph, Stacked & Sharpened | A body whose period breaks T² ∝ a³; Venus without a full phase cycle; retrograde motion at the wrong time. | from the ground A 100-mm telescope, a clock, and two weeks on Jupiter’s moons. |
| 7 | Stars are distant, not infinitely so. The stars are suns at light-years, with small but measured motions: Barnard’s star moves 10.4″ a year; parallaxes are measured to microarcseconds by Gaia; the Plough’s shape changes over tens of thousands of years. Barnard, “A small star with large proper-motion,” AJ 29 (1916) 181; Gaia Collaboration, “Gaia Data Release 3,” A&A 674 (2023) A1. | Two photographs of Barnard’s star a year apart in a 100-mm telescope show the shift; the same spectral lines in starlight as in sunlight. On this site: The Sky Turns at One Rate | Any star at a dome-scale distance (kilometres to thousands of kilometres) measured by parallax; stars with no proper motion at all. | tools, not agencies A small telescope and a camera, two nights a year apart, on Barnard’s star; a grating for the spectrum. |
| 8 | Gas obeys Newton’s third law. Exhaust is mass; momentum is conserved whatever the phase. A rocket’s thrust is the mass rate times the exhaust speed and does not need anything outside to push on. Sutton & Biblarz, Rocket Propulsion Elements (9th ed., Wiley 2017), ch. 2–3, the thrust equation and its pressure term; the in-flight measurements on Thrust, Measured in Flight. | Thrust measured in a vacuum chamber, and in flight above the atmosphere where ambient pressure is zero; thrust that rises as the ambient pressure falls, as the nozzle equation says. On this site: Rockets Don't Push Against Air, Thrust, Measured in Flight, Action Lab Footage | Thrust that falls to zero as the chamber pressure does; a measured engine that pushes only when there is air behind it. | from the ground A balloon cart in a bell jar under a pump, with the question the run must answer set out on the rockets page. |
| 9 | Gravity: proportional to mass, and universal. Weight is proportional to mass everywhere (g = 9.81 m/s² at the surface); the same inverse-square law, with one constant, gives the Moon’s month, every satellite’s period, and two tides a day with springs and neaps on the lunar cycle. g is 9.780 m/s² at the equator and 9.832 at the poles. Moritz (2000), as above, for the normal-gravity formula; Kater’s pendulum, Phil. Trans. 108 (1818) 33; NOAA Tides & Currents for the springs and neaps. | A spring scale against a balance shows weight tracking mass. A seconds pendulum runs about 2 ms per swing slower at the equator than at 60° N, 2.7 ms against the pole — 1,000 swings against a stopwatch shows it — and the 0.53 % pole-to-equator difference decomposes into 0.35 % from the spin and 0.18 % from the bulge, so one pendulum carried north tests claims 1, 2 and 9 together. The fall-off with distance is the part that needs a friend at altitude or the satellite periods. On this site: Hydrostatics, Equator Flight Data | Weight not proportional to mass; the same g at every latitude; a satellite or moon whose period does not match GM/r³; tides not on the lunar cycle. | from the ground A pendulum, a stopwatch and a trip north; a tide table and a lunar calendar. |
| 10 | Axial tilt, and the seasons. The axis is tilted 23.44° to the orbit. The Sun’s noon altitude swings ±23.44° through the year; the midnight sun runs poleward of 66.56° in both hemispheres; day and night are equal everywhere at the equinox. The Astronomical Almanac (USNO/HMNAO), section C, for obliquity and the Sun’s declination; the USNO rise/set tables for the azimuth swing. | A noon shadow measured on the solstices and equinoxes; the sunrise azimuth swinging 47° along the horizon through the year at mid-latitudes; sunrise and sunset tables. On this site: Sunlight & Shadows, The Compass in Antarctica | A southern midnight sun that does not occur; unequal days at the equinox; a noon-Sun swing of other than 46.9°. | from the ground A stick and its noon shadow on four days of the year. |
| 11 | The Sun and the Moon. The Sun is 149.6 million km away and 1.39 million km across, a constant 0.53° wide from sunrise to sunset; the Moon is 384,400 km away and 3,474 km across, a sphere lit by the Sun, its phase set by the Sun–Moon–Earth angle and its terminator a curve only a lit sphere shows. The Astronomical Almanac, sections C and D; the angular-size measurement on The Sun Does Not Shrink and the terminator geometry on Is the Moon a Ball?. | The Sun’s angular size measured through the day with a filter; the Moon’s phase against the Sun’s direction; the lunar distance from a timed occultation seen from two places. On this site: The Sun Does Not Shrink, Is the Moon a Ball?, Full-Moon Lighting, Where's the Moon's Silhouette? | A Sun that visibly shrinks toward the horizon; a phase inconsistent with the Sun’s position; a terminator that is not the edge of a lit sphere. | from the ground A solar filter and a camera at two times of day; a month of Moon photographs with the Sun’s position noted. |
| 12 | Eclipse geometry. A lunar eclipse shows a circular Earth shadow about 2.7 Moon-diameters wide whatever the Moon’s altitude; a solar eclipse’s track of totality is 100–270 km wide and is predicted to the kilometre and the second centuries ahead. Espenak & Meeus, Five Millennium Canon of Lunar Eclipses (NASA TP-2009-214172); the eclipse pages on this site. | The next lunar eclipse, photographed; the next solar eclipse, met at the predicted place and time. On this site: The Selenelion, The Eclipse Trajectory, Where's the Moon's Silhouette? | A non-circular umbra; a shadow edge whose curvature changes with the Moon’s altitude; a predicted eclipse that fails to occur where and when stated. | from the ground A camera at the next lunar eclipse; a map and a clock at the next solar one. |
| 13 | Two hemispheres, symmetric. The southern sky turns clockwise about a southern pole with its own circumpolar stars; Coriolis reverses sign; the Sun crosses the northern sky right to left; both hemispheres see a midnight sun at the same latitude band. The Coriolis drifter analysis on this site, split by hemisphere; any southern-hemisphere star-trail photograph with its location. | A star-trail photograph from the southern hemisphere; a southern sunrise-to-sunset time-lapse; the drifter data by hemisphere. On this site: Coriolis Drifters, Celestial Globes Exposed | A southern sky that behaves like a stretched northern one: no southern pole of rotation, or the same sense of rotation at both poles. | from the ground A camera on the night sky from anywhere south of the equator, or a friend there with one. |
| 14 | Surface geometry is spherical. Distances and directions on the surface follow spherical, not planar, geometry: great circles are the shortest routes, a triangle’s angles sum to more than 180°, and Sydney–Santiago is 11,300 km, flown non-stop in 12–14 hours. Any airline timetable against the great-circle distance (the haversine calculator at Movable Type computes it); Torge & Müller, Geodesy (4th ed., de Gruyter 2012), ch. 2, on spherical excess in triangulation. | Published flight times against great-circle distances on many routes at once, southern routes included; surveyed triangulation networks that close only with spherical excess. On this site: Equator Flight Data, The Compass in Antarctica | A flight time inconsistent with the spherical distance; a large surveyed triangle whose angles sum to exactly 180°. | from the ground A timetable and the haversine formula, for as many routes as patience allows. |
Numbers are from the references in each row and are re-derived in the test suite where arithmetic is involved (the pendulum difference, the centrifugal share, the hidden-height figures). “On this site” links the pages where the book’s own challenge to that claim is answered.
3 · How one of these gets falsified
The right-hand columns are the point of the page. A claim with a number can be wrong by an amount, and a claim with a stated test can fail it. That is what makes the list answerable, and it is also what a falsification has to meet. Each of these fourteen sits on measurements that agree with each other across different instruments, different centuries and different trades that would notice if they failed — the surveyor’s curvature correction, the radio engineer’s earth-bulge term, the airline’s inertial alignment, the eclipse timetable. Against evidence that robust, a falsification cannot be vague. It has to be specific: which measurement is wrong, by how much, and why every other measurement that agrees with it is wrong in the same direction. “Refraction could explain it” is not a falsification of claim 4; a measured gradient that the solver turns into a different picture would be. “Nobody has proven the Sun’s light is straight” is not a falsification of claim 5; a measured aberration of zero would be.
That is the standard this site holds itself to on every claim page — the number, the working, and a “where this could be wrong” section — and it is the standard the book’s closing chapter asks for. The two sides of the ledger can now be read against each other: the book’s three claims and the Rampion observation on the map, and these fourteen here. The difference between them is not who is asking harder questions. It is that each of these fourteen names the observation that would end it.
Where this page could be wrong
- The choice of fourteen is ours. A reader could split gravity into two claims or fold the Sun and Moon into the orbit; the count would move and the structure would not. What should not be arguable is that nothing here is a particular fact dressed as a structural one.
- The reader-check levels are judgements about kit, not about difficulty. Bessel’s parallax is hard to do well; it is still done with tools a hobbyist can own, which is the distinction being drawn.
- The falsifiers are written as clean negatives. A real anomaly would arrive with an error bar, and the question would be whether it is outside the model’s or inside the instrument’s; the claim pages show how that is decided case by case.
- The inverse-square fall-off in claim 9 is marked as needing help for a reason: from one place on the ground it is not measurable, and the page says so rather than pretending a pendulum does it.
Sources & further reading
- Per-claim references are in the table. The ones worth having to hand: Moritz, “Geodetic Reference System 1980” (J. Geod. 74, 2000) for every shape and gravity constant; the Astronomical Almanac for the Sun, Moon and obliquity; the U.S. Standard Atmosphere 1976 for the pressure profile; Sutton & Biblarz for the thrust equation.
- The table is generated from build_globe_claims.py, which holds every claim, test, falsifier and reference in one place.
- The book’s own list and its scoring rule: The Optimized Debate, Mapped; the chapter it comes from: Extraordinary Evidence.