…but how do we know?
Last checked 1 November 2026 · every number in the episode has a row below
The percentages are our editorial estimate of how settled each answer is, not a measurement or a statistic. The ranges are the sources' own.
Earthquake waves: S waves vanish beyond about 104° from a quake, P waves leave a shadow from 104° to 140°, and faint P waves turn up inside it. The whole planet rings after big quakes. Average density 5,513 kg/m³; moment of inertia 0.3308.
A liquid outer core from about 2,890 km down and a solid inner core from about 5,150 km, known to within a few kilometres; the core holds 16 % of the volume and 32 % of the mass; roughly 85 % iron and 5 % nickel by weight.
Which light elements, and how much. The temperature at the inner core's surface (published estimates 4,800–6,300 K). Whether the changing waves mean a turning inner core, a reshaping surface, or both.
| Row | Move | Evidence that moved it | Rows |
|---|---|---|---|
| What's there | ? → 5 % | The deepest hole: 12,262 m, 0.19 % of the way. Less than a millimetre on a globe one metre across. | L1, L6 |
| What's there | 5 → 40 % | The planet is about twice as dense as its crust, and its moment of inertia (0.3308) is below a uniform ball's 0.4: mass is crowded toward the centre. | L7–L10 |
| What's there | 40 → 70 % | Oldham 1906: waves through the deepest part arrive late. Gutenberg: the core starts about 2,900 km down. Missing S waves beyond about 104°; Jeffreys 1926: the core is liquid. | L11–L14 |
| What's there | 70 → 82 % | Lehmann 1936 proposes an inner core from faint P waves in the shadow; Dziewonski & Gilbert 1971: the planet's ringing after the 1964 Alaska quake fits only a solid inner core. | L15, L16 |
| What's there | 82 → 92 % | Two independent Earth models (PREM 1981, ak135 1995) put the boundaries within 0.5 km and 4.0 km of each other. | L17–L20 |
| Made of | ? → 40 % | Density jumps from 5.57 to 9.90 g/cm³ at the core and reaches about 13 at the centre: about eleven times water on average, which says metal. | L21, L22 |
| Made of | 40 → 65 % | Meteorites vs Earth's rocks: the missing iron and nickel; but the core is 6–8 % (outer) and 4–5 % (inner) lighter than pure iron, so something lighter is mixed in. | L30, L31 |
| How it moves | ? → 30 % | Song & Richards 1996: waves through the inner core changed over three decades; about 1° a year faster than the mantle. | L23 |
| How it moves | 30 → 45 % | Wen 2006 (surface moved outward by 1–1.75 km); Wang & Vidale 2022 (about 0.1° a year); Yang & Song 2023 (recently paused); Wang et al. 2024 (ahead 2003–2008, back 2008–2023, relative to the mantle). | L24–L27, L41 |
| How it moves | 45 → 35 % | Vidale et al. 2025: changes near the surface that rotation can't explain; the title says both. A 2026 model ties the swings to the length of day but uses the turning as input, so the row stays. | L27, L28, L40, L42 |
| How hot | ? → 45 % | The inner core's surface is freezing alloy, so its temperature is a melting point at 330 GPa. ESRF 2013: pure iron melts at 6,230 ± 500 K; with light elements about 5,500 K. | L32, L33 |
| How hot | 45 → 60 % | 2024 laser shock: 5,345 K at 240 GPa, which caps the inner core's surface at 6,202 K. Published estimates still span 4,800–6,300 K. | L34, L35 |
R. D. Oldham, “The Constitution of the Interior of the Earth, as Revealed by Earthquakes”, Quarterly Journal of the Geological Society 62 (1906), p. 456
beyond these facts little can be said to be known
his core: about 0.4 of the radius → today 0.55
K. Hirose, B. Wood & L. Vočadlo, “Light elements in the Earth's core”, Nature Reviews Earth & Environment 2 (2021)
The exact composition of the core remains unknown.
Oldham 1906 is in the public domain; the episode shows the sentence typeset from the GeoSphere Austria scan. Other papers appear as typeset title cards, not as scans.
| # | Claim | Value | Tag | Source · where |
|---|---|---|---|---|
| L1 | The deepest hole ever drilled (vertical depth) is 12,262 m, on the Kola Peninsula in the far north of Russia; that depth was reached in 1990 and has not been surpassed. | 12,262 m; 1990; 69°25′ N, 30°44′ E | OBSERVED | Lobanov, Chicherov & Sharov, "The 50th Anniversary of the Start of Drilling the Kola Superdeep Well", Arctic and North no. 44 (2021), DOI 10.37482/issn2221-2698.2021.44.267 — arcticandnorth.ru/upload/iblock/61f/44_224_… PDF header p. 224 (Abstract, Introduction: "the world's deepest scientific well"); p. 228 (location); p. 231 ("In 1990, the maximum depth was reached — 12262 m, which has not been surpassed so far"); p. 230 ("As of 05/01/1991, the well depth was 12262 m"). Journal licence: CC BY-SA (landing page), so typeset card only. Note: widely read secondary sources give 1989 / ended 1995 / 180 °C; we follow Lobanov (primary). Longer deviated wells (Z-44 Chayvo, 12,376 m total depth, 2012) are longer, not deeper. |
| L2 | The standard picture shows crust, mantle, a liquid outer core and a solid inner core. | — | OBSERVED | USGS Earthquake Science Center, "Crust, Mantle, and Core of the Earth" (Public Domain) — usgs.gov/media/images/crust-mantle-and-core… caption: "core (liquid in light gray, solid in dark gray)" |
| L3 | The centre of the Earth is about 6,371 km down. | 6,371.000 km (volumetric mean radius) | OBSERVED | NASA NSSDCA, Earth Fact Sheet — nssdc.gsfc.nasa.gov/planetary/factsheet/ear… Bulk parameters, "Volumetric mean radius (km)" |
| L4 | Drilling began in 1970 and stopped in 1992. | 24 May 1970; 1992 | OBSERVED | Lobanov et al. 2021 (as L1) p. 228 ("Drilling was started on May 24, 1970"); p. 231 ("Drilling was stopped in 1992") |
| L5 | At the bottom the rock was 212 °C; the plan had assumed about 120 °C. | 212 °C at ≈ 12 km (180 °C at 10 km); assumed 120 °C | OBSERVED | Lobanov et al. 2021 (as L1) p. 235 ("at a depth of about 12 km, the temperature values reached 212°С, instead of the assumed 120°С"); repeated in Conclusion p. 237 |
| L6 | 12.262 km out of 6,371 km is 0.19% of the way to the centre; on a globe 1 m across, less than a millimetre. | 0.19 %; 0.96 mm | INFERRED | L1 + L3 — |
| L7 | Divide the Earth's mass by its volume: average density 5,513 kg/m³, five and a half times water. | 5,513 kg/m³ = 5.51 × water | OBSERVED INFERRED | NASA Earth Fact Sheet (as L3) Bulk parameters, "Mean density (kg/m3) 5513" |
| L8 | The crust, in the standard Earth model, is 2.6 to 2.9 times water; the planet is about twice as dense as its own skin. | 2.600 and 2.900 g/cm³; ratio 1.9–2.1 | OBSERVED INFERRED | Dziewonski & Anderson, "Preliminary reference Earth model", PEPI 25 (1981) 297–356, DOI 10.1016/0031-9201(81)90046-7 — PDF lweb.cfa.harvard.edu/~lzeng/papers/PREM.pdf Table I, p. 308, rows "Crust 6346.6–6356.0: 2.900" and "6356.0–6368.0: 2.600" |
| L9 | Mass near the middle turns easily. A uniform ball scores 0.4, a hollow shell about 0.67; the Earth scores 0.3308. Lower than uniform: mass crowded toward the centre. | I/MR² = 0.3308 (Earth); 0.4; 0.667 | OBSERVED INFERRED | NASA Earth Fact Sheet (as L3); textbook formulas for a uniform solid sphere (2/5) and thin spherical shell (2/3) Bulk parameters, "Moment of inertia (I/MR2) 0.3308" |
| L10 | The Earth's score is measured from the slow circling of its spin axis (precession) plus satellite tracking. | method; (source's own value: polar C/MR² = 0.3307007, not shown on screen) | OBSERVED | J. G. Williams, "Contributions to the Earth's obliquity rate, precession, and nutation", Astronomical Journal 108 (1994) — ntrs.nasa.gov/citations/19950028691 Abstract: precession rate gives (C−A)/C, "in combination with a satellite-derived J2" gives C/MR² |
| L11 | In 1906 Richard Oldham summed up what was known (size, average density, denser toward the centre, and hot), then wrote "beyond these facts little can be said to be known". Waves through the deepest part arrived late: a core in roughly the central four-tenths of the radius. The size was off; it's more than half. | 0.4 of radius (Oldham); 0.546 (today, PREM) | OBSERVED INFERRED | R. D. Oldham, "The Constitution of the Interior of the Earth, as Revealed by Earthquakes", Quarterly Journal of the Geological Society 62 (1906) 456–475, DOI 10.1144/GSL.JGS.1906.062.01-04.21 — scan opac.geologie.ac.at/ais312/dokumente/Oldham… p. 456 (Introductory: mean density "about 5½ times that of water", "beyond these facts little can be said to be known"); p. 472 (conclusion: "central four-tenths of the radius", first-phase waves "slightly reduced", second-phase "not transmitted at all, or … at about half the rate") |
| L12 | Within a few years, Beno Gutenberg put the top of the core about 2,900 km down. | ≈ 2,900 km (Brush); 2,890 km in 1911/12 (EGU) | OBSERVED INFERRED | S. G. Brush, "Discovery of the Earth's core", American Journal of Physics 48 (1980) 705–724, DOI 10.1119/1.12026 — ui.adsabs.harvard.edu/abs/1980AmJPh..48..70… ; EGU, "Beno Gutenberg" — egu.eu/egs/gutenberg.htm Brush abstract ("Beno Gutenberg established the depth of this boundary as 2900 km"); EGU page ("In 1911/12 … determined the core-mantle boundary of the Earth at a depth of 2890 km") |
| L13 | P waves push and pull, through solids and liquids; S waves shake sideways and only travel through solids. More than about 104° around the globe from an earthquake, over a quarter of the way round, the S waves never arrive: a liquid stops them. P waves are missing from 104° to 140°. | S shadow > ~104° (~103° in text); P shadow 104°–140° | OBSERVED INFERRED | IRIS (EarthScope Consortium), Animation 121 "Seismic Shadow Zone: Basic Introduction" — iris.edu/hq/inclass/animation/121 (CC BY 4.0) Description and animation text ("S waves are stopped by the liquid outer core"; "not detected beyond 104°"; "not detected between 104° and 140°") |
| L14 | Missing waves alone didn't settle it; in 1926 Harold Jeffreys knocked down the case for a solid core. | 1926 | OBSERVED | Brush 1980 (as L12) Abstract ("failure to detect propagation of S waves through the core was not sufficient evidence"; "Not until 1926 did Harold Jeffreys refute the arguments for solidity and establish that the core is liquid") |
| L15 | Faint P waves turned up in the shadow. In 1936 the Danish geophysicist Inge Lehmann, studying a big 1929 earthquake in New Zealand, proposed a smaller core inside the core. Her paper was titled "P′". | 1929; 1936 | OBSERVED | APS News, "This Month in Physics History" (L. Boatman, 10 Aug 2023) — aps.org/apsnews/2023/08/seismologist-lehman… ; EARTH Magazine, A. Mills, "March 29, 1936: Notes on Earth's Inner Core" (30 Jan 2015) — earthmagazine.org/article/march-29-1936-not… APS: P waves "few and faint" in shadow zones; 1929 New Zealand earthquake; "in September 1936 … published a paper, succinctly called 'P′'"; "a core nested within a core". EARTH: "the 1929 earthquake in Buller, New Zealand" |
| L16 | In 1971 Adam Dziewonski and Freeman Gilbert found the first direct evidence that the inner core is solid, from the way the whole planet rang after the 1964 Alaska earthquake; only a solid inner core fitted. | 1971; 1964 | OBSERVED | Dziewonski & Gilbert, "Solidity of the Inner Core of the Earth inferred from Normal Mode Observations", Nature 234, 465–466 (24 Dec 1971), DOI 10.1038/234465a0 — nature.com/articles/234465a0 ; Nature 234, 438 (1971) "The Earth's Cores" — nature.com/articles/234438a0.pdf ; B. Romanowicz, "Adam M. Dziewonski (1936–2016)", Eos 97 (2016) — eos.org/articles/adam-m-dziewonski-1936-2016 Nature title + abstract; News & Views ("normal modes … excited by earthquakes"; "no model with a liquid inner core can satisfy all the data"); Eos (normal-mode data "from the 1964 Alaska earthquake", "first direct proof … (1971)") |
| L17 | PREM (1981): the core starts at radius 3,480.0 km, the inner core at 1,221.5 km — about 2,891 km and 5,150 km down. | 3,480.0 km; 1,221.5 km; depths 2,891.0 km, 5,149.5 km | OBSERVED INFERRED | PREM 1981 (as L8) Table I, p. 308 (region radii; x = r/a, a = 6371 km) |
| L18 | ak135 (1995) puts the top of the core within half a kilometre of PREM, the inner core within four. | CMB depth 2,891.50 km; ICB depth 5,153.50 km → r = 3,479.5 and 1,217.5 km; Δ 0.5 km, 4.0 km | OBSERVED INFERRED | AusPass, "AK135 Travel Time Tables" / model ak135-F (Kennett, Engdahl & Buland, GJI 122 (1995) 108–124, DOI 10.1111/j.1365-246X.1995.tb03540.x) — auspass.edu.au/research/ak135/ak135f.html model table, rows at depth 2891.50 and 5153.50 |
| L19 | The 1981 model was built to fit the Earth's mass and spin score, so it matches them by design. | radius 6371 km; mass 5.974 × 10²⁴ kg; I/MR² 0.3308 | OBSERVED | PREM 1981 (as L8) §3 "The gross Earth data set", p. 300 ("were used as constraints") |
| L20 | Its boundaries come from earthquakes: about a thousand ringing tones and 1.75 million wave arrival times. | ≈ 1000 normal-mode periods; 500 summary travel times + 1.75 × 10⁶ ISC travel times | OBSERVED | PREM 1981 (as L8) Abstract, p. 297 |
| L21 | In that model, a sixth of the volume holds a third of the mass: about eleven times water. Not rock. | volume 16.3 %; mass 32.5 % (PREM, our integration; Hirose et al. 2021 give 32.3 %); mean core density 10.99 g/cm³ | INFERRED | PREM 1981 Table I (as L8); cross-check Hirose et al. 2021 (as L30) ms p. 3 ("a core that constitutes 32.3% of the mass of the Earth") Table I density polynomials |
| L22 | At the core's edge the density jumps from about five and a half to almost ten. | CMB 5.57 → 9.90 g/cm³; ICB 12.17 → 12.76 g/cm³; centre 13.09 g/cm³ (PREM). ak135: 5.7721 → 9.9145; 12.1391 → 12.7037 | OBSERVED INFERRED (PREM) (ak135) | PREM 1981 Table I (as L8); ak135 table (as L18) PREM: polynomials for lower mantle, outer core, inner core; ak135: rows at 2891.50 and 5153.50 (densities from ak135-F, density model credited to Montagner & Kennett; boundary depths are ak135) |
| L23 | In 1996 Song and Richards found waves crossing the inner core had changed slowly over three decades; their explanation: it spins about one degree a year faster than the mantle and crust. | "on the order of 1° per year" | OBSERVED | X. Song & P. G. Richards, "Seismological evidence for differential rotation of the Earth's inner core", Nature 382, 221–224 (18 Jul 1996) — nature.com/articles/382221a0 Abstract |
| L24 | In 2006 Lianxing Wen compared two earthquakes from the same spot, nearly ten years apart (9.76 yr); waves bouncing off the inner core came back 40 to 70 ms early (at least 39–70 ms); beneath middle Africa its surface had moved outward by 0.98–1.75 km. Either a bump rotating into place or the surface growing. | 39–70 ms; 0.98–1.75 km; 1 Dec 1993 and 6 Sep 2003 | OBSERVED INFERRED | L. Wen, "Localized temporal change of the Earth's inner core boundary", Science 314, 967–970 (10 Nov 2006), DOI 10.1126/science.1131692 — researchconnect.stonybrook.edu/en/publicati… Abstract (doublet, "at least from 39 to 70 milliseconds earlier", radius "enlarged locally beneath middle Africa by 0.98 to 1.75 kilometers"; differential motion or rapid growth) |
| L25 | A 2022 study of old nuclear test explosions: the inner core slipped back from 1969 to 1971, then ran ahead about three tenths of a degree by 1974 — roughly a tenth of a degree a year. | ≥ 0.1° (1969–71, back); ~0.29° (1971–74, ahead); ≈ 0.1°/yr; model period 6 yr | OBSERVED INFERRED | W. Wang & J. E. Vidale, "Seismological observation of Earth's oscillating inner core", Science Advances 8 (10 Jun 2022), DOI 10.1126/sciadv.abm9916, PMC9187225 — doi.org/10.1126/sciadv.abm9916 (PMC9187225) Abstract ("superrotation of ~0.29° from 1971 to 1974"; "The model, which has a 6-year period") |
| L26 | In 2023 Song's own group said the extra spin had recently paused, part of a swing of about seventy years (the 2022 model swings every six). | ≈ 7-decade oscillation, turning points ~2009–2011 and early 1970s; vs 6-yr period (L25) | OBSERVED | Y. Yang & X. Song, "Multidecadal variation of the Earth's inner-core rotation", Nature Geoscience 16, 182–187 (23 Jan 2023) — nature.com/articles/s41561-022-01112-z Abstract ("differential inner-core rotation has recently paused"; "an approximately seven-decade oscillation, with another turning point in the early 1970s") |
| L27 | In 2024, 121 repeating earthquakes showed the inner core running ahead from 2003 to 2008, then drifting back along the same path, two to three times slower. Relative to the mantle; it never stopped turning with the planet. The paper says its data can't resolve changes at the inner core's surface. | 121 events, 1991–2023; turn 2008.5 ± 0.18; 2–3× slower | OBSERVED INFERRED | W. Wang, J. E. Vidale et al., "Inner core backtracking by seismic waveform change reversals", Nature 631, 340–343 (12 Jun 2024), DOI 10.1038/s41586-024-07536-4, CC BY 4.0 — nature.com/articles/s41586-024-07536-4 Abstract ("gradually super-rotated from 2003 to 2008", then "two to three times more slowly back through the same path"); Results (121 events, 143 pairs, 1991–2023); Methods (2008.5 ± 0.18; "about 2.5 times slower"); main text just before Methods, PMC11236701 ("Our data do not resolve changes at the IC boundary or in the OC") |
| L28 | In 2025 the same group found changes near the surface that rotation can't explain; their title says both: the rotation rate and the surface. | non-rotational changes at YKA, 2004–2008 | OBSERVED UNKNOWN (: rotation vs. reshaping) | J. E. Vidale, W. Wang, R. Wang, G. Pang & K. Koper, "Annual-scale variability in both the rotation rate and near surface of Earth's inner core", Nature Geoscience 18, 267–272 (10 Feb 2025), DOI 10.1038/s41561-025-01642-2 — nature.com/articles/s41561-025-01642-2 Title (Crossref); abstract via secondary coverage ("non-rotational changes at YKA but not ILAR between 2004 and 2008"). The phrase "tentatively both" could not be verified word for word (nature.com blocked), so it is not quoted. |
| L29 | Kitchen Check (animated EXAMPLE, not filmed): spin a raw and a hard-boiled egg, stop each with a fingertip and let go; what you should see: the boiled egg stays stopped, the raw egg starts turning again on its own. NASA's lesson: the finger stops the shell, not the liquid inside, and the moving liquid can drag the shell back into motion. NASA JPL uses the same test in a lesson on how InSight would check Mars for a liquid core from its wobble. | — | OBSERVED EXAMPLE (method and expected result) | NASA/JPL Education, "Planetary Egg Wobble and Newton's First Law" — jpl.nasa.gov/edu/teach/activity/planetary-e… procedure ("momentarily place a finger on top of the egg to stop the spin, then release"; "the liquid center of a raw egg will continue spinning"); InSight/RISE section |
| L30 | Primitive meteorites are thought to be the raw material of planets; the Earth's rocks are short of iron and nickel; put the missing metal in the core: about 85% iron, 5% nickel. | ~85 wt% Fe; 5 wt% Ni | OBSERVED | K. Hirose, B. Wood & L. Vočadlo, "Light elements in the Earth's core", Nature Reviews Earth & Environment 2, 645–658 (2021), DOI 10.1038/s43017-021-00203-6 — accepted ms: discovery.ucl.ac.uk/10134605/1/Vocadlo_Hiro… ms p. 3 ("CI chondrites are thought to represent the basic planet-building material"; "~85 wt% Fe and 5 wt% Ni"); ms p. 7 (Fe and Ni depleted in silicate Earth; mass balance) |
| L31 | Pure iron would be too heavy: the liquid outer core is 6–8% lighter than pure iron would be down there, the inner core 4–5%. Something lighter is mixed in: some combination of sulfur, silicon, oxygen, carbon and hydrogen. "The exact composition of the core remains unknown." | 6–8 %; 4–5 % | OBSERVED UNKNOWN (: which light elements) | Hirose et al. 2021 (as L30) ms p. 1 (abstract: light elements S, Si, O, C, H; quote); ms p. 4 ("6–8% density deficit"); ms p. 5 ("less dense by 4–5% than pure Fe") |
| L32 | The inner core's surface is where liquid alloy is freezing, so its temperature is that alloy's freezing point at 330 GPa, 3.3 million atmospheres. | 330 GPa = 3.3 million atm | OBSERVED | Hirose et al. 2021 (as L30); ESRF Highlights 2013 (as L33) Hirose ms p. 10 ("TICB corresponds to the liquidus temperature … of an outer core liquid at 330 GPa"); ESRF: "330 GPa (3.3 million atmospheres)" |
| L33 | In 2013 a team at the European synchrotron squeezed iron between diamonds, laser-heated it and watched it melt with X-rays: 6,230 ± 500 K, a thousand degrees above an experiment twenty years earlier; allowing for light elements, about 5,500 K at the inner core's surface. | 6,230 ± 500 K (pure Fe, 330 GPa); ≈ 5,500 K at ICB | OBSERVED | ESRF, "The Earth's centre is 1000 degrees hotter than previously thought" (Highlights 2013) — esrf.fr/home/UsersAndScience/Publications/H… ; paper: Anzellini et al., Science 340, 464–466 (2013), DOI 10.1126/science.1233514 ESRF text ("finding a temperature of 6230 +/- 500 K"; "1000 degrees hotter than in a previous experiment run 20 years ago"; light elements lower the liquidus by ~700 ± 200 K, "a temperature around 5500 K at the inner core boundary"); figure caption (X-ray diffraction) |
| L34 | In 2024, a laser shock instead of a slow squeeze: iron melted at 5,345 K at 240 GPa, which caps the inner core's surface at 6,202 K. | 5,345 K at 240 GPa; ≤ 6,202 K at ICB | OBSERVED | APS Physics, "Taking the Temperature of Earth's Core" (Synopsis, v17 s139) — physics.aps.org/articles/v17/s139 ; paper: Balugani et al., Phys. Rev. Lett. 133, 254101 (2024), DOI 10.1103/PhysRevLett.133.254101 Synopsis text (laser compression + X-ray absorption; melting at 240 GPa, 5345 K; "must be no higher than" 6202 K) |
| L35 | Published estimates still span 4,800 to 6,300 K, because the recipe sets the melting point. | 4,800–6,300 K (span 1,500 K) | OBSERVED INFERRED UNKNOWN (: ICB temperature) | Hirose et al. 2021 (as L30) ms p. 10 ("A wide range of previous TICB estimates (from 4,800 K … to 6,300 K) includes large uncertainties derived from unknown light element composition"; liquidus "depends largely on its light element composition") |
| L36 | A liquid outer core from about 2,890 km down, a solid inner core from about 5,150 km, known to within a few kilometres. | ≈ 2,890 km; ≈ 5,150 km; models differ by 0.5 and 4 km | INFERRED | L17 + L18 — |
| L37 | About 5,500 K is near the temperature of the Sun's visible surface. | Sun effective temperature 5,772 K (photosphere 4,400–6,600 K) | OBSERVED INFERRED | NASA NSSDCA, Sun Fact Sheet — nssdc.gsfc.nasa.gov/planetary/factsheet/sun… Solar atmosphere: "Effective temperature: 5772 K" |
| L38 | Which lighter elements are in the core, and how much of each, nobody knows; several mixtures fit. | outer core range: Fe + 5 % Ni + 1.7 % S + 0–4.0 % Si + 0.8–5.3 % O + 0.2 % C + 0–0.26 % H (wt %) | UNKNOWN | Hirose et al. 2021 (as L30) ms p. 1 (abstract; "The exact composition of the core remains unknown") |
| L39 | The temperature at the inner core's surface is not pinned down: give or take several hundred kelvin. | estimates 4,800–6,300 K | UNKNOWN | Hirose et al. 2021 (as L30) ms p. 10 |
| L40 | Whether the changing waves mean a turning inner core, a reshaping surface, or both is an open argument. | — | UNKNOWN | Vidale et al. 2025 (as L28: title, rotation rate and near surface); Wang et al. 2024 (as L27: data "do not resolve changes at the IC boundary") abstracts / main text |
| L41 | The estimated extra spin has settled to around a tenth of a degree a year (the estimated drift). | 0.05–0.15° per year | OBSERVED | Wang et al. 2024 (as L27), PMC11236701 Introduction ("The inferred rate of super-rotation has settled to about 0.05–0.15° per year") |
| L42 | This September, a model in Nature tied the inner core's back-and-forth to the length of our day: changes of a few milliseconds, over decades. It's built on the turning, so it can't prove it. | a few ms; multidecadal | OBSERVED (model) | Zhang & Dumberry, "Gravitational torque drives multidecadal variations in length of day", Nature (published 23 Sep 2026), DOI 10.1038/s41586-026-10999-2 — doi.org/10.1038/s41586-026-10999-2 ; ScienceDaily, 29 Sep 2026 — sciencedaily.com/releases/2026/09/260928100… abstract / press summary: inner-core oscillation coupled to the mantle by gravitational torque reproduces multidecadal length-of-day changes of a few milliseconds; uses seismic reconstructions of inner-core rotation as input |
One raw egg, one hard-boiled. Spin each on a table, stop it with a fingertip for a moment, and let go. The boiled egg stays stopped. The raw egg starts turning again on its own: the finger stopped the shell, not the liquid inside, and the moving liquid drags the shell back into motion.
In the episode this is an animated example, labelled EXAMPLE on screen, not footage of a test. It shows what NASA/JPL's lesson says you should see. The limits: this is not how we know Earth's outer core is liquid (that took earthquake waves), the yolk is not a model of the inner core, and the estimated drift of the inner core is about a tenth of a degree a year. Try it with yours and tell us what you got.
None yet. Any change to a number in the episode is listed here with its date and reason.
Note: many secondary sources give 1989 for Kola's deepest point, 1995 for the end of drilling and 180 °C at the bottom. We follow the primary account (Lobanov et al. 2021): 12,262 m reached in 1990, drilling stopped in 1992, 212 °C at about 12 km (180 °C at 10 km).
Sure So Far · how we know what we know, and where knowing stops. The Certainty Meter is our editorial estimate, not a measurement.