On this page
Your geography or physics textbook says the mantle ends about 2,900 km down. Nobody has drilled anywhere near that deep, so the number had to be worked out from earthquake waves. The person who first pinned it down precisely was Beno Gutenberg. This post explains how, gives you a simple geometry estimate to try, and shows why your estimate comes out wrong in an instructive way.
Who was Beno Gutenberg?
The National Academy of Sciences biographical memoir by Leon Knopoff gives his dates as 4 June 1889 to 25 January 1960, and calls him "the foremost observational seismologist of the twentieth century". He was born in Darmstadt, Germany, and studied geophysics at Göttingen under Emil Wiechert, earning his doctorate in 1911.
After the First World War he worked for years in his family's soap factory, doing seismology in his spare time, and lectured at Frankfurt. In 1930 he moved to the California Institute of Technology, and the Seismological Laboratory's rapid growth began with his arrival. There he worked with Charles Richter on magnitude scales and on the statistics of earthquakes.
Why was the depth of the core uncertain?
Scientists already suspected a core, but disagreed badly about its size. The memoir notes that Wiechert proposed in 1897 that an iron core began about 1,400 km down, while R. D. Oldham in 1906 put its top near 3,900 km. A difference of 2,500 km is not a detail; it changes the core's volume several times over.
How did Gutenberg measure it?
At Göttingen, Gutenberg studied records of distant earthquakes. Knopoff describes how he noticed that a wave called P′ grew stronger at about 143° from the earthquake, and interpreted the gap between about 95° and 143° as a shadow cast by a low-velocity core deep inside the Earth.
He then built new, accurate travel-time curves for P and S waves at large distances and calculated the times of waves reflected from and refracted through the core's surface. From these he put the top of the core at 2,900 km and showed that the boundary is sharp. That was in 1914.
The memoir calls the precision "astounding". More than twenty years later, Gutenberg and Richter used waves reflected off the core and got the same depth. In 1939 Harold Jeffreys found 2,898 ± 3 km.
The NCERT Class 11 chapter Interior of the Earth (2026-27 reprint) still says the mantle "extends from Moho's discontinuity to a depth of 2,900 km". It also gives the shadow zone as 105° to 145° for P-waves and everything beyond 105° for S-waves. Our post on Inge Lehmann and the inner core explains how the waves seen inside that zone led to the next discovery.
Worked example: estimate the core from the shadow zone
Here is a deliberately simple model. Assume waves travel in straight lines. A wave that just grazes the core is a chord of the Earth that touches the core's surface at its midpoint. If that chord reaches the surface at an angular distance Δ from the earthquake, the geometry of the right triangle at the midpoint gives:
core radius = R × cos(Δ/2)
With NCERT's radius R = 6,378 km and the shadow zone starting at Δ = 105°:
Δ/2 = 52.5°, and cos 52.5° = 0.609
core radius = 6,378 × 0.609 ≈ 3,883 km
depth to core = 6,378 − 3,883 ≈ 2,495 km
That is about 400 km shallower than Gutenberg's 2,900 km. Why?
Real waves are not straight. Wave speed in the mantle generally rises with depth, so rays bend and curve back up towards the surface. A curved grazing ray reaches the surface sooner than a straight one would, so a given shadow edge corresponds to a smaller, deeper core. Gutenberg handled this properly by using measured travel times at many distances rather than straight lines.
One more check: with the core's top at 2,900 km, the core radius is 6,378 − 2,900 = 3,478 km, about 0.545 of the Earth's radius. Volume scales as radius cubed, so the core is about 0.545³ ≈ 0.16, roughly 16 percent of Earth's volume.
What mistakes do students make with this topic?
Mixing up the shadow zones. S-waves vanish beyond 105° because they cannot pass through the liquid outer core. The P-wave shadow is a band, because P-waves are refracted.
Using arc length where the geometry needs an angle. The core estimate uses the angle Δ at Earth's centre, not the surface distance in km.
Trusting a simple model too far. A straight-ray model is a good first step, but always ask which assumption makes it wrong.
If refraction at a boundary is hard to picture, Fermi's Simulation Lab has interactive simulations for ideas that are easier to see than read, which can help before you return to the ray diagram.
What can students learn from Gutenberg's way of working?
Gutenberg's strength was patient work with records: collecting travel times, building better curves, and only then drawing conclusions. He also kept working on science through years when it was not his job. And his result held up because he tested it against several kinds of waves, not just one.
Next time you meet a textbook number, ask how it could have been measured, then try a rough model like the one above and see how close you get.
