How a librarian measured the whole Earth with two sticks around 240 BC, what that single proof can and cannot show, and why every modern repeat across many latitudes still lands on a globe.
Around 240 BC a scholar in Egypt named Eratosthenes measured the size of the whole Earth with two sticks and one distance. At noon on the longest day of the year, the Sun stood straight over the city of Syene. It lit the bottom of a deep well, and an upright pole threw no shadow. On that same day, 500 miles north in Alexandria, an upright pole did throw a shadow. The shadow placed the Sun about 7.2 degrees away from straight up. Since 7.2 degrees is one fiftieth of a full circle, the road from Syene to Alexandria is one fiftieth of the way around the Earth. That makes the Earth about fifty times 500 miles around. His result came within a few percent of the value we use today.
The shadow angle at Alexandria equals the angle between the two cities at the center of the Earth. Measure the ground distance, find what fraction of a circle that angle is, and you have the whole circumference.
A flat-Earth reader can accept every step above and still ask a fair question. Why does a shadow difference prove a ball, rather than a flat map lit by a nearby Sun? The honest reply is that one measurement, on its own, does not. The proof appears in the shape of the pattern once you add more sticks in more places.
Three different worlds are on the table, and each predicts something you can check:
So you measure a third city, and a fourth, spread far apart, and you read the shape. A straight climb means a round Earth and a far Sun. A climb that bends and flattens means a nearby Sun over flat ground. Across the real world, from the equator toward the poles, it climbs in a straight line.
The three worlds predict three different shapes. Real shadow angles, measured from many places, land on the straight line and stay there. Both flat cases fail: a far Sun predicts no shadow at all, and a near Sun predicts a curve that bends away from the data.
One honest catch deserves plain statement. Over a short hop, like Syene to Alexandria, the straight line and the arctangent curve sit almost on top of each other. Two nearby cities cannot separate them. You need points spread across a wide band of latitude before the curve pulls away from the line. Eratosthenes had two cities, so his own measurement, taken alone, does not shut the nearby-Sun door. Measurements across whole continents do.
This is no longer a story about one scholar and two cities. Every year, on the two equinoxes, large groups repeat the measurement together across the whole planet. The Eratosthenes Experiment, organized in Greece by Ellinogermaniki Agogi and the Hellenic Mathematical Society, runs twice a year and posts its data in the open. In its March 2024 round, 51 schools from 29 countries took part, pooling shadow readings across many latitudes and both hemispheres. Online the reach is wider still. On the June 2025 solstice, the science channel SciManDan asked viewers to measure the shadow of a vertical object and report their latitude, the height of the object, and the length of the shadow. In all, 1,013 people from both hemispheres responded. Plotted against latitude, their readings matched a round Earth and not a flat one, and three points along one meridian rebuilt the circumference to within 4 percent. These are not the only groups doing it. The European Association for Astronomy Education has run the measurement with schools across Europe and South America year after year, and a Greek national program repeats it across that country each equinox. Each time this is run, the round-Earth pattern appears and the nearby-Sun result does not. Hundreds of strangers, scattered over thousands of miles, cannot all coordinate the same mistake, and the answer comes out round every single time.
Sources. The Eratosthenes Experiment, organized by Ellinogermaniki Agogi with the Hellenic Mathematical Society, holds its collaborative measurement on both equinoxes and publishes participant data: eratosthenes.ea.gr (the March 2024 round drew 51 schools from 29 countries). The European Association for Astronomy Education has organized the measurement with schools across Europe and South America, annually since 2015: eaae-astronomy.org (its June 2011 European round reconstructed the meridional circumference to within 0.36 percent). In Greece, a national program run by PANEKFE with the Institute for Astronomy of the National Observatory of Athens and the Greek ESERO office repeats it across the country each equinox: greekreporter.com (March 2026). The 2025 crowd measurement was run on the June solstice by the channel SciManDan: youtube.com/watch?v=JSSaMNktHbU, and independently documented by science educator Dr. Jay Wile, a participant, who reported 1,013 contributors worldwide and a reconstructed circumference within 4 percent: blog.drwile.com (4 August 2025).
Two more facts sat within reach of the ancient world, and each defeats the nearby Sun by itself. The Sun holds the same width in the sky all day and from every city. A lamp hanging low overhead would look large above you and small near the horizon, and the Sun does not. The Sun also truly sets, sinking fully below the horizon. A nearby light above a flat floor can fade with distance, but it can never drop below the floor. Aristotle noted a third clue: during every lunar eclipse the Earth casts a round shadow on the Moon, from every angle. Only a ball throws a round shadow from every side.
There is a real limit to the stick test, and it is worth naming. It measures how the ground curves from north to south, and nothing else. A shape that curves the same way north to south but stays flat east to west, like a barrel resting on its side, would give the very same shadows. Ruling that out, and reaching a true globe, took more measurements from more directions.
A single run of the stick test does not prove a sphere, and careful people should not claim that it does. What it does prove, the moment you add one more distant stick, is that the Earth is not flat. The flat map argues against itself: a far Sun predicts no shadow difference, and a nearby Sun predicts a bending curve that the measurements refuse to follow. Everything that came after, from a canal in England to a camera beyond the Moon, only sharpens the same answer and fixes the shape. The Earth is a globe, spinning, a little flattened at the poles.