Two sticks measure a planet.

Eratosthenes knew that at noon on the summer solstice, the sun sat directly overhead at Syene (near modern Aswan) — a vertical stick cast no shadow at all. On the same day, a stick in Alexandria, about 800 km north, cast a shadow 7.2° off vertical. Since the sun's rays arrive essentially parallel (it's 150 million km away), that 7.2° can only come from the ground itself curving between the two cities. A 7.2° wedge is 1/50th of a full circle, so the whole circumference must be 50×800 km = 40,000 km — within a few tenths of a percent of the true value, using nothing but sticks and shadows. The diagram is schematic, not drawn to true curvature or scale; the small globe on the left shows the wedge these two sticks actually correspond to. This also assumes both sticks sit on the same meridian (line of longitude) and measure at the same solar noon, which Eratosthenes' setup approximated but didn't hit exactly — part of why his answer, while remarkably close, wasn't exact. Read more on Wikipedia.