Primes lining up on diagonals nobody can fully explain.
Stanisław Ulam, bored in a 1963 lecture, doodled the integers in a square spiral and circled the primes. The primes did not scatter. They crowded along visible diagonals. Why primes prefer certain quadratic forms over others is one of the deepest unsolved problems in number theory — Ulam saw it on a napkin.
Write 1 at the centre, then walk the integers around it in a square spiral — right, up, left, left, down, down, right, right, right. Now fill only the cells whose number is prime. A pattern should not be there. One is.
Stanisław Ulam doodled this during a tedious seminar on the back of a napkin. The next year he, Stein and Wells published it in the American Mathematical Monthly, and Martin Gardner's Scientific American column made the 65 000-cell picture famous the same year. The diagonals had been hiding in the integers the whole time.
Each diagonal is a quadratic 4n² + bn + c. That so many of them are prime-rich is a window onto Bunyakovsky's 1857 conjecture and the Hardy–Littlewood prime-tuple program — open problems that still surprise us today.
Write 1 in the middle of squared paper. Step right for 2, up for 3, left for 4 and 5, down for 6 and 7, right for 8, 9, 10, then up for 11, 12, 13. Keep walking outward in a square spiral — every integer lands in exactly one cell.
Numbers along any straight diagonal of the spiral obey a quadratic of the form 4n² + bn + c. The visible streaks of primes are exactly the diagonals whose polynomial happens to be unusually prime-rich. The geometry is just bookkeeping; the arithmetic is the surprise.
Amber = prime. Pink = the diagonal carved by the chosen quadratic. Cyan outline marks 1 at the centre.
In 1772 Euler noticed that n² − n + 41 is prime for every integer n from 0 to 40: forty-one consecutive prime values from a single quadratic, first failing at n = 41 (which lands on 41²). On the spiral this is a single uninterrupted diagonal streak. No other simple quadratic has yet beaten it.
Prime for every n from 0 to 40: a single uninterrupted diagonal streak.
Each cell shows n on top and a·n² + b·n + c below. Amber cells are prime. Euler's quadratic scores a perfect 41 / 41, with 41 itself appearing at both n = 0 and n = 1. Try your own.
Viktor Bunyakovsky conjectured in 1857 that any irreducible polynomial in ℤ[x] with no fixed prime divisor and positive leading coefficient produces infinitely many primes. For n² + 1, for n² − n + 41, for almost every specific quadratic the spiral makes visible — we still cannot prove it.
Hardy and Littlewood's 1923 conjecture F predicts, for each quadratic, the exact asymptotic density of primes along its diagonal. The conjectures are unproven, yet the numerical match with Ulam's spiral is near-perfect — the picture confirms the arithmetic, and the arithmetic explains the picture.
Behind the diagonals lives the analytic number theory of quadratic forms, the Sato–Tate distribution of trace data, and modular forms. Ulam's napkin doodle is a window onto one of the most stubborn corners of modern mathematics — accidentally drawn by anyone with squared paper.
The Explorer pushes the grid up to 501 × 501 — a quarter-million integers — and lets you sweep through quadratics to find your own prime-rich diagonals.
→ Open the ExplorerHow and where this technique lives in the world today.
Topics in the same vein.
