Prime-Field Census

On a prime's multiplication clock, pick a number (let's say 2) and keep multiplying by it: 2, 4, 8, … Each step hops around the clock. Sometimes the hops eventually visit every number before looping back; sometimes only some of them. (On that clock, the orange dots are exactly the full-tour numbers, for one prime. This page asks how common they are across all primes.)

The surprise is what happens when you check this across thousands of different prime clocks: a number like 2 takes the full tour about 37% of the time, and almost any number you pick lands near that same 37%. Meanwhile, how often a number is a perfect square on the clock settles at exactly half. So each curve below is one number, drawn as more and more primes are counted; they all drift toward those steady values.

Watch g = 5 run stubbornly high. Some numbers just turn out special. (Curiosity: that steady ~37% is a real unsolved riddle, Artin's constant ≈ 0.3739; the exact 50% is provable, from a classic law called quadratic reciprocity.)