Leo Kadanoff, at the University of Illinois at Urbana-Champaign, publishes in 1966 "Scaling Laws for Ising Models near Tc" in the short-lived journal Physics, founded by Philip Anderson and Bernd Matthias. Kadanoff proposes dividing the Ising model — a simplified model of magnetism consisting of a lattice of spins that can point up or down — into microscopically large but much smaller-than-correlation-length cells or "blocks", and treating the total magnetization of each block as a new collective variable. Iterating this "block spin" transformation repeatedly, the system becomes self-similar: it looks the same at different observational scales near the critical point. This argument provides the first physical — not merely empirical — justification of the scaling hypothesis Benjamin Widom had proposed the previous year (1965): Kadanoff's calculation exactly reproduces Widom's relation among the critical exponents. Furthermore, by assuming that the block process converges to a single "fixed point" independent of the lattice's microscopic detail, Kadanoff's argument explains why very different physical systems (a magnet, a fluid, a binary alloy) share the same critical exponents near their respective phase transitions — the phenomenon known as universality. Kadanoff himself called the subdivisions of his lattice "cells", not "blocks"; the name "block spin" would become popular later. This idea becomes the direct basis on which, five years later, Kenneth Wilson builds the formal renormalization group (1971).