Physicists Samuel Edwards (Cambridge) and Philip Anderson (Bell Labs) publish "Theory of spin glasses", proposing the first theory for a new class of dilute, disordered magnetic alloys called spin glasses, explaining the experimentally observed cusp in magnetic susceptibility. Their argument is that the interaction between spins, oscillating in sign with distance, produces no average ferromagnetic or antiferromagnetic order, but does produce a ground state with spins frozen in apparently random but time-fixed directions. To mathematically describe this new type of order they invent their own parameter, q (the Edwards-Anderson parameter), and a calculation method known as the "replica trick", both derived from an analogy with polymer gelation and not from a direct citation of Landau's 1937 theory. The work opened an entire field of statistical physics of disorder, laying the groundwork Giorgio Parisi would formalize in 1979.