Bertrand I. Halperin and David R. Nelson, at Harvard University, extend Kosterlitz and Thouless's concept of topological order to the specific problem of how a two-dimensional crystal melts. They show that, unlike conventional melting in three dimensions, a two-dimensional solid can melt in two distinct stages mediated by topological defects: first, thermal dissociation of dislocation pairs destroys long-range positional order but preserves orientational order, producing an intermediate phase — which the authors themselves call "hexatic" — with properties midway between a solid and a liquid; then, at higher temperature, dissociation of disclinations also destroys orientational order, giving rise to the conventional isotropic liquid. This hexatic phase has no strict analog in three dimensions and was experimentally confirmed two years later in liquid crystal films. The work, together with a parallel, independent development by A.P. Young, completes what is today known as KTHNY theory.