ABSTRACT
Early theoretical arguments of Imry and Wortis showed that disorder would result in a rounding of a first order transition, with the broadened transition remaining first order for a small disorder. Since the correlation length for a first order transition is finite, the transition proceeds through nucleation of the finite regions of the second phase. The major qualitative deviation this caused in phenomenology, from both the Ehrenfest classification and the modern classification, is that disorder-broadened first order transition cannot support the concept of an observable latent heat. The chapter discusses some of the new protocols that test this proposal of a disorder-broadened first order transition being interrupted midway and before completion. The transition is not accompanied by diffusive motions, and the arrest of transition cannot be convincingly related to diffusivity or viscosity. The latent heat of first order magnetic transitions corresponds to ordered magnetic structures, and this magnetic order could be due to localized orbital magnetism or itinerant electron magnetism.
