The literature reveals that the mechanisms of some solid-state dehydrations are more complicated than has been generally accepted. Reactions at a thin advancing reactant-product interface provide the geometric models on which the most widely employed rate equations are based. For some systems, this "thin interface" model is a simplification of observed behaviour. Elimination of water from crystallographic sites may occur to a significant extent within a much thicker zone of reactant towards which the active interface is progressing. Consequently the region of chemical change may not coincide with the region of structural transformation. Limited initial dehydration may occur across all crystal faces prior to the onset of a nucleation and growth process that is usually regarded as the dominant rate process in the dehydrations of many large crystals. Experimental observations for solid-state dehydrations are discussed and reaction mechanisms with different rate controlling processes are distinguished. Studies of dehydrations have contributed substantially to the theory of solid-state reactivity, and advances in understanding may have wider application to other solid-state reactants.