The pressure-temperature stability diagram of proteins and the underlying assumptions of the elliptical shape of the diagram are discussed. Possible extensions, such as aggregation and fibril formation, are considered. An important experimental observation is the extreme pressure stability of the mature fibrils. Molecular origins of the diagram in terms of models of the partial molar volume of a protein focus on cavities and hydration. Changes in thermal expansivity, compressibility and heat capacity in terms of fluctuations of the enthalpy and volume change of the unfolding should also focus on these parameters. It is argued that the study of water-soluble polymers might further our understanding of the stability diagram. Whereas the role of water in protein behaviour is unquestioned, the role of cavities is less clear.