Profiles of squared normalized structure factors, <|E|(2)>(d*), have been computed for a large number of proteins and nucleic acids. These are interpreted in terms of their underlying structural features. It is also shown that the 'solvent dip' at around 6.3 A resolution is to a large extent a protein secondary-structure effect that is enhanced by the water structure. A hierarchical classification of protein structures based on their <|E|(2)>(d*) profiles is briefly outlined, together with the use of <|E|(2)>(d*) profiles as an improvement over Wilson absolute scale estimation and as a novel solvent-modelling method.