X-ray diffuse scattering from protein crystals is, at the moment, the only available experimental process to be directly sensitive to long-range correlations between protein-atom displacements. It is shown here that calculations based on independent rigid-body displacements of individual molecules yield a description in good agreement with the experimental diffuse-scattering pattern displayed by tetragonal crystals of hen egg-white lysozyme (HEWL) In particular, it appears that molecular rigid-body translations and rigid-body rotations appear roughly in the same proportion as the average atomic mean-square positional fluctuations. The crystallographic temperature-factor analysis by TLS (translation/libration/screw) refinement, performed by Sternberg, Grace & Phillips [Sternberg, Grace, & Phillips (1979). J. Mel. Biol. 130, 231-253], is then confirmed and completed by a quantitative estimation of the molecular rigid-body translation contributions. The major contribution of molecular rigid-body displacements to the average atomic mean-square positional fluctuations, contradicts a previous analysis of the tetragonal HEWL diffuse-scattering data by Clarage, Clarage, Phillips, Sweet & Caspar [Clarage, Clarage, Phillips, Sweet & Caspar (1992). Proteins Struct. Funct. Genet. 12. 145-157] which concluded that short-range correlations dominate. The origin of these opposite conclusions mostly lies in the different hypotheses made to model diffuse scattering, underlying the limits of the 'homogeneous disorder' model.