Temperature factors present a larger problem. With 1.6Å data one
can refine individual temperature factors without resorting to restraints
that depend on stereochemical relationships among atoms. At the other extreme,
with only 3Å data nothing is gained by refinement of individual temperatures
factors. Handling the ground between is subject to judgement. The decisions
depend on the resolution of the diffraction data and its quality. Table
ii lists my personal criteria for the application of restraints to
the temperature factors of a model.