Missed Symmetry and How to Avoid Being Marshed. A.L. Spek, Bijvoet Centre for Biomolecular Research, Utrecht University, Padualaan 8, 3584 CH, Utrecht, The Netherlands.
In the late seventies it became clear that a significant number of the exponentionally growing number of published structural studies are described with too low crystallographic symmetry. This fact was brought to general attention with papers by R.E. Marsh and others with titles like 'Some Incorrect Space Groups in Inorganic Chemistry, Volume 16'.
The reason for publishing a structure in the wrong space group can be traced to the fact that provisional decisions on symmetry have to be made at various stages of a structure determination, i.e. data collection, data reduction, phase determination, refinement and publication. It is often a matter of experience to recognise problems in the later stages and to trace them back to incorrect decisions in an earlier stage of the study and to correct for them.
The problem cases can be divided in two classes. The first one in general contains poorly defined structures where a centrosymmetric description should have been used in place of a non-centrosymmetric one. Signals are usually poor refinement, strange geometry and terrible ORTEP (often discretely hidden with a PLUTO drawing). The second class involves additional symmetry elements other than an inversion centre. The only signal for higher symmetry can be a ratio higher than 1 for the number of crystallographically and chemically independent species.
A number of tools have been developed to investigate the problem. A very powerful one is the MISSYM algorithm (Y. Le Page). Other tools address both the problem of higher crystallographic symmetry and non-crystallographic symmetry (e.g. BUNYIP, S.Hall et al). A number of the available tools will be discussed, including those implemented in the program PLATON (MISSYM, NEWSYM, LEPAGE and NONSYM) (copies by anonymous ftp: xraysoft.chem.uu.nl).
Papers submitted to Acta Cryst are now routinely checked for missed symmetry. Similar measures are to be expected soon from other journals and databases accepting data in CIF-format.