We review the application of Brandon's criterion to identifying so-called "special'' grain boundaries. The underlying principles of the Brandon criterion, and others that have followed it, are explained, and the choices of particular parameters within the criteria are considered in the light of experimental information in the literature. It is suggested that varying choices of the parameters may be appropriate for differing applications. An experimental study of the stability of CSL-related triple junctions is used to evaluate the applicability of Brandon-like criteria to these microstructural features.