In terms of the Langford—Gray mechanistic classification, simple ligand substitution in octahedral complexes in donor solvents may be described as an I process in all cases; it is doubtful whether the criteria for D or A mechanisms could ever be met. It is argued, however, that no meaningful, unequivocal assignments along the D/ D1/Id/Ia/A lines can be made for octahedral substitution, and that attempts to make them are actually counterproductive. Instead, we should seek means of expressing the selectivity shown by various substrates towards nucleophiles, and also the lifetimes of putative intermediates, at least semi-quantitatively. Computer modelling can provide important insights into the mechanism of reaction, particularly into the key role played by the solvent in the activation process. Certain complexes of metals of the second transition series show anomalous lability; the effect requires experimental and theoretical clarification, but appears to have relativistic origins.