This review, covering the period 1987-1993, deals with quantum chemical studies that employ the supermolecular method with some form of counterpoise to calculate the interaction energies of molecular complexes. A theorem is provided showing that the standard counterpoise approach, in which, at a given geometry, energies are evaluated using the full basis set of the complex, yields a pure (i.e. BSSE-free) interaction energy. Artifacts resulting from secondary BSSE have been found to become negligible before the basis set limit of this interaction energy is attained. Alternative methods in which counterpoise is achieved by restricting the fragment description in the calculation of the:full complex to the level used for the isolated fragment are formally nearly correct but have been found to give less satisfactory results, mainly because the relevant fragment properties converge only slowly to the basis set limit. Some practical aspects are considered as well, such as the ambiguities that arise in calculating the binding energy of a cluster of molecules.