This paper deals with the seismic behaviour of bracing members. An experimental investigation involving shake table tests on idealised concentrically braced single-storey frames is decribed. The testing arrangement, specimen configuration and choice of seismic input are summarised. Typical results are presented and the main experimental observations are highlighted. Elastic tests were first conducted to examine basic dynamic characteristics, followed by large amplitude seismic tests to assess the inelastic behaviour. The frames incorporated hollow steel bracing members with three different square and rectangular cross-sections sizes in order to provide a range of section and member slenderness. In addition, a number of hollow members were in-filled with mortar for comparison purposes.An assessment of a number of key design aspects is undertaken on the basis of experimental observations and findings as well as consideration of underlying response mechanisms. The main differences in the interpretation and simplification of response within codes of practice are discussed, particularly in terms of dealing with the brace buckling in compression. The implications of these differences on the overall behaviour, primarily in terms of the expected lateral overstrength as well as the level of inelastic ductility demand on the bracing members, are pointed out.The limiting criterion for the bracing members was illustrated in some of the tests, in which brace failure occurred through fracture of the cross-section following the on-set of local buckling. Depending on the section and member slenderness level, beneficial effects may be realised by in-filling the hollow braces which could delay or inhibit local buckling. The tests also indicated that bracing members with slenderness exceeding the limits imposed by some seismic codes demonstrated generally satisfactory performance. Despite their lower energy dissipation capabilities, there are several practical and design advantages that merit their utilisation.