The ability of metals to store or trap considerable amounts of energy, and thus exist in a non-equilibrium or metastable state, is very well known in metallurgy; however, such behaviour, which is intimately connected with the defect character of metals, has been largely ignored in noble metal surface electrochemistry. Techniques for generating unusually high energy surface states for gold, and the unusual voltammetric responses of such states, are outlined. The surprisingly high (and complex) electrocatalytic activity of gold in aqueous media is attributed to the presence of a range of such non-equilibrium states as the vital entities at active sites on conventional gold surfaces. The possible relevance of these ideas to account for the remarkable catalytic activity of oxidesupported gold microparticles is briefly outlined.From a technological viewpoint heterogeneous catalysis is one of the most important areas of chemistry; in new technology such processes are responsible for over 90% of chemical and petroleum products by value (1). It is widely accepted (1) that a vital objective in this area is to identify the nature and mode of operation of surface active sites; this task was designated recently by Roberts (2) as "the ultimate in catalytic research". The problems involved appear quite daunting as, despite extensive investigation involving a vast array of sophisticated techniques, the general impression (for the vast majority of catalytic processes) is that "catalytic sites remain unidentified" (3).The active site theory of heterogeneous catalysis was proposed by Taylor (4) in 1925. Along with the suggestion that only a small percentage of surface atoms participate in such reactions, he pointed out that the sites in question "manifest an extraordinary sensitivity to heat treatment". He attributed the latter to the incomplete crystallization of the active site material, ie the latter (as discussed here later) usually exists in a metastable, non-equilibrium state. By inducing crystallization, the heat treatment eliminates the active sites and the resulting more stable material produced at the surface is relatively inactive. Taylor's suggestions are in agreement with current ideas in the surface science approach to surface catalysis, eg Somorjai (5) pointed out recently that, for the same surface, catalytic processes occur much more rapidly at defects (the metal atoms (Med) associated with defects are intrinsically active, ie μ(Med) > μ°(Me)), such as kinks and ledges, than on terraces; he also stressed that "rough [disordered or non-equilibrated] surfaces do chemistry", ie they are the most active from a catalytic viewpoint.In the present article the main emphasis is on electrocatalysis, ie the catalysis of faradaic reactions, and in particular on the behaviour of gold in aqueous media. This work may be regarded as an extension of an earlier 2-part review (6, 7) of the electrocatalytic properties of the same metal. Attention is focused on such topics as metastable, non-equilibrium or defect states of metals, me...