Gold sols are orange to violet, display electron dense properties and are capable of strong emission of secondary electrons. These properties enable gold particles to be used as specific markers in microscopy both at the low and high resolution level (light and fluorescent microscopy, scanning and transmission electron microscopy). Monodisperse colloidal gold can be produced by several methods in a size range of 5 nm to 150 nm. As a consequence, the gold method is well suited for multiple marking experiment at the high resolution level. Since gold markers bind non-specifically to a very low extent, the technique has found application in TEM for marking intracellular components on thin sections. Both the one step and the two step marking procedures have been utilized in the various modes of microscopy. Under appropriate conditions, gold particles can be labelled with a variety of macromolecules (polysaccharides, glycoproteins, proteins, lectins, antibodies), presumably through a noncovalent binding process. Generally the probes acquire the specific activity of the adsorbed macromolecule and their stability upon storage is good. A number of factors which influence the adsorption process are discussed in relation to the more general problems of adsorption of macromolecules onto metallic surfaces. The stability of gold markers is also best understood by the DLVO theory for disperse systems. The preparation, labelling, stabilization, stability and binding characteristics of gold markers are reviewed. Since the binding of gold probes to cell surfaces is primarily determined by the size of the particle, several problems related to steric hindrance and quantification of the method are also discussed. The advantages of the method over others are compared. The different modes of microscopy and the several gold methods available for marking cell surface and intracellular components are illustrated by micrographs.