We have investigated the sequential changes of acetylcholine receptor (AChR) distribution on identified Xenopus laevis muscle cells in culture before and after innervation. AChRs on muscle cells were stained with tetramethylrhodamine-conjugated alpha-bungarotoxin and the distribution of AChR clusters was examined on a fluorescence microscope using an image intensifier. Large receptor clusters were identified on muscle cells and their fate was followed afterward. In muscle cells cultured without neural tube cells, about one-half of the identified AChR clusters survived for 2 days. In nerve-muscle cocultures, preexisting AChR clusters survived longer on non-nerve-contacted muscle cells than on muscle cells cultured without nerve. However, in nerve-contacted muscle cells the great majority of preexisting AChR clusters dispersed within 2 days. The dispersal of preexisting AChR clusters preceded receptor accumulation along the path of nerve contact by about 12-16 hr. Therefore, an accelerated dispersal of receptor clusters in innervated muscle cells is not a consequence of receptor accumulation along the nerve. The preexisting AChR clusters located near and far from the nerve contact sites dispersed along a similar time course. Protease inhibitors, trasylol and leupeptin, reduced the nerve-induced dispersal of the preexisting AChR clusters in the period before AChR accumulation at the nerve contact sites but did not do so during the period when AChRs began to accumulate at nerve-muscle contact. The significance of the dispersal of preexisting receptor clusters is discussed with regard to neuromuscular junction formation.