The clustering of acetylcholine receptors (AChR) in the postsynaptic membrane of newly innervated muscle fibers is one of the earliest events in the development of the vertebrate neuromuscular junction. Here, we describe two hypotheses that can account for AChR clustering in response to innervation. The "trophic factor" hypothesis proposes that the neuron releases a soluble factor that interacts with the muscle cell in a specific manner and that this interaction results in the local accumulation of AChR. The "contact and adhesion" hypothesis proposes that the binding of the nerve to the muscle cell surface is itself sufficient to induce AChR clustering, without the participation of soluble factors. We present a model for the molecular assembly of AChR clusters based on the contact and adhesion hypothesis. The model involves the sequential assembly of three distinct membrane domains. The first domain to form serves to attach microfilaments to the cytoplasmic surface of the muscle cell membrane at sites of muscle-nerve adhesion. The second domain to form is clathrin-coated membrane; it serves as a site of insertion of additional membrane elements, including AChR. Upon insertion of AChR into the cell surface, a membrane skeleton assembles by anchoring itself to the AChR. The skeleton, composed in part of actin and spectrin, binds and immobilizes significant numbers of AChR, thereby forming the third membrane domain of the AChR cluster. We make several predictions that should distinguish this model of AChR clustering from one that invokes soluble, trophic factors.