The preceding discussion documents the diverse ways in which monoclonal antibodies have contributed to neuroscience research. They provide highly specific reagents to membrane-associated proteins, such as pumps, channels, receptors, and cell-adhesion molecules, that are useful for purifying these proteins, studying their structures at high resolution, and mapping their distributions. In many cases, the specific reagents were obtained using only partially purified antigens. Monoclonal antibodies to cytoskeletal proteins, organelles, and protein kinases have revealed that specific molecules are concentrated in anatomically distinct regions of the cell. A protein kinase has been shown to be a major postsynaptic constituent in many synapses. Individual proteins, such as actin, tubulin, and calmodulin appear to have different antigenic epitopes shielded in different parts of the cell. Monoclonal antibodies have provided a diversity of cell-type-specific reagents in both vertebrate and invertebrate nervous systems. They seem likely to be useful in identifying functionally related subpopulations of neurons and describing neural cell lineages. They will also serve to identify molecules that are important in regulating cell migration in the cerebellum, in marking cell position in the retina, and directing axon growth. This review also documents many purposes for which monoclonal antibodies are poorly suited or must be used with caution: A monoclonal antibody to a protein does not always reveal every place where that molecule is located. Pre- or post-translational microheterogeneity can expose different epitopes on the protein, such as may occur on the Na+-channel. Other proteins within the cell may shield antigenic sites on proteins such as calmodulin. Monoclonal antibodies can bind to epitopes on unrelated molecules (Nigg et al 1982, Lane & Koprowski 1982). This is revealed in some cases as multiple bands on immunoblots. Some cross-reactivity, however, may have a functional basis. For example, structural homology is clearly the basis for the antigenic epitopes that are shared among the five classes of intermediate filaments (Pruss et al 1981). The epitope that appears to be shared between the muscarinic and alpha 1-adrenergic receptors may be conserved because the two receptors modulate common effectors. The cross-reactivity between these receptors was only recognized because very specific and sensitive assays exist for each. It is quite possible that these same antibodies also bind sites on many other types of receptors. Mapping the distribution of this epitope may therefore have little relationship to the actual distribution of the muscarinic receptor.(ABSTRACT TRUNCATED AT 400 WORDS)