An attempt was made to identify some of the proteins of the postsynaptic density (PSD) fraction isolated from dog cerebral cortex. The major protein has been tentatively labeled "neurofilament" protein, on the basis of its 51,000 mol wt correspondence to a protein found in neurofilament preparations. Other proteins are akin to some dog myofibrillar proteins, on the basis if immunological crossreaction and equal sodium dodecyl sulfate (SDS)-gel electrophoretic mobilities. While a protein similar to dog muscle myosin is not present in the PSD fraction, a major protein present is actin, as evident from reactivity with antiactin serum, from SDS-gel mobility, and from amino acid composition. Only very little tubulin may be present in the PSD fraction, as determined by gel electrophoresis. Various treatments of the PSD fraction were attempted in order to extract some proteins, as revealed by gel electrophoresis, and to observe the structural changes of the PSD fraction residue after extraction of these proteins. The PSD is remarkably resistant to various extraction conditions, with only 4 M guanidine being found to extract most of the proteins, except the 51,000 mol wt protein. Disulfide reducing agents such as dithiothreitol (DTT), blocking agents such as p-chloromercuribenzoate (PCMB) (both in the presence of deoxycholate [DOC]), a Ca++ extractor, ethylene glycol-bis (beta- aminoethyl ether) N,N,N',N'-tetraacetate (EGTA), and guanidine caused an opening up of the native dense PSD structure, revealing approximately 10-nm filaments, presumably consisting of "neurofilament" protein. Both DTT-DOC and PCMB-DOC removed chiefly actin but also some other proteins. EGTA, in greatly opening up the structure, as observed in the electron microscope, revealed both 10-nm and 3- to 5-nm filaments; the later could be composed of actin, since actin was still in the residue after the treatment. EGTA removed a major 18,000 mol wt component and two minor proteins of 68,000 and 73,000 mol wt. Based on the morphological and biochemical evidence, a picture is presented of the PSD as a structure partly made up of 10-nm and 3- to 5-nm filaments, held together through Ca++ interaction and by bonds amendable to breakage by sulfhydrylblocking and disulfide-reducing reagents; either removal of Ca++ and/or rupture of these disulfide bonds opens up the structure. On the basis of the existence of filamentous proteins and the appearance of the PSD after certain treatments as a closed or open structure, a theory is presented with envisages the PSD to function as a modulator in the conduction of the nerve impulse, by movements of its protein relative.