The activity-dependent mechanism that refines the topography of the retinotectal projection in frogs is mediated by the NMDA receptor. Earlier studies found that chronic treatment of the optic tectum with the NMDA receptor antagonist DL-2-amino-5-phosphonovaleric acid (DL-AP5) desegregated eye-specific stripes in three-eyed frogs, while chronic treatment with NMDA sharpened stripe borders (Cline et al., 1987; Cline and Constantine-Paton, 1990). We now report that this same chronic treatment with NMDA decreases the electro-physiologically measured sensitivity of the optic tectum to applied NMDA: acute application of a given concentration of NMDA decreased the evoked tectal potential to a lesser extent in animals chronically treated with NMDA than it did in normal and sham-treated animals. This is observed as a shift to the right in the NMDA dose-response curves for both the positive and negative postsynaptic components of the evoked tectal response. We believe that this decreased NMDA receptor effectiveness further restricts the intermingling of axon branches from the two eyes by limiting synapse stabilization to areas where afferent activity is most correlated. This would account for the anatomical sharpening of stripe borders (i.e., increased afferent segregation). Quantitative autoradiographic analysis of 3H-glutamate binding to NMDA receptors indicated that binding densities within the tectum did not differ between control groups and NMDA chronically treated groups. We suggest that in the experimental animals the response to NMDA may be decreased by a change in the effectiveness of individual NMDA receptors rather than by decreases in receptor number. This experimentally induced change may be analogous to naturally occurring decreases in receptor function that correlate with the end of some periods of visual plasticity in mammals.