1. Stable N-methyl-D-aspartic acid (NMDA) receptor-mediated currents in cultured mouse hippocampal neurons were evoked by 20 ms pressure pulse applications of L-aspartate, repeatedly applied at 30 or 40 s intervals, to the cell body region of the neurone. We have characterized the voltage- and use-dependent blockade of the currents by three dissociative anaesthetics: ketamine, phencyclidine (PCP) and MK-801 in mouse hippocampal neurones grown in dissociated tissue culture. 2. We have used a simple model of the blockade, based on the 'guarded receptor hypothesis' to interpret our data. The model assumes that receptors are maximally activated at the peak of the response with an open probability (Po) approaching 1, that there is no desensitization and that the blocking drug only associates with, or dissociates from, receptor channels which have been activated by agonist (e.g. open channels). 3. The model allows us to estimate forward and reverse rate constants for binding of the blockers to open channels from measurements of the steady-state level of blockade and the rate of change of the current amplitude per pulse during onset and offset of blockade. As predicted by the model, the estimated reverse rate was independent of blocker concentration while the forward rate increased with concentration. Changing the level of positively charged ketamine (pKa 7.5) tenfold by changing pH from 6.5 to 8.5 caused a corresponding change in the forward rate while having no effect on the reverse rate. Most of the voltage dependence of the blockade could be accounted for by reduction of the reverse rate by depolarization. 4. Estimated forward rate constants for ketamine, PCP and MK-801 were similar to one another when measured under similar conditions and were 3 x 10(4) - 3 x 10(5) M-1 S-1. Most of the differences in potency of the three blockers could be accounted for by differences in the reverse rate constants which were approximately 0.2, 0.03 and 0.003 s-1 for ketamine, PCP and MK-801, respectively. The estimated rate constants actually are the product of the rate constants and 1/Po. Suggestions that maximum Po is much less than 1 for NMDA channels imply that both forward and reverse rate constants of blockade may in fact be larger than we have calculated. However, their magnitudes, relative to one another, are unaffected by this consideration. 5. The reverse rate constant of blockade increased at positive potentials. This increase was prevented when the neurone was loaded with N-methyl-D-glucamine, an impermeant cation which prevented outward currents.(ABSTRACT TRUNCATED AT 400 WORDS)