Acute and subchronic administration of N-methyl-D-aspartate antagonists to rats in the early postnatal period has been reported to produce widespread and selectively cortical neurotoxicity, respectively. To resolve this apparent discrepancy, we sought to clarify these data by determining the dose and temporal and regional characteristics of acute and subchronic phencyclidine (PCP)-induced neurotoxicity. Measurement of degenerating neurons with the cupric silver technique following a single dose of PCP on postnatal day (PN) 7 revealed that neurodegeneration increased in all areas measured (frontal, parietal and cingulate cortices, striatum, hippocampus, subiculum, and thalamus) within 9 hr. Silver staining peaked at 9-16 hr and was then not detectable or was greatly reduced after 24 hr depending on the specific region. Dose-response analysis at 9 hr showed that the lowest effective dose was 1, 3, and 10 mg/kg for the frontal cortex, hippocampus, and striatum, respectively. However, repeated PCP administration (10 mg/kg) on PN 7, 9, and 11 elicited an increase in silver staining only in the frontal cortex. To determine whether the loss of effect in the striatum and hippocampus was due to a "tolerance" mechanism or to a developmental phenomenon, we compared the effects of PCP given on PN 7, 9, or 11 with those of two doses given on PN 7 and 9 or three doses administered on PN 7, 9, and 11. Analysis of these experiments shows that both developmental factors and unknown mechanisms of tolerance underlie the apparent selective cortical neurotoxicity observed following subchronic PCP administration in perinatal rat pups.