This study was designed to analyze possible differences in the binding of [3H]flunitrazepam ([3H]FNZP) and [3H]ethyl ‐ β ‐ carboline ‐ 3 ‐ carboxylate ([3H]β‐CCE), to rat brain membranes, in various experimental conditions. In cerebral cortex, hippocampus, cerebellum, and orain stem the number of binding sites for [3H]β‐CCE was higher than for [3H]FNZP; both were displaced by clonazepam. Until the 7th day of postnatal brain development the numbers of [3H]FNZP and [3H]β‐CCE sites were equivalent; but later on, the β‐carboline sites increased to a higher level. Noradrenergic denervation by 6‐hydroxydopamine was followed in the hippocampal formation. Already after 2 days, there was a decrease in [3H]FNZP sites, which reached 70% of control after 14 days. Similar results were obtained with DSP‐4 denervation. This change was only in Bmax and not in KD, In contrast, the [3H]β‐CCE sites did not change with denervation. Neonatal injection of l ‐ 2,4,5 ‐ trihydroxyphenylalamine or DSP‐4 produced in the adult a decrease in [3H]FNZP sites in the cerebral cortex, in parallel with the noradrenergic denervation. On the other hand, there was an increase in the cerebellum and brain stem, in correspondence with the hyperinnervation by sprouting. In these rats, the number of sites for [3H]β‐CCE did not change in the different brain regions. With 0.1% Triton X‐100, applied to synaptosomal membranes, [3H]FNZP binding was reduced by 35%, while that of [3H]β‐CCE was not significantly changed. These results suggest that there is heterogeneity of binding sites for benzodiazepine receptors in rat brain. A tentative interpretation of the experiments involving noradrenergic denervation and hyperinnervation, as well as those with Triton X‐100, is that [3H]FNZP binds to pre‐ and postsynaptic receptors, while [3H]β‐CCE binds mainly to postsynaptic benzodiazepine receptors.