Nitric oxide synthase (NOS) is the enzyme responsible for the production of nitric oxide, a putative neurotransmitter. The distribution of this enzyme in rat brain (Vincent & Kamura, 1992) and spinal cord (Valtschanoff et al., 1992) neurones has previously been studied using the NADPH-diaphorase technique. However, this method may not be specific for NOS and therefore a more selective tool would be valuable. Nitroarginine is an inhibitor of NOS which has recently become available in a tritiated form. We have shown that the binding of this radioligand to the soluble enzyme can be prevented by L-arginine (Michel et al., 1993) and have now examined the regional distribution of its binding in the rat cenitral nervous system (CNS) using autoradiography. Frozen sections (20 gin) of unfixed brain and spinal cord, on coated slides, were preincubated in 50 mM Tris.HCI, 3 mM CaCl2 and 0.025% Triton X100 butter, pH 7.4 for 30 min at 25°C, then incubated for 60 min at 4°C with-7 nM [3H]nitroarginine (Michel et al., 1993) with or without 1 jM nitroarginine for non-specific binding, washed for 3 x 10 min and dipped for 10 sec in ice-cold distilled water. The dried slides were apposed to X-ray film for 3 months and the images were quantified using suitable standards. The highest levels of specific binding (grey values, mean ± s. e. mean, n 2 19) were found in the tenia tecta (0.155 ± 0.009), throughout the amygdaloid complex, in particular in the anterior (0.107 ± 0.007) and medial (0.125 ± 0.006) nuclei, the amygdalopiriform transition (0.134 ± 0.005) and in the anterior cerebellum (0.087 ± 0.002). There was also binding in the thalamic bed nucleus of the stria terminalis (0.067 ± 0.005), the dentate gyrus (0.033 ± 0.001), the ventral hippocampal CA3 area (0.051 ± 0.002), layers 1 and 2 of the frontal cortex (0.064 ± 0.003) and the superior gray layer of the superior colliculus (0.053 ± 0.002). The paraventricular nucleus of the hypothalamus and the superficial layers of the dorsal hom of the thoracic and lumbar spinal cord were also labelled but this was not quantifiable at this time point. Binding in the striatum and the rest of the cortex was diffuse and therefore not quantified. There appeared to be few or no specific binding sites in the pons and medulla. The regional distribution of binding sites for [3H]itroargiine in the rat CNS was, in general, similar to that seen with the diaphorase method. There were, however, differences in the cortex and the hindbrain which may be due to the existence of more than one isoform of NOS in the CNS. Thus, [3H]nitroarginine appears to be a very useful ligand for studying the distribution of NOS in the CNS and may also be capable of distinguishing between potential isoforms. More studies are plamned using dual labelling techniques to determine whether [3H]nitroarginine binds to cells, labelled either with the diaphorase method or a specific antibody to the cloned braini isoform of NOS.