This study focused on two points concerning the histochemical and immunohistochemical detection of neurons that produce nitric oxide (NO): (a) the effect of fixation and other methodological parameters on the staining pattern of both NADPH-diaphorase (NADPH-d) histochemistry and nitric oxide synthase (NOS) immunohistochemistry, and (b) the possibility that neurons display immunoreactivity against NOS antisera obtained from non-neuronal sources. Frontal sections of rat brains, fixed with 4% paraformaldehyde according to different protocols, were processed for single and double labeling using NADPH-d histochemistry and neuronal (nNOS), macrophagic (macNOS), and endothelial (eNOS) NOS immunohistochemistry. Our results show that variations in the fixative schedule, even within standard parameters, produce qualitative and quantitative changes in NADPH-d labeling. The effect of fixative on weakly stained neurons is different from that on heavily stained neurons. In subfixed brains, a large number of NOS-positive neurons lose their NADPH-d activity, whereas NOS immunolabeling remains unaltered. This finding may be particularly interesting in morphological studies that compare NADPH-d activity under experimental conditions that can affect brain perfusion. On the other hand, many cortical and subcortical neurons show macNOS immunoreactivity, most of it colocalized with nNOS.