We describe a method for estimating the average proportion of GABAergic terminal area relative to total nerve terminal area with confocal microscopy. Nerve terminal regions were identified with dual colour immunofluorescence on Vibratome sections with an antibody to synaptophysin (SYN), and GABAergic processes, including axon terminals, were identified with an antibody to glutamic acid decarboxylase (GAD). Sections were viewed in an Olympus AX70 microscope attached to a Biorad 1024 MRC scanning confocal system. Images were collected with a 100 x objective from the same tissue locations and imported into the NIH-Image program, where black and white binary images were obtained for co-localisation and quantitation. Measurements were made separately of areas of SYN/GAD (GABAergic terminals) and SYN labelling (all terminals). The relative proportion of GABAergic terminal areas in visual cortex (6.1+/-1%; mean +/- SE), CA1 hippocampus (2.6+/-0.5%) and deep cerebellar nuclei (46.6+/-3%) are consistent with what is known of the relative levels of inhibitory input to these structures. The assumptions that SYN labelling is restricted to axon terminals, and that SYN labels all axon terminals was tested by ultrastructural localisation of SYN in the three brain regions examined. Only 7.4+/-0.4% of SYN-labelled profiles could not be positively identified as synaptic vesicle-containing axon terminals, and between 93.4 and 99.2% of vesiculated axon profiles were SYN-positive. These results suggest that SYN is a very reliable marker for axon terminals, and validates the confocal analytical approach. The confocal method allows rapid sampling of many brain regions and would be suitable for examining terminals containing any neurotransmitter that can be detected immunocytochemically.