Traditional histochemical detection of wheat germ agglutinin-horseradish peroxidase (WGA-HRP) can impose substantial technical limitations on studies requiring co-localization of neurotransmitters, receptors and other neural antigens. The goal of our experiments was to establish the ideal conditions and reagents for immunohistochemical detection of WGA-HRP. WGA-HRP was injected into the tongues and vibrissae pads of adult rats to characterize labeling of somas and synapses, respectively. Rats were perfused with either 4% paraformaldehyde (for light microscopy, LM) or 4% paraformaldehyde/0.15% glutaraldehyde (for electron microscopy, EM) after survival times of 2, 3, 4, 5 or 6 days. For LM, brainstem tissue was cut on a cryostat at 20 microm and collected onto glass slides. For EM, tissue was sectioned with a vibratome at 50 microm and processed free floating. For LM, WGA-HRP was detected with goat anti-HRP, goat anti-WGA, biotinylated goat anti-HRP or biotinylated goat anti-WGA antibodies. For EM, WGA-HRP was detected with biotinylated goat anti-WGA and anti-HRP antibodies. Survival intervals of 3 days were ideal for staining of hypoglossal neurons, whereas an interval of 4 days produced the strongest staining of synapses within the spinal trigeminal nucleus. For LM, the biotinylated antibodies resulted in better signal-to-noise ratios than the unconjugated antibodies. At both the LM and EM levels, the biotinylated antibody to WGA produced better quality staining than the biotinylated antibody to HRP.