Previous studies in this laboratory have demonstrated 2,5-dimethylpyrrole adduct formation during in vitro exposure of protein amino groups to the neurotoxic n-hexane metabolite 2,5-hexanedione (2,5-HD). The present investigation reports in vivo pyrrole adduct formation in neural and nonneural protein from 2,5-HD-treated animals. Adult, White-Leghorn hens were given daily doses of either 200 or 70 mg 2,5-HD/kg, po, for up to 55 or 135 days, respectively. Additional animals were given 70 mg/kg for 63 days and then allowed to recover for 72 more days. Protein separation by gel electrophoresis followed by staining with a pyrrole-specific reagent yielded evidence of widespread adduct formation in protein from serum, liver, kidney, brain, and purified myelin. Binding was particularly strong in serum albumin nd myelin basic protein. Quantitation of the adduct in these tissues revealed that its formation reached peak levels at 20 days in high dose and 30 days in low-dose animals. Levels subsequently declined, suggesting the presence of a clearance mechanism capable of removing altered protein during continuing 2,5-HD exposure. Protein from animals on the recovery regimen contained no detectable pyrrole adduct. Pyrrole adduct formation was also detected in neurofilament protein preparations, although protein yields were too low to allow assessment of clearance. Hens at both dosages displayed clinical signs indicative of CNS and PNS neuropathy. Histologic findings included axonal swelling and degeneration in peripheral nerve and some spinal cord nerve tracts. A hypothesis is proposed involving differential clearance of pyrrole adduct from neural vs nonneural tissue to explain the mechanism of action and target organ specificity of 2,5-hexanedione.