In rats, fevers induced by moderate-to-high doses of intravenous lipopolysaccharide consist of three phases (phases 1, 2 and 3) with body temperature peaks at approximately 1, 2, and 5 h postinjection, respectively. In this study, the effects of bilateral truncal subdiaphragmatic vagotomy and intraperitoneal capsaicin desensitization on febrile phases 1-3 were assessed in adult Wistar rats. Surgical vagotomy was performed approximately 30 d before the experiment; this procedure interrupts both afferent and efferent vagal fibers. Capsaicin was administered intraperitoneally in two consecutive injections (2 and 3 mg/kg, 3 h apart) 1 week prior to the experiment; this procedure desensitizes afferent fibers, primarily within the abdominal cavity, and does not lead to the known thermal effects of systemic capsaicin desensitization. At a neutral ambient temperature, the rats were given Escherichia coli lipopolysaccharide (10 microg/kg) through a preimplanted jugular catheter, and their colonic temperature wes measured by thermocouples for 7 h. The control rats exhibited the typical triphasic febrile responses. Confirming our earlier studies, subdiaphragmatic vagotomy did not affect phases 1 and 2; it did, however, result in a 2.5-fold reduction of phase 3. Capsaicin desensitization modified the febrile response differently: phases 2 and 3 were unaffected, but phase 1 disappeared. We suggest that neural afferent fibers (nonvagal but perhaps vagal as well) play an important role in the early febrile response (phase 1) by transducing peripheral pyrogenic signals to the brain. We also suggest that vagal efferent fibers are likely to participate in the later febrile response (phase 3) via an unknown mechanism.