Neurones within the rostral ventromedial medulla project to the intermediolateral cell column of the spinal cord where they may influence sympathetic preganglionic neuronal activity controlling cutaneous vascular beds. Here we assess whether such neurones contribute to cutaneous sympathetic vasoconstrictor activity in a fever-like state induced by i.c.v. injection of E-series prostaglandin. In urethane-anaesthetised rats, we recorded population sympathetic activity to the tail (an index of vasoconstrictor discharge regulating cutaneous thermoregulatory circulations). A survey of the effects of GABA microinjections (200 mM; 60-80 nl; 111 sites in 57 rats) demonstrated that those into the rostral ventromedial medulla (in the region of raphe pallidus and magnus; approximately bregma -10 to -12 mm) markedly decreased (51%-100%) population sympathetic cutaneous vasoconstrictor activity during "normothermic control." In contrast, injections at sites dorsal and lateral to this region tended to produce either a smaller decrease or have no effect. In heat-clamp (nine animals: body temperature 40-41 degrees C) cutaneous vasoconstrictor activity was decreased by 83+/-5%. I.c.v. prostaglandin E(1) (100 ng and above) restored activity to, or above, control levels in these animals and where body temperature was maintained at control levels (12 animals: body temperature 35.5-36.5 degrees C). The depressant action of GABA was sustained in both conditions. GABA did not significantly influence concurrently recorded splanchnic nerve activity and heart rate in any condition although both were increased following i.c.v. prostaglandin E(1) (500 ng). This study is the first to demonstrate that inhibition of neuronal activity within the rostral ventromedial medulla decreases sympathetic cutaneous vasoconstrictor activity during normothermic control and following i.c.v. prostaglandin E(1) (both with and without heat-clamp). Therefore, sympathetic premotor neurones in this area contribute to vasoconstrictor drive in these conditions. In contrast, we were unable to demonstrate that the same area had a substantial involvement in the control of splanchnic nerve activity or heart rate, even when these were enhanced following i.c.v. prostaglandin E(1).