Sympathetic vasoconstriction of cerebral vessels has been proposed to be a protective mechanism for the brain, limiting cerebral perfusion and microcirculatory pressure during transient increases in arterial pressure. To furnish direct neural evidence for this proposition, we aimed to develop a method for recording cerebral sympathetic nerve activity (SNA) from the superior cervical ganglion (SCG). We hypothesized that SNA recorded from the SCG increases during imposed hypertension, but not during hypotension. Lambs (n = 11) were anesthetized (-chloralose, 20 mg/kg/hr) and ventilated. SNA was measured using 25µm tungsten microelectrodes inserted into the SCG. Arterial blood pressure (AP) was pharmacologically raised (adrenaline, phenylephrine or angiotensin II, 1-50 µg/kg i.v.), mechanically raised (intravascular balloon in the thoracic aorta), or lowered (sodium nitroprusside, 1-50 µg/kg i.v.). In response to adrenaline (n = 10), mean AP increased 135 ± 10 % from baseline (mean ± S.E.) and the RMS value of SNA (Square Root of the Mean of the Squares, SNA_RMS) increased 255 ± 120 %. In response to mechanically induced hypertension, mean AP increased 43 ± 3 % and SNA_RMS increased 53 ± 13 %. Generally, (9 of 10 animals), SNA_RMS did not increase as AP was lowered with sodium nitroprusside. Using a new model for direct recording of cerebral SNA from the SCG, we have demonstrated that SNA increases in response to large induced rises, but not falls, in AP. These findings furnish direct support for the proposed protective role for sympathetic nerves in the cerebral circulation.