We reported previously that the angiotensin II (AII) – superoxide (O₂^–) signalling pathway modulates, in the rostral ventrolateral medulla (RVLM), the acute hypertensive response to emotional (air-jet) stress in rabbits (Mayorov & Head, 2003; Mayorov et al. 2004). This study determined (i) whether AII and O₂^– regulate the hypertensive response to stress in the dorsomedial hypothalamus (DMH), and (ii) whether AII and O₂^–, in the DMH or RVLM, also modulate cardiovascular arousal associated with appetitive eating behaviour. Two weeks before the experiment, under fluothane anaesthesia, rabbits were implanted with guide cannulae for bilateral microinjections into the DMH or RVLM. At the beginning of the experiment, a catheter was placed into the central ear artery under local xylocaine anaesthesia to record blood pressure. The air-jet stress was induced by directing a fine stream of compressed air to the face of the rabbit for 8 min. Animals were monitored during stress and the experiment terminated if excessive distress was observed. Feeding (which lasted 5-8 min) was initiated by presenting a small bunch of straw (~10 g) to rabbits. Upon completion of the experiment, rabbits were humanly killed with an overdose of pentobarbitone. Data (means + S.E.M.) were analysed using ANOVA and significance taken at P<0.05. Air-jet stress and feeding elicited similar tachycardic responses (+51+6 bpm and +47+4 bpm, respectively, n=17). However, the pressor response to feeding (+9+1 mmHg) was lower than that caused by stress (+16+1 mmHg, P<0.05). Inhibition of the DMH with muscimol (500 pmol; n=8) attenuated pressor and tachycardic effects of stress by 56+11% and 63+24%, respectively (P<0.01), and evoked anorexia. Local injection of a glutamate receptor antagonist kynurenate (10 nmol; n=4) decreased pressor and tachycardic responses to stress by 46+9% (P<0.01) and 82+36% (P=0.08), respectively. The cardiovascular response to feeding was similarly reduced by kynurenate, although rabbits showed no apparent changes in eating behaviour. Microinjection of a selective AT₁ receptor antagonist candesartan (500 pmol; n=9) into the DMH attenuated pressor and tachycardic responses to stress by 31+5% and 33+11%, respectively (P<0.05). Local injection of an O₂^– scavenger tempol (20 nmol; n=7) decreased these responses by 50+13% and 36+8%, respectively (P<0.05). Neither candesartan nor tempol altered cardiovascular arousal caused by feeding. Likewise, microinjection of tempol (20 nmol; n=6) into the RVLM decreased the hypertensive response to air-jet stress (-49+5%, P<0.01), but not to feeding (-3+10%, n.s.). These results indicate that AII and O₂^– modulate neuronal responsiveness in the DMH and RVLM in a stimulus-dependent fashion. In particular, local AII – O₂^– signalling may specifically subserve neural mechanisms that regulate cardiovascular arousal associated with the defence response, but not with appetitive feeding behaviour.