1. Cats were anaesthetized with urethane (1100-1200 mg/kg, i.p.), supplemented with halothane inhalation during surgery. Responses in terms of systemic arterial pressure, mean systemic arterial pressure(MSAP), heart rate, mean blood flow of the common carotid artery and femoral artery, amplitude and frequency of vocalizations, thoracic-abdominal contractions and limb movements were recorded. 2. Microinjection of N-methyl-D-aspartate (NMDA; 20 mmol/L, 200 nL) into the pretentorial periaqueductal grey (PAG) produced two classes of response: (i) cardiovascular responses and vocalization; and (ii) cardiovascular responses without vocalization. 3. For class 1 responses, five types of vocalization concomitant with pressor (VPR) or depressor (VDPR) responses were observed: (i) type 1 VDPR and VPR, elicited in the rostral and caudal part of the dorsal PAG, produced vocalization of slight hissing, with or without limb movement, moderately increased flow of the common carotid and slightly increased flow of the femoral arteries; (ii) types 2 and 3 VPR, elicited in the dorsolateral and intermediate-lateral PAG, produced hissing-howling and growling, increased respiratory movement, with or without repetitive burst limb movements or stretching of paws, slightly decreased common carotid artery flow and inconsistent changes (increased or decreased slightly) in femoral artery flow; (iii) type 4 VPR, elicited in the dorsomedial and intermediate-medial PAG, produced meowing-crying but without limb movements, common carotid artery flow was increased, but the femoral artery flow was slightly decreased or increased markedly; and (iv) type 5 VPR, elicited in the ventromedial PAG, produced meowing-screaming with or without limb movements, common carotid artery flow increased moderately and femoral artery flow increased markedly. Vocalization was loud and wild in type 4 and 5 responses. 4. For class 2 responses, two types of responses were observed: (i) a pressor response (PR) alone, elicited in the dorsolateral and intermediate-lateral PAG, produced inconsistent changes in common carotid and femoral artery flow, which increased, decreased or underwent no change; and (ii) a depressor response (DPR) elicited in the ventrolateral PAG produced moderate increases of common carotid and femoral artery flow. 5. The reduction of resistance in the femoral artery was more prominent (P < 0.05) in type 1 VDPR than in DPR. Similar changes occurred in the femoral artery among types 3 (P < 0.05) and 5 VPR (P < 0.005) and PR. In addition, the frequency of vocalization was significantly positively correlated with the increase in MSAP (P < 0.05) and mean common carotid and femoral artery flow (both P < 0.01) in all types of VPR. 6. These results suggest the presence of neurons involved in the expression of defence reactions in a longitudinal, complicated functional organization in the entire PAG column. In particular, NMDA stimulation of the medial, dorsal and dorsolateral PAG may elicit five different types of defence reactions, expressed by various forms of cardiovascular alterations concomitant with vocalization responses.