1 The depolarizations elicited by seven neurokinin receptor agonists were examined in both rat and guinea-pig superior cervical ganglia by use of grease-gap methodology in the presence of tetrodotoxin (0.1-mu-M). Responses were normalised with respect to 1-mu-M eledoisin. 2 The rank order of agonist potency in the rat ganglia was senktide > substance P > substance P methyl ester = eledoisin = Sar-Met-substance P > neurokinin B > neurokinin A, whereas in guinea-pig superior cervical ganglion (SCG) the rank order was senktide > Sar-Met-substance P > neurokinin B = eledoisin = substance P methyl ester. The concentration-effect curves for substance P and neurokinin A in guinea-pig ganglia were biphasic which precluded the determination of meaningful potency values. 3 The maximal depolarization achieved by subtype selective ligands was different between these two species. On rat and guinea-pig SCG, the NK3-selective ligand, senktide, produced a maximal depolarization of 27% and 274% respectively, whereas the NK1-selective ligand, substance P methyl ester, produced depolarizations of 77% and 64% respectively. 4 The depolarizations induced by substance P methyl ester and senktide in either species were unaffected by atropine (1-mu-M), suggesting a lack of involvement of presynaptic neurokinin receptors in the generation of the response. 5 The potency of substance P methyl ester, senktide, and neurokinin A were unaffected by pretreating ganglia with the peptidase inhibitors bacitracin (40-mu-g ml-1), leupeptin (4-mu-g ml-1), and chymostatin (2-mu-g ml-1). Similarly, these peptidase inhibitors had no effect on the maximal depolarizations achieved by any of these agonists. 6 It is evident that rat and guinea-pig superior cervical ganglia possess both NK1 and NK3 receptors, but that their net contribution to depolarizations are different between the two species. The depolarizations in guinea-pig SCG are mediated predominantly by an NK3 subtype and in rat SCG by an NK1 receptor subtype.