The effects of superfused epinephrine (E) and norepinephrine (NE) on the membrane potential of primary afferent fibers of the isolated frog spinal cord were studied by sucrose gap recordings from the dorsal root. In all preparations both E and NE, applied in concentrations ranging from 0.1 μM to 1.0 mM, produced a hyperpolarization of afferent terminals. In many instances this was followed by a slow depolarization and, in a small number of cords, a small depolarization preceded the increase in membrane potential. E- and NE-induced hyperpolarizations were blocked by the selective α 2-antagonists yohimbine and piperoxan, but not by the selective α 1-antagonists prazosin and corynanthine or by the ß-blockers propranolol and sotalol. The α 2-agonists clonidine, α-methylnorepinephrine and guanabenz also hyperpolarized terminals, causing a change in potential that was reduced by yohimbine and piperoxan. Taken together, these results suggest that α 2-receptors mediate the hyperpolarizing effects of E and NE. The ß-agonist isoproterenol evoked a slow depolarization similar to that produced by E and NE. The isoproterenol-depolarization was antagonized by propranolol. Sometimes, application of E and NE after superfusion with yohimbine produced only a depolarization of the dorsal root and this depolarization was sensitive to propranolol. It would appear therefore that the late depolarization seen after the application of E and NE is produced by activation of ß-receptors. In contrast, the α 1-agonist phenylephrine elicited a short latency, short duration depolarization similar to those seen preceding approximately 10% of the E- and NE-hyperpolarizations. Such short-latency depolarizations were blocked by prazosin and corynanthine. The major component of the response to both E and NE is indirectly mediated through a synaptic process: application of Mn 2+, Mg 2+, procaine or tetrodotoxin in concentrations sufficient to block synaptic transmission substantially reduced, but never eliminated, the actions of the catecholamines. Interneurons are probably involved because mephenesin, which reduces interneuronal transmission, significantly decreased the E and NE effects. Furthermore, interneurons which secrete excitatory amino acids and/or GABA may mediate the indirect effects of the catecholamines on afferent terminals because (−)baclofen and d,l-α-aminoadipate decrease, and picrotoxin and bicuculline increase, the dorsal root (DR) effects of E and NE. These results indicate that catecholamines can alter the membrane potential of afferent fibers indirectly and probably directly and' that such effects are produced by activation of both α- and ß-receptors. These results may be important in understanding the pharmacological basis of previous reports which have shown that catecholamines can affect the transmission of sensory information in the spinal cord.