During a recent comparison of iontophoretically applied angiotensin II (AII) and angiotensin III (AIII) in the paraventricular nucleus of the rat, we observed that the response latency for AIII was much shorter than that for AII. This suggested that AII may have to be converted to AIII before it becomes active. To test this hypothesis we performed 3 experiments. (1) We examined the effects of bestatin, an aminopeptidase B inhibitor, on the activity of applied AII and AIII. (2) Next, we monitored the effects of amastatin, a specific aminopeptidase A inhibitor, on the action of co-applied AII or AIII. (3) And, finally, we examined the response to the aminopeptidase-resistant analog Sar1-AII, both applied alone and in combination with AII or AIII. Bestatin, while having no activity of its own, dramatically enhanced the actions of both AII and AIII. Amastatin, on the other hand, had little effect on AII's action and diminished or totally blocked AII-dependent activity. Like bestatin, amastatin had no effect alone. Sar1-AII reduced spontaneous activity of angiotensin-sensitive neurons and inhibited the actions of AII and AIII in a reversible manner. The same cells were also blocked by the recognized angiotensin antagonist Sar1, Ile8-AII. In total these results strongly support the notion that AII must be converted to AIII in the brain before it is activated.