Cortical acetylcholine (ACh) release and content were measured in non-anesthetized pretrigeminally sectioned or in Dial-anesthetized cats. In 28 pretrigerninally sectioned cats a very highly significant negative correlation between ACh content and output was found. In the same preparations, 15 mg/kg pentobarbital or electrolytic lesion in the rostral midbrain decreased ACh output and increased content. Atropine (1 mg/kg i.v. or 1 μg/ml topically) increased ACh output fourfold without significantly altering content. A larger dose of atropine (25 mg/kg i.v.) increased output ninefold and decreased ACh content. In Dial-anesthetized preparations, picrotoxin-induced ACh release was accompanied by a decrease in ACh content. Pretreatment with atropine in the same preparation resulted in an enhanced effect of picrotoxin on both output and content while atropine alone (1 mg/kg i.v.) raised ACh output without decreasing content. In pretrigerminally sectioned non-anesthetized preparations, hemicholinium-3 (HC-3) did not reduce ACh output significantly but reduced content. In the presence of atropine, HC-3 rapidly reduced the high output and decreased ACh content faster than without atropine. Cortical ACh content determined following cholinesterase inhibition (surplus ACh) did not follow changes in ACh output. It is concluded that cortical cholinergic nerve endings do not maintain constant ACh content at various output rates in contrast to peripheral cholinergic nerve endings. Atropine enhances ACh output probably both by increasing ACh synthesis and by blocking an inhibitory loop. The amount of ACh released per minute varies from 1/500 to 1/13 of the amount present depending on the activity of cholinergic fibers. ACh collected from the surface of the brain does not originate directly from surplus ACh.