Behavioral effects of the beta-carboline derivatives ZK 93423 (6-benzyloxy-4-methoxymethyl-beta-carboline-3-carboxylic acid ethyl ester) and ZK 91296 (5-benzyloxy-4-methoxymethyl-beta-carboline-3-carboxylic acid ethyl ester) were compared with those of lorazepam and suriclone in squirrel monkeys. Two groups of monkeys were trained to respond under a fixed-interval schedule of food presentation. In one group, responding was suppressed (punished) by superimposing a fixed-ratio schedule of response-produced electric shock. Dose-effect curves were determined for all drugs by administering single doses i.m. 15 min before the start of the fixed-interval schedule. Low and intermediate doses of ZK 93423 (0.03-0.3 mg/kg), ZK 91296 (0.3-10.0 mg/kg), lorazepam (0.03-0.3 mg/kg) and suriclone (0.003-0.03 mg/kg) produced dose-related increases in the rates of both suppressed and nonsuppressed responding. The increases in response rates normally produced by maximally effective doses of ZK 93423, ZK 91296, lorazepam or suriclone were eliminated when the drugs were given in combination with the benzodiazepine antagonist Ro 15-1788. Although many beta-carboline derivatives have been found to act as "inverse agonists" at benzodiazepine recognition sites, the present results show that ZK 93423 and ZK 91296 have clear benzodiazepine-like (agonist) effects on schedule-controlled behavior in squirrel monkeys.