The reinforcing effects of diazepam (DZ), triazolam (TZ) and placebo were examined in eight male subjects with histories of sedative abuse. DZ (40 or 80 mg), TZ (1.0 or 2.0 mg) and placebo were each individually available for oral self-administration using a double-blind within-subject design. After an initial sampling exposure, subjects could self-administer a single dose of drug on each of 6 days by completing a progressively increasing bicycle riding requirement. All subjects initially self-administered DZ and TZ but a decreasing number of subjects continued to self-administer drugs on the remaining days; there was no difference between DZ and TZ in the total number of self-administrations. Placebo was self-administered only by one subject on two occasions. Performance measures showed that TZ produced greater memory impairment than DZ and that DZ produced residual psychomotor impairment on the next day. With repeated dosing, evidence of tolerance was seen for both drugs on psychomotor and memory performance and subject ratings of drug liking. A few modest correlations of drug self-administration and subject-ratings were obtained, suggesting some correspondence of subject verbal and drug self-administration behaviors, but these measures did not covary in a completely consistent manner.