Central muscarinic cholinergic involvement in classical conditioning of eyeblink responses was determined in trace and delay paradigms. Rabbits were trained on a trace procedure in which a 250-ms tone conditioned stimulus (CS) and a 100-ms air-puff unconditioned stimulus (UCS) were presented with a 500-ms trace interval. Each training session day consisted of ten tone alone, ten air-puff alone and 80 paired CS-UCS trials. Scopolamine hydrochloride at doses of 0.03 and 0.1 mg/0.5 ml per kg, s.c. dose-dependently disrupted acquisition of conditioned responses. Rabbits that were treated with scopolamine and failed to learn showed a gradual increase in conditioned responses during an additional training period with saline injections and no transfer from earlier training. Scopolamine methyl bromide, which does not appreciably cross the blood-brain barrier, showed no effects in the trace conditioning paradigm at a dose of 0.1 mg/kg, s.c., indicating central cholinergic blockade is responsible for the suppressive effect of scopolamine. Scopolamine hydrochloride at a dose of 0.1 mg/kg, s.c. did not block acquisition in the delay procedure with a 250-ms inter-stimulus interval, although the rate of acquisition was somewhat reduced by the drug. These data are the first to demonstrate that classical conditioning of the eyeblink response in the trace procedure is highly sensitive to central cholinergic deficits.