The greatest loss of memory shown by mice 24 hr after learning was found to occur with cycloheximide (CXM) (120 mg/kg) administered subcutaneously 30 min before training. With injection at this time the extent of the amnesia was dose dependent (30-150 mg/kg) and the resultant amnesia was found to be relatively constant when tested at 1, 7 or 14 days. An attempt was made to follow the development of this amnesia with 100 and 120 mg/kg CXM. However, the saline controls showed an unexpectedly low avoidance 6 hr after training. This was interpreted as a possible interaction between the stress of the injection and the 6 hr interval. An experiment designed to test this possibility showed that mice injected with 0.1 ml of 1% lignocaine gave high avoidance at 6 hr but mice receiving only a needle puncture of the skin gave performances similar to mice receiving saline injections. It was felt that these findings cast doubt on the usefulness of the passive avoidance task in the assessment of drug action on short term memory.