In rats, conditioned place preferences are produced by morphine and conditioned place aversions produced by naloxone. In the present studies, several issues concerning the demonstration and interpretation of place conditioning findings were examined in a two-compartment (black and white) tilt box: (1) the responses of naive rats to testing, (2) place conditioning in rats with strong unconditioned biases to one of the sides, and (3) modifications of the testing situation so that naive rats respond to the black and white sides with a minimum of initial bias. Experiments involving manipulation of the conditions of training and testing, use of pentobarbital, and use of a three-compartment test box helped to control for morphine's ability to produce state dependent learning as an explanation of its conditioned place preference. In addition, we examined previous place conditioning studies that failed to show aversive effects of naloxone. These negative findings were suggested to be due to the use or procedures insensitive to aversive stimuli and to the IP administration of naloxone. Finally, in the course of the experiments, novel data on general parameters of the place conditioning were provided. Dose-response curves for subcutaneous (SC) morphine (0.04-5.0 mg/kg) and naloxone (0.02-5.0 mg/kg) were established. Conditioned preferences were also shown to occur after at three pairings of SC drug, and they were retained for at least 1 month.