Mice were rendered tolerant to morphine by the subcutaneous implantation of one 75 mg morphine pellet. Seventy-two hours post-pellet implantation, the animals were evaluated in the tail-flick assay for analgesic tolerance and cross-tolerance to subcutaneously administered morphine, normorphine, methadone, etorphine and intracerebroventricularly administered morphine. With the pellet remaining in situ during testing, there was the expected analgesic tolerance to peripherally administered morphine and analgesic cross-tolerance to normorphine. However, with the pellet in situ during testing, there was a surprising lack of analgesic tolerance to intracerebroventricular administered morphine and no analgesic cross tolerance to peripherally administered etorphine or methadone. In contrast, removal of the morphine pellet 3 hours prior to the analgesic evaluation apparently unmasked the expression of tolerance and cross-tolerance as evidenced by a three fold, parallel shift to the right of the analgesic dose-response curve for subcutaneously administered etorphine and methadone and a seven fold shift to intracerebroventricularly administered morphine. These data emphasize that a more rigorous evaluation of tolerance development methodologies need be explored and support the suggestion that removal of the morphine-inducing pellet prior to analgesic determinations results in a distinct state of "tolerance" quite different from that observed with the pellet remaining in situ during testing.