Morphine (15 mg/kg i.p.) produces a biphasic effect: hypokinesia, followed by hyperkinesia and stereotyped behaviour. After repeated administration, the signs of the second phase more and more predominate. In the following study, it was evaluated, to which degree "classical" pharmacodynamic tolerance to hypokinesia or, alternatively, conditioning phenomena contribute to this shift. In particular, it was studied whether not only hyperkinesia but also stereotypies could occur as conditioned responses either in the presence or in absence of morphine. Rats were conditioned 8 times with morphine (15 mg/kg i.p.) in the presence of various, defined conditioned stimuli (auditory, olfactory and tactile), another group was "pseudoconditioned", i.e. they were exposed to the same treatment schedule of morphine and stimuli, but with no positive association between drug and stimuli, a third group ("naive rats") was treated with saline instead of morphine, but exposed to the same stimuli as both other groups. All groups were tested for conditioned responses in the presence of the conditioned stimuli. One series of experiments was performed with saline after a break of 2 days after the end of the conditioning period, a second series was tested with saline after a break of 7 days, a third series with morphine (15 mg/kg i.p.) after a break of 2 days, a fourth series with the same dose of morphine after a break of 7 days. The results showed that when morphine was used after a break of 2 days, "classical" pharmacodynamic tolerance, but not conditioning phenomena could explain the shift in behaviour, whereas under the three other protocols described, some conditioned behavioural effects could be observed, either in presence or in absence of morphine, at least in part of the parameters used (locomotor activation, decrease in hypokinesia, sniffing, gnawing, rearing, but not in licking). Accordingly, development of pharmacodynamic tolerance and, to a lesser degree, conditioning contribute to the shift in behaviour after repeated administration of morphine. The conditioned effects could not be attributed to any alteration in striatal or mesolimbic dopamine turnover.