Abstract Electrical stimulation of the ventral midbrain periaqueductal grey (PAG) elicits an opioidergic antinociception against noxious heat and pressure in freely moving rats. Recurrent stimulation was associated with a gradual decline and eventual loss of this stimulation-produced antinociception (SPA). This could be reinstated by an increase in current intensity and this reinstatement was preventable by naloxone. The current intensity-antinociception (dose-response) curve was shifted to the right in recurrently stimulated rats and parallel to that in naive animals. The loss of SPA upon repetitive simulation did not represent a conditioning phenomenon. Thus, tolerant rats exposed to all cues which accompanied stimulation revealed no (compensatory) hyperalgesic response — but rather a slight antinociception. Further, SPA recovered spontaneously in tolerant rats, Moreover, ‘extinction’ by repeated exposure to all cues accompanying stimulation did not restore or accelerate the recovery of SPA in tolerant animals. Tolerant rats showed no depletion in midbrain PAG or other CNS or hypophyseal pools of β-endorphin. Met-enkephalin or dynorphin indicating that a depletion of endogenous opioid peptides does not underlie the tolerance which develops to stimulation. In fact recurrently stimulated rats did not show any of the pronounced effects upon CNS pools of opioid peptides which are seen with long-term stress. Moreover, repetitively stimulated rats revealed no indications of stress as judged by a diversity of stress-sensitive parameters: basal nociceptive threshold, core temperature, ingestive behaviour, body weight, adrenal weight and hypophyseal secretion of β-endorphin and prolactin. The data offer two major conclusions. Firstly, the gradual loss of analgesia upon recurrent stimulation of the midbrain PAG does not reflect a generalized debilitation or stress and neither a conditioning phenomenon nor a depletion of pools of endogenous opioid peptides. Rather it closely corresponds to the pharmacological definition of tolerance and may reflect a process occurring at the level of the opioid receptor and coupled processes. This finding explains the cross-tolerance which we observe recurrently stimulated rats to display to morphine. Secondly, this SPA is not a form of stress-induced analgesia and rats undergoing recurrent stimulation reveal no indications of stress as judged by biochemical, physiological and behavioural parameters.