1. Recordings were made from thirty-nine convergent neurones in the lumbar enlargement of the rat spinal cord. These neurones were activated by both innocuous and noxious stimuli applied to their excitatory receptive fields located on the extremity of the ipsilateral hind paw. Transcutaneous application of suprathreshold 2 ms square-wave pulses to the centre of the receptive field resulted in responses to A- and C-fibre activation being observed; a mean of 18.8 +/- 1.8 C-fibre latency spikes was evoked per stimulus. This type of response was inhibited by applying noxious conditioning stimuli to heterotopic body areas; immersing the tail in a 52 degrees C water-bath caused a mean 54.5 +/- 2.3% inhibition of the C-fibre-evoked response; such inhibitory processes have been termed diffuse noxious inhibitory controls (d.n.i.c.). 2. The effects of microinjections of morphine (5 micrograms; 0.2 microliter) on both the unconditioned C-fibre-evoked response and inhibitory processes triggered from the tail were investigated in an attempt to answer two questions: (a) does morphine increase tonic descending inhibitory processes and (b) what are the effects of morphine on descending inhibitory processes triggered by noxious stimuli? 3. The predominant effect of periaqueductal grey matter (p.a.g.) morphine on the C-fibre-evoked responses was a facilitation: 51% of cells had their C-fibre-evoked responses increased by morphine (by roughly 50%); 31% of cells were not influenced while the remaining 18% of units were depressed; however the cells classified as depressed were only marginally so. No clear relationships were found either between the microinjection sites in the p.a.g. and their corresponding effects or between the number of C-fibre-spikes evoked in the control sequences and the subsequent effect of morphine. 4. While d.n.i.c. was not altered by morphine in 56% of cases, it was clearly reduced in the remaining cells. The effects were immediate but peaked at 40 min following the microinjection (a mean 77% reduction) and then returned towards control values. All but three of the corresponding microinjection sites were such as to include the medio-ventral p.a.g. including the nucleus raphé dorsalis. In contrast none of the cases where d.n.i.c. was unaltered included microinjection sites in this region. 5. No relationship was found between the changes in d.n.i.c. and the number of spikes evoked in the control sequences, or the changes in the C-fibre responses. 6. Autoradiographic controls using [3H]morphine showed a large diffusion of the drug within an area of about 0.75 mm around the tip of the cannula.(ABSTRACT TRUNCATED AT 400 WORDS)