Neurones in the dorsolateral/lateral (DL/L) and ventrolateral (VL) columns of the midbrain periaqueductal grey (PAG), that mediate active and passive coping strategies respectively (Bandler et al., 2000), are activated by nociceptive inputs that originate in different peripheral domains, after a relay in the superficial dorsal horn. The aim of the present study was to determine whether the columnar organisation of nociceptive input to the PAG extends to differences in the pattern of C- and Aδ- nociceptor inputs that relay in the dorsal horn. In anaesthetised rats (Sagatal 60 mg kg1 i.p.) the retrograde tracer cholera toxin B (CTb; 50-100nl) was injected at sites in the VL- (n=12) or DL/L- (n=14) PAG at which prior injection of DL-homocysteic acid (50nl, 0.05M) evoked depressor or pressor responses respectively. Animals were allowed to recover and, one week later, were re-anaesthetised (Sagatal 60 mg kg^-1 i.p.) and slow (2.5^oC s-1; n=13) or fast (7.5^oC s^-1; n=13) ramps (30-55^oC) of contact heat applied (six times in each animal) to the dorsal surface of the left hind paw to preferentially activate C- or Aδ- nociceptors respectively (Yeomans & Proudfit 1996). Two hours later, the animals were killed humanely. Control animals (n=4) received no peripheral stimulation. 40µm sections of the PAG were processed immunocytochemically to visualise injection sites and 40µm sections of the spinal cord were processed to visualise neurones labelled retrogradely and those in which Fos protein was evoked by the heat stimulus. Unpaired Mann-Whitney U test was used to assess levels of significance. There were no significant differences in the total numbers of spinal neurones activated by the different heat stimuli. However, in lamina II, significantly (P< 0.05) more neurones were activated by slow compared to fast ramp stimulation. No such differences were found in other laminae or in the lateral spinal nucleus (LSN). Neurones that projected to the contralateral PAG were located in Lamina I and LSN; 11.95% and 35.8% of heat-activated neurones respectively. Significantly (P<0.05) more lamina I neurones activated by slow ramp stimulation projected to the VL- than to the DL/L- PAG. No differences were found in the columnar organisation of projections of neurons activated by fast ramp stimulation.These data support the view that activation of Cnociceptors may preferentially activate passive coping strategies that are co-ordinated by neurons in the VL-PAG. In contrast, Aδ-mediated nociception is equally represented in DL/L and VL sectors.