The principle finding of the present study is that there are two types of mechanical hyperalgesia developing in human hairy skin following injurious stimuli. Mechanical hyperalgesia comprises a dynamic component (brush-evoked pain, allodynia) signalled by large myelinated afferents and a static component (hyperalgesia to pressure stimuli) signalled by unmyelinated afferents. While the static component is only found in the injured area, the dynamic component also extends into a halo of undamaged tissue surrounding the injury. The irritant chemicals, mustard oil or capsaicin, were applied transdermally in 20 subjects to a patch (2 × 2 cm) of hairy skin. Both substances evoked burning pain and hyperalgesia to mechanical stimuli. While stroking normal skin with a cotton bud was perceived only as touch prior to chemical stimulation, there was a distinctly unpleasant sensation afterwards. This component of mechanical hyperalgesia persisted for at least 30 min and was present in the skin exposed to the irritants (primary hyperalgesia) as well as in a zone of untreated skin surrounding the injury (secondary hyperalgesia) measuring 38 ± 4 cm2 after capsaicin. Pressure pain thresholds dropped to 55 ± 8% of baseline level after mustard oil and to 46 ± 9% after capsaicin. However, this drop of thresholds was short-lived, lasting 5 min following mustard oil but persisting more than 30 min following capsaicin treatment. The reduction of pressure pain thresholds was only observed for treated skin areas, but not in the surrounding undamaged tissue from where brush-evoked pain could be evoked. When pressure pain thresholds were lowered, the pain had a burning quality which differed distinctly from the quality of brush-evoked pain. On-going burning pain and both types of mechanical hyperalgesia were critically temperature dependent. Mildly cooling the skin provided instant relief from on-going pain, abolished brush-evoked pain and normalized pressure pain thresholds. Rewarming resulted in a reappearence of on-going pain and hyperalgesia. The effect of a nerve compression block of the superficial radial nerve on these sensations was tested in 14 experiments. When the ability to perceive light touch had been abolished, there was also no touch-evoked pain, indicating that this component of mechanical hyperalgesia is mediated by large-diameter primary afferents. At a later stage of the block when the subjects' ability to perceive cold stimuli had also been lost, application of cool stimuli still eliminated on-going burning pain, suggesting that pain relief afforded by cooling the skin acts at the peripheral receptor level and not by central masking. Importantly, at this stage of the block, when only unmyelinated primary afferents conducted, neither spontaneous pain, nor hyperalgesia to heat, nor the lowered pressure pain threshold had changed significantly. Based on the differences in quality of sensations, in spatial and temporal profiles and in susceptibility to differential nerve blocks, we conclude that irritant chemicals induce a dynamic and static component of mechanical hyperalgesia signalled by large-diameter or unmyelinated fibres, respectively. While the static component may be mediated by sensitized peripheral nociceptors, the dynamic component is probably the consequence of an altered processing of large diameter primary afferent input in the central nervous system subsequent to a maintained barrage of nociceptor activity. The parallel fluctuation of brush-evoked and burning background pain therefore suggest that on-going activity from nociceptors is required to maintain a central state that permits dynamic mechanical hyperalgesia to be expressed in humans.