Sixteen patients with tension headaches were treated with acupuncture, needles being inserted in the hand and foot and then electrostimulated. All patients were seen at four 2-h sessions with intervals of 1 week between each. The first two sessions were used for clinical examination only, thus serving as control, while acupuncture was administered during the last two sessions. Urine samples were collected immediately before and after each session and later analyzed for catecholamine content. During the 1-week intervals between sessions, records were kept of headache attacks and medicine intake. Following acupuncture two-thirds of the patients showed reduction in number of headache attacks, and about one half of the patients reduced their use of pain relieving drugs considerably. Urinary excretion of adrenaline and noradrenaline was always higher in samples collected before sessions than after sessions. However, this reduction in catecholamine excretion was significantly greater when acupuncture had been administered.