One hundred and three patients referred to a neurological outpatient clinic were examined to assess the relationship between persistent headache, not due to significant physical illness, and emotional disturbance. Overall, the patients showed slightly more evidence of emotional disturbance than a general practice population but less than psychiatric outpatients. Thus, with cut-off points of 4/5 and 9/10 on the General Health Questionnaire (GHQ 28) the whole group had 52% or 20% of psychiatric 'cases' respectively. On the Crown-Crisp Experiential Index the 70 females had mean total scores of 37.19 +/- 11.11 and the 33 males had scores of 31.79 +/- 11.36. In addition the childhood experiences measured by the Parental Bonding Instrument appeared to be normal. Seven patients had significant depressive illness, according to the Levine-Pilowsky Depression Questionnaire. Statistically significant differences in psychological state did not emerge between the diagnoses of cluster headache, classical migraine, common migraine, tension headache or combined headache. However, negative correlations were found between the duration of illness and measures of anxiety. It is concluded that although the emotional state contributes to the development of pain and headache in some patients, there are others in whom comparable headaches are unlikely to be due to emotional factors. Selection effects are held to be important and some of the emotional changes will vary at different phases of a chronic disorder. A new symptom may initially cause anxiety but when a condition persists some patients will be increasingly concerned or depressed whilst others develop tolerance for the situation.