Thirty seven patients with Anxiety Neurosis were selected from the patients attending the Day Centre, Whitchurch Hospital, Cardiff, between April 1972 and June 1973, following formal diagnosis of a neurotic Anxiety State by a consultant psychiatrist. A relatively homogeneous sample of subjects was selected using the following criteria to confirm diagnosis: apprehension, palpitations, panic attacks, excessive perspiration, inability to relax, and feelings of impending disaster. Patients were excluded from the parent population if they had been treated by E.C.T. in the previous six months, or had had a leucotomy, gross agitation, physical tremors, congenital blindness, or an abnormally low I.Q. Twenty-eight patients were subjected at weekly intervals to a series of individually tailored stimuli composed of stressful events pertaining to each patient's life history. Skin conductance variables (i.e. maximum deflection, baseline change, and frequency of spontaneous fluctuations) heart rate variables (i.e. changes in mean and peak heart rate), mood and personality, (E.P.I. measures of E. and N.) were examined for changes associated with recovery. It was predicted that no significant change in trend would occur in the physiological responses of non-recoverers over time, whereas there would be significant positive trends in those who recovered. The results were contrary to prediction, as significant negative trends were displayed in four out of five physiological measures in the non-recovery group, and no significant trends were recorded for recoverers, despite a tendency towards a positive trend in all physiological measures. Comparison with a control group (non-experimental) indicated that habituation effects could not account for the outcome, and also that the procedure per se had no influence on the speed of recovery. Mood scores, subjective ratings of distress, and ratings of improvement, confirm that physiological changes were related to recovery from Anxiety Neurosis. Recovery was also found to be linked with a decrease in N. scores and an increase in E. scores. It is possible to distinguish the physiological responses of recoverers and non-recoverers from Anxiety Neurosis on the basis of their reaction to highly stressful and relevant stimuli in terms of change in maximum deflection and spontaneous fluctuation frequency. The recovery curves should be generalisable to all groups of Anxiety State patients regardless of the type of treatment received. Support for the cognitive mediation of stress during Anxiety Neurosis is presented.