We have examined the applicability of the Hamilton Anxiety Scale (HAS) and our Melancholia Scale (MES) on a group of patients suffering from cardiac disease. They were assessed before surgery and again 1, 3, and 6 months after the operation. The results showed that the median for this group of patients even preoperatively was below the cut-off scores for both rating scales with a monotonous fall in score on the follow-ups. Item analysis showed that it was the psychic or cognitive symptoms of the HAS rather than the somatic symptoms of anxiety that explained the score variation. On the MES it was symptoms like depressed mood, psychic anxiety, pains, emotional and intellectual retardation rather than symptoms of guilt, motor retardation or suicidal impulses that were present. From both scales a total of 10 items emerged which were considered to measure generalized anxiety (Generalized Anxiety Scale). When the patients were classified into groups by a global assessment according to their ability to verbalize preoperative anxiety it was found that patients who were less able to verbalize had the lowest rating scale score. It was argued that preoperative anxiety is an introspective perception of fluctuating somatic manifestations of anxiety, whereas the rating scale procedure focuses on the persistent or cognitive part of anxiety which mostly is retrospectively perceived.