During the waiting time for upper gastrointestinal endoscopy 165 patients with dyspepsia completed a questionnaire and a diary for daily measurements of the symptoms pain, heartburn, and global complaints. 23 patients (14%) had peptic ulcer disease (PUD), 18 oesophagitis (11%), and the rest were labelled nonulcer dyspepsia (NUD). NUD was further subdivided into ulcer-like, reflux-like, dysmotility, and essential NUD by means of predefined symptom profiles. 39 (24%) patients were on H2 receptor antagonist treatment. In general, the intensity of the daily symptoms was rather low, and except for a higher rating of heartburn in oesophagitis, there were no significant differences between PUD, oesophagitis, and NUD--treated or untreated. NUD patients with reflux-like dyspepsia had significantly more heartburn than the group with essential NUD; otherwise there were no differences between the subgroups of NUD. The individual daily ratings for abdominal pain, heartburn, and global symptoms varied by an average standard deviation of 64%, 97% and 47% of the mean values, respectively, and were independent of treatment or diagnoses. There was an approximately 40% probability that two successive days had different levels of symptoms. Only 10% of the patients showed stable symptoms, and the patients were completely symptom-free for 20% of the observation period. Symptoms in dyspepsia patients disclosed low intensity and high variability in this study. Such factors may be important sources of bias in clinical trials.