To determine the factors responsible for ulcerative colitis relapse a cohort of 92 patients (18 to 78 years, 50 men) with clinically inactive disease have been followed for over 48 weeks. At 12 weekly intervals patients were asked, by means of standardised questionnaires, about infections, compliance with maintenance medication, new drug treatment, dietary changes, episodes of non-bloody diarrhoea, life stresses, and feelings of anxiety and depression. Thirty five patients (38%) relapsed (median interval 17 weeks, range three to 46 weeks). Patients who relapsed had a higher previous relapse rate than non-relapsers (p less than 0.001) and a shorter time from previous relapse to trial entry (p less than 0.05). Other clinical characteristics were equally matched in the two groups. Between and within group comparisons revealed that upper respiratory tract symptoms, antibiotic ingestion, analgesic intake, diarrhoeal episodes and stressful life events were no more common in the four weeks before relapse than before routine attendance. Anxiety and depression ratings were also similar in the two groups. The timing of ulcerative colitis relapse showed a clear seasonal pattern with 26 patients relapsing from August to January and only nine from January to July (p less than 0.001). In addition, a retrospective case note analysis revealed significant seasonality of onset of ulcerative colitis. We conclude that seasonal factors may contribute to both onset and relapse of ulcerative colitis.