The temperature at which people chose to take a hot drink was measured in 59 patients with endoscopically proven peptic disorders of the upper gastrointestinal tract and 65 asymptomatic controls. The patients in the disease group drank significantly hotter tea or coffee than the control group (medians 62 degrees and 56 degrees Celsius respectively, P less than 0.0001). The median temperatures of choice for subgroups of patients with oesophageal, gastric or duodenal disease were significantly higher than the control group (63.5 degrees, 63 degrees, 60.5 degrees C respectively). There was no relationship between a preference for hotter drinks with either the sex or smoking habits of the patient. In the control group the temperature of choice tended to decrease with age though linear regression just failed to reach statistical significance (p = 0.06); this trend was not apparent in the disease group (p = 0.64). Thermal injury as a result of drinking hot fluids may be a causative factor in some peptic disorders.