Ambulatory patients frequently ingest liquids or smoke just before temperature measurement. The change in body temperature measurements over time following ingestion of ice water, hot water, and smoking were investigated. Twenty-two healthy, afebrile study subjects sequentially ingested temperature-controlled water and smoked a cigarette. Simultaneous oral and auditory canal temperatures were measured over 15 minutes following ingestion. Auditory canal temperatures were obtained with an infrared detection probe; we designated this process a tympanic membrane-derived (TMD) temperature. To determine the correlation between rectal and TMD temperatures, 100 patients had simultaneous measurements at both sites. Mean initial temperatures were rectal, 37.1 +/- 0.5 degrees C (mean +/- S.D.); oral, 36.4 +/- 0.4 degrees C; and TMD, 37.4 +/- 0.4 degrees C. Maximal mean oral temperature change was greatest at 1.5 minutes after hot, +0.9 +/- 0.1 degrees C, and cold, -1.2 +/- 0.2 degrees C, water. This change was statistically significant for seven minutes at the 95% confidence level (analysis of variance test with Dunnett's multiple range test for significance). There was no significant change in the TMD temperature with any ingestion. The Pearson correlation coefficient for 107 pairs of rectal and TMD temperatures, r = 0.90 (P less than .001), was excellent. It was concluded that hot and cold liquids significantly influence oral temperature measurement for seven to nine minutes following ingestion. TMD temperature is unaffected by liquid ingestion and may allow accurate measurement of body temperature. Further studies are needed to determine the accuracy of TMD temperature over a wide range of body temperature in diverse clinical settings.