Various methods have been developed to confirm proper intubation of the trachea. This blind, randomised study evaluates some of these quantitatively and qualitatively. Forty patients had both their trachea and oesophagus intubated. A procedure that included auscultation of the upper abdomen and lungs was 100% reliable independent of which tube was ventilated. Auscultation of the lungs resulted in a wrong conclusion in 15% (6-30, 95% confidence limits) of the cases when the oesophagus was ventilated: the sounds were misinterpretated as normal breath sounds. Suction on the tubes with a 60-ml syringe was also a reliable test. Other methods assessed were observation for condensation of water vapor, and abnormal movements of the upper abdomen; these were unreliable.