Two hundred and fifty patients were assessed preoperatively using the Mallampati classification and by measuring their thyromental distances. The ease or difficulty of direct laryngoscopy was assessed at the time of induction of anaesthesia. Retrognathia was seen in 15.6% of patients and the incidence of difficult laryngoscopy without external laryngeal pressure was 8.2%. It was found that both assessments predicted less than two in three difficult laryngoscopies and had high false positive rates. It was found that external laryngeal pressure often improved the view of the glottis in difficult laryngoscopies.