Sleep screening was used to discover the incidence of sleep apnoea in 50 children undergoing routine adenotonsillectomy for recurrent upper respiratory tract infections, randomly selected from the waiting list. Preoperative assessment included a detailed parental history, physical examination, and lateral cephalometry, in order to identify factors that might alert the clinician to a diagnosis of obstructive sleep apnoea. There were 2 equal groups of snorers and non-snorers (grade 0); 1 patient was found to have the sleep apnoea syndrome (IV), 9 patients had obstructive snoring with apnoeic episodes (III), 3 patients had snoring with a disrupted sleep pattern (II), and 12 patients snored with no disruption of sleep (I). In identifying patients with apnoea, a history of snoring was unhelpful, whereas one of breathing irregularities was found to be highly specific. Nasal obstruction correlated poorly; however, there was a significant relationship between tonsillar position and size and sleep grade (Chi-squared P less than 0.01). Stepwise regression analysis showed a large contribution to the grading was made by the size of the oropharyngeal airway measured by lateral cephalometry. The children in grade II-IV were re-studied 3 months post-operatively and all reverted to grades 0 or I.