We have studied the extent to which learning fibreoptic nasotracheal endoscopy first helped anaesthetists to learn fibreoptic orotracheal endoscopy later, and vice versa. After preliminary training on a bronchial tree model, 30 anaesthetic trainees were randomly allocated to the nasal first/oral second group, who performed 10 nasal intubations followed by 10 oral intubations, or the oral first/nasal second group, who performed 10 oral intubations followed by 10 nasal intubations, in anaesthetised, ASA group I or II patients undergoing elective oral or general surgery. Each type of endoscopy was taught in a standard manner, with the aid of an endoscopic video-camera system, under the supervision of experienced instructors. Performing nasal endoscopy second (average 70.8 s) took significantly less time than performing it first (average 84.4 s) and performing oral endoscopy second (average 35.2 s) took significantly less time than performing it first (average 48.5 s). The mean (SD) total endoscopy time for all the endoscopies (both nasal and oral) in the nasal first/oral second group [1196 (162) s] was not significantly different from that for all the endoscopies in the oral first/nasal second group [1193 (188) s]. Because there is no advantage or disadvantage to be gained in starting to learn either type of endoscopy first, graduated training programmes can be planned according to the availability of suitable patients for fibreoptic intubation, without instructors needing to consider whether trainees make better progress if they learn one technique before the other.