Before the introduction of computed tomographic (CT) scanning, aortography was the investigation of choice for acute aortic dissection. Between 1978 and 1982, 24 patients were referred to the Brompton Hospital with suspected acute thoracic aortic dissection; all had aortography with diagnosis confirmed at surgery (n = 12) or necropsy (n = 2) or supported by clinical outcome (n = 8). One patient in whom aortography was negative had type B dissection at necropsy and another patient was lost to follow up. CT scanning became available in this unit in 1983 and between 1983 and 1987 was used as the only imaging investigation in 32 patients with suspected acute dissection of the thoracic aorta while in a further 22 patients aortography was used alone. Results were confirmed at surgery (n = 18), necropsy (n = 3), or supported by clinical outcome (n = 31). Two patients were lost to follow up. In an additional 16 patients both aortography and CT scanning were performed with concordant findings in 10. In six in whom the results were discordant, aortography was normal in three in whom subsequent CT scanning showed type B dissection and CT scanning was normal in three patients in whom aortography showed type A dissection. Both CT scanning and aortography are reliable techniques for assessment of suspected acute dissection of the thoracic aorta. Both techniques misdiagnose occasionally and the frequency of misdiagnosis will be minimised by performing both investigations in patients where the level of clinical suspicion is high and the initial investigation negative. CT scanning tends to miss type A dissection and in view of the success of surgery in this condition this failing has the more serious clinical consequences.