In 40 patients middle cerebral artery trunk (M1) flow velocity was recorded just before 54 carotid angiography in 54 cases exhibiting vasospasm after aneurysm rupture. Angiographic vasospasm distribution was studied; cases of symptomatic vasospasm were noted and were compared with transcranial Doppler data. Angiographic vasospasm was present in M1 in 41/54 carotid angiograms. Postulating that all the cases of M1 angiographic vasospasm should be identified by transcranial Doppler, the theoretical sensitivity of TCD was 76%. In this series however the real sensitivity of TCD in vasospasm diagnosis was only 70%: besides 13 cases where vasospasm was not present in M1 (mainly after ACoA Aneurysm rupture), TCD failed to identify 3 cases of M1 angiographic vasospasm. Vasospasm may not be located in M1 even when severe and symptomatic (4 cases in this series). Transcranial Doppler remains a mediocre tool for identifying vasospasm after anterior communicating artery aneurysm rupture (sensitivity: 55%). Its reliability is better after internal carotid aneurysm rupture (sensitivity: 72%) and excellent after middle cerebral artery aneurysm rupture (sensitivity: 93%). In order to test the drugs or methods used to prevent or combat vasospasm, angiography has to be considered when during the vasospasm risk period TCD does not demonstrate vasospasm in M1, either in patients in whom clinical deterioration is occurring without other obvious explanation, or in all patients.