Bilateral examination of supraorbital and frontal arteries in 26 patients were examined by the Doppler flow technique. In 70% of the patients, ipsilateral flow was diminished when compared to contralateral flow, during interim periods. An additional 5 patients were tested during the cluster attack and in 10 to 15 minute intervals, for I hour, following administration of sublingual ergotamine tartrate. Measurements of blood flow velocity revealed: ipsilateral decrese during the attack; marked increase 10 to 15 minutes following ergot administration; and a slow decline toward baseline by one hour. Doppler flow examination was supplemented by facial thermography in an additional 20 patients, of whom 7 had not used medication, 6 had been treated prophylactically without success, and 7 successfully treated. The results indicated that when treatment was successful, velocity values were higher on the ipsilateral side in most patients (p<.01), as were thermographic values (p<.001). Most patients, either untreated or unsuccessfully treated had lower Doppler and thermographic values ipsilaterally (p<.001), in agreement with data obtained earlier in this study. Although the results of this study clearly challenges the suggestion that the cluster attack is due to vasodilatation, it does not provide an explanation for pain nor establishes the presence of absolute vasoconstriction. It does suggest that extracranial vasodilatation and increased CBF observed during attacks, may be compensatory responses to internal carotid artery spasm.