Migraine has long been considered as a "vascular headache" but clearly neurological mechanisms are involved. The pathophysiology appears to somehow involve serotonin, both peripherally and centrally, but its involvement may be just epiphenomenal. Adding to the enigma it is apparent that many of the presently available drugs for the treatment of migraine interact in one way or another with serotonin receptors. However, they tend to have a number of other unrelated actions and they are only of limited clinical value. Interestingly a promising new drug for the treatment of the acute attack, sumatriptan, has a very selective action as an agonist at a specific 5-HT1-like receptor sub-type, mediating vasoconstriction, which is localized on cranial blood vessels. Its action may, or may not, be independent of any involvement of serotonin in the genesis of migraine. Hopefully though, current attempts to determine sumatriptan's mechanism of action will shed further light on the pathology of migraine itself and the putative involvement of serotonin.