''Menstrual'' migraine, a term misused by both patients and doctors, l
acks precise definition. This dissertation critically reviews papers o
n the subject and examines the problem from a clinical perspective. A
definition is proposed that the term ''menstrual'' migraine should be
restricted to attacks exclusively starting on or between day 1 +/- 2 d
ays of the menstrual cycle; the woman should be free from attacks at a
ll other times of the cycle. This definition, unlike many used previou
sly, links to a specific mechanism; the timing is consistent with oest
rogen withdrawal. If this is correct, ''oestrogen withdrawal'' migrain
e may be a better term. Future studies, necessary to support or refute
these proposals, are suggested.