OBJECTIVE To define the term "menstrual" migraine and to determine the prevalence of "menstrual" migraine in women attending the City of London Migraine Clinic. METHODS Women attending the clinic were asked to keep a record of their migraine attacks and menstrual periods for at least 3 complete menstrual cycles. RESULTS Fifty-five women completed the study. "Menstrual" migraine was defined as "migraine attacks which occur regularly on or between days -2 to +3 of the menstrual cycle and at no other time". Using this criterion, 4 (7.2%) of the women in our population had "menstrual" migraine. All 4 women had migraine without aura. A further 19 (34.5%) had an increased number of attacks at the time of menstruation in addition to attacks at other times of the cycle. Eighteen (32.7%) had attacks occurring throughout the cycle but with no increase in number at the time of menstruation. Fourteen (25.5%) had no attacks within the defined period during the 3 cycles studied. CONCLUSIONS A small percentage of women have attacks only occurring at the time of menstruation, which can be defined as true "menstrual" migraine. This group is most likely to respond to hormonal treatment. The group of 34.5% who have an increased number of attacks at the time of menstruation in addition to attacks at other times of the month could be defined as having "menstrually related" migraine and might well respond to hormonal therapy. The 32.7% who have attacks throughout the menstrual cycle without an increase at menstruation are unlikely to respond to hormonal therapy. The 25.5% who do not have attacks related to menstruation almost certainly will not respond to hormonal therapy.