Ambulatory blood pressure was recorded and analyzed in 13 migraineurs during headache-free periods and in 11 healthy subjects. Systolic blood pressure, mean blood pressure, diastolic blood pressure, pulse pressure, and pulse rate were recorded for 48 hours (three times every hour from 6 AM to mid-night, once an hour at night). Circadian variation of blood pressure was analyzed using single cosinor analysis and group mean cosinor analysis methods. Single cosinor analysis identified significant circadian rhythm of systolic blood pressure in 10 of 11 control subjects (90.9%) and 5 of 13 migraineurs (38.5%, P < 0.02 versus controls, Fisher's exact test). Incidences of significant circadian rhythm of mean blood pressure and diastolic blood pressure were 100% and 100% in controls, 46.2% and 53.8% in the migraineurs (P < 0.01, P < 0.02 versus controls). Incidences of significant rhythm of pulse pressure were 36.3% in controls and 38.5% in the migraine group (difference was not significant). Group mean cosinor analysis identified significant circadian rhythm in both the migraine group and the controls. The MESOR (midline estimating statistic of rhythm) values of systolic, mean, and diastolic blood pressures showed no significant differences between the migraine group and the controls. Acrophase amplitudes of systolic, mean, and diastolic blood pressure were 4.2, 5.2, and 5.9 mm Hg in the migraine group, respectively; and 7.2, 7.3, and 7.5 mm Hg in the controls, respectively. These amplitudes of systolic, mean, and diastolic blood pressures in the migraine group were significantly smaller than those in the controls. These data suggest that some migraineurs lose or alter their circadian blood pressure rhythm. Evaluating migraineurs as a group, significant circadian rhythm of blood pressure can be identified and oscillation amplitudes of blood pressures are decreased. The present results suggest that migraineurs may be subject to dysfunction of the circadian rhythm generator and the autonomic nervous system. Possible involvement of serotonergic projections from the raphe to the suprachiasmatic nuclei of the hypothalamus in migraine is discussed.