Eighteen continuous 48 h monitoring studies are reported from 17 patients with epilepsy. The numbers of epileptiform discharges over corresponding epochs of the 2 days were significantly positively correlated in 16 studies. However, this was explicable by masking due to the sleep/wake cycle and when waking and sleep were considered separately a minority of studies showed significant correlations. The difference in total 24 h production of discharges between the 2 days ranged from 1.3 to 30.3%, mean 15.1%. The maximum discharge rate in 75% of the studies occurred during sleep; during waking the distribution of discharges was random. Even in the waking state the 0.5 h discharge rate was extremely variable and in few patients could a single 30 min epoch be regarded as a reliable sample of the mean rate over the waking day. The intervals between events showed a Poisson distribution during 9 days and 5 nights, but there was no within-patient consistency between the first and second 24 h period. The occurrence of discharges was periodic significantly more often at night than during the day, but the periodicities did not clearly correspond to the REM cycle. Discharges increased overall during sleep in 14 studies, were unchanged in 2 and decreased in 4. The time of occurrence of maximal discharge rate during sleep was consistent from night I to night II only in patients exhibiting generalized regular spike-wave activity but random in the others. A negative correlation between antiepileptic drug levels and discharge rate was rarely observed.