Seizures can evolve sequentially into different clinical phases. For example, a seizure may start as an aura (first phase), then evolve into a tonic seizure (second phase), and evolve further into a generalized tonic-clonic semiology (third phase). It is currently unknown whether specific seizure evolutions cluster at particular times of the day and/or during sleep/wakefulness. We aimed to describe the distribution of the clinical evolution of seizures across time of day and sleep/wake state. We included all patients with at least two seizure phases admitted for long-term electroencephalogram monitoring during a 5 year period. Two-hundred-and-fifteen patients (866 seizures) presented with two different phases and 87 patients (324 seizures) evolved into a third clinical phase. During phase two, evolution into clonic seizures differed across time (p = 0.047) with peaks at 0-3 h and 6-9 h and during sleep (p < 0.001), evolution into automotor seizures peaked during wakefulness (p = 0.015), evolution into tonic seizures differed across time (p = 0.005) with peaks at 21-12 h and during sleep (p = 0.0119), and generalized tonic-clonic seizures peaked during sleep (p = 0.0067). Findings remained statistically significant after multivariable analysis adjusting, separately, for potential confounders (semiology of the first phase, age, gender, days in the long-term electroencephalographic monitoring unit, abnormal neuroimaging, number of antiepileptic medications, and seizure localization). During phase three, seizure evolutions followed the same pattern of distribution as during phase two but differences did not reach statistical significance. Our data demonstrate that the evolution of seizures into different phases cluster at specific times of day and at specific phases of the sleep/wakefulness cycle.