In the past, febrile seizures were considered to be a sign of epilepsy, a disorder characterised by recurrent unprovoked seizures. Currently, febrile seizures are considered to be a benign seizure syndrome that is distinct from epilepsy. This distinction has been possible largely because of the epidemiological evidence which is presented here in the form of a two-part argument. If febrile seizures are epilepsy one might expect that: (1) following a first febrile seizure, the risk of a second febrile seizure should be similar to the risk of an unprovoked seizure (in fact, the risk of a recurrent febrile seizure is approximately 34%, whereas the risk of an unprovoked seizure after having had a febrile seizure is approximately 2% to 3%); (2) the factors that predict recurrent febrile seizures should also predict subsequent unprovoked seizures. From the available literature, young age at the time of the first febrile seizure and a family history of febrile seizures predict recurrent febrile seizures, but do not predict subsequent unprovoked seizures. By contrast, a family history of epilepsy, complex febrile seizures and neurological abnormality are associated with an increased risk of subsequent epilepsy but are not consistently associated with the risk of a recurrent febrile seizure.