The objective of this study was to define the clinical semiology of seizures in temporal lobe epilepsy according to the age at onset. We analyzed 180 seizures from 50 patients with medial or neocortical temporal lobe epilepsy who underwent epilepsy surgery between 1997-2002, and achieved an Engel class I or II outcome. We classified the patients into two groups according to the age at the first seizure: at or before 17 years of age and 18 years of age or older. All patients underwent intensive video-EEG monitoring. We reviewed at least three seizures from each patient and analyzed the following clinical data: presence of aura, duration of aura, ictal and post-ictal period, clinical semiology of aura, ictal and post-ictal period. We also analyzed the following data from the clinical history prior to surgery: presence of isolated auras, frequency of secondary generalized seizures, and frequency of complex partial seizures. Non-parametric, chi-square tests and odds ratios were used for the statistical analysis. There were 41 patients in the "early onset" group and 9 patients in the "later onset" group. A relationship was found between early onset and mesial temporal lobe epilepsy and between later onset and neocortical temporal lobe epilepsy (p = 0.04). The later onset group presented a higher incidence of blinking during seizures (p = 0.03), a longer duration of the post-ictal period (p = 0.07) and a lower number of presurgical complex partial seizures (p = 0.03). The other parameters analyzed showed no significant differences between the two groups. We conclude that clinical and semiological differences exist between patients with temporal lobe epilepsy according to the age at onset. [Published with video sequences].