The authors made the review of a thousand of patients with clinical syndrome of epilepsy, in whom clinical and EEG studies were obtained. These one thousand cases were selected among 4,700 cases of epileptic patients with EEG diagnosis. From all cases, detailed history of the type of seizure, of past history (particularly obstetrical history, traumatic lesions and other cerebral lesions), of familial history, and of the beginning of convulsive seizures, have been obtained. In every case the onset of the disease was considered as the first convulsive seizure, even when between the first and the next there was a large period of time. In every case clinical, neurological and psychiatric examinations were obtained. In many of the cases spinal fluid and radiological (craniogram, pneumoencepbalogram and arteriogram) examinations were performed. The clinical manifestations were classified according to Penfield and col. These clinical features were compared with the EEG records (in 202 cases with barbiturate sleep and in 55 cases with metrazol), and projected in graphs and tables. In these the patients were classified according to age in three different groups (from 11 to 20 years of age, from 21 to 35, and with more than 36 years of age). From this study, the authors call the attention to the following data: 1. The EEG showed changes of lesional type in 36.8 per cent of the cases, of endogenous type in 32.5 per cent and did not show epileptic changes in 30.7 per cent. 2. Patients with lesional type of EEG could be classified according to age in the following manner: 39.8 per cent in the first group (from 11 to 20 years of age), 36.75 per cent in the second group (from 21 to 35 years of age) and 32.30 per cent in the third group (with more than 36 years of age). 3. The same patients could be classified according to age of the first seizure: 53.9 per cent in the group from zero to 19 years, 31 per cent in the group from 11 to 20 years, 29.1 per cent in the group from 21 to 35 years, and 31.8 per cent in the group with more than 36 years. 4. Patients with endogenous changes in the EEG were classified according with the age in the following manner: 42.25 per cent in the first group, 33.75 per cent in the second group, and 16 per cent in the third group. 5. These patients were distributed according to age of the first seizure: 30.20 per cent in the group from zero to 10 years, 41.50 per cent in the group from 11 to 20 years, 31.9 in the group from 21 to 35 years, and 8.5 per cent in the group with more than 36 years. 6. The patients with non epileptic changes in the EEG could be classified: 17.94 per cent in the first group, 29.5 per cent in the second and 51.7 per cent in the third. 7. These patients were classified according to the first seizure in: 15.9 per cent in the group from zero to 10 years of age, 27.5 per cent in the group from 11 to 20 years of age, 39 per cent in the group from 21 to 35 years of age and 59.7 per cent in the group with more than 36 years of age. 8. The EEG changes and the topography of these changes, when they were present, were coincident with the clinical syndrome in high percentage. 9. Bravais-Jacksonian seizures had negative EEG in high percentage of cases (36 in 91), in discordance with the other types of focal seizures. Temporal seizures were negative in 10 among 132 cases, and the other motor seizures with lateralization were negative in 18 among 125 cases. 10. GM seizures were the type with higher incidence of negative EEG (236 among 572 cases). PM seizures indicated the diagnosis of endogenous lesion (43 in 44 cases). 11. Temporal seizures presented lesional type of changes in the EEG in high percentage of cases (118 among 132 cases). These lesional type of changes were of focal type in temporal lobes in 86 among 118 cases. 12. The other types of motor seizures with lateralization showed high incidence of lesional changes in the EEG (88 among 127 cases). 13. The incidence of epilepsy among relatives was higher in lhe group of endogenous epilepsy (90 in 219 cases). 14. The neuro-psychiatric signs of cerebral organic diseases were coincident with lesional type of changes in the EEG in high incidence (103 among 130 patients). 15. The barbituric sleep was used when the routine EEG or the EEG with hyperventilation did not show evident changes. In this way it was possible to bring out positive findings in 65 among 202 cases. The barbituric sleep was helpful in 65 cases of temporal seizures and negative in 9. 16. The metrazol was used in 65 patients, with the same criteria. It was helpful in 46. 17. The majority of the 307 patients with EEG showing no epileptic patterns was older than 35 years. This probably indicates the existence of other co-factors (as chronic intoxications, anoxia due to cerebral arteriosclerosis, hypoglycemia, etc.), when cerebral lesions are not present, in the genesis of paroxistic disturbances.