An evaluation was made of results of electrocochleography (ECoG) in a group of 75 children, tested during the last four years in the ENT department of the Leiden University Hospital. Eleven years was taken as the upper age limit; 79% of our group were younger than three and a half years. In only 60% of the children the cause of the hearing loss was clarified. In this group maternal rubella followed by meningitis and retardation were the most frequent cases of deafness. Cochlear microphonics (CM) could be measured in all the children. Examples are given of cochlear responses of children belonging to 6 different threshold categories. The responses of the largest category (56 children) with the highest loss (larger than 80 dB) were very abnormal. For the categories with less hearing loss the ECoG was not restricted to threshold determination and the origin of the hearing loss could be demonstrated, while a contribution of conduction loss could be derived from a horizontal shift of the latency-intensity function to higher intensity levels. The potentialities and limitations of ECoG and brainstem electric response audiometry (BERA) for threshold determination are discussed, particularly against the background of the time consumption of both procedures and the invasive character of ECoG.