The extent of resection was assessed in 45 temporal lobectomies for medically intractable epilepsy with mapped temporal lobe foci. Postoperative magnetic resonance imaging (MRI) in the coronal plane was used to quantify the extent of resection of superior lateral, inferior lateral, basal, and medial structures, including the amygdalohippocampal complex. A new 20-compartment model of the temporal lobe was used for this assessment. Blinded interobserver variability was minimal. Intraoperative measurements and maps routinely overestimated the actual extent of resection, especially of medial structures. One year after surgery, 70% of patients remained seizure-free (except for auras). Seizure-free outcome was accomplished despite varying degrees of resection, but was more likely achieved with more extensive resections in all compartments. Among patients with mesiobasal foci, seizure-free outcome correlated significantly with extent of resection of amygdalohippocampal complex. We conclude that assessment of extent of resection by postoperative MRI provides an objective basis of evaluating outcome after temporal lobectomy. It allows a rational approach to understanding of operative failures and is potentially useful in comparing efficacy of various surgical approaches.