Last winter, on the Children's Medical Service of the Massachusetts General Hospital, we had the opportunity of studying a few patients with epilepsy that were undergoing, as a therapeutic measure, prolonged periods of fasting. Since for some time it has been generally conceded that this method of treatment has met with considerable success in decreasing the number and severity of the epileptic seizures, we were interested in ascertaining, if possible, what metabolic changes occurred, coincident with improvement, that might reasonably serve as a basis for explaining the effect or influence of fasting on this disease. Our attention was directed principally toward the study of changes in the blood and urine. These observations included determinations of the certain chemical constituents both before, during and following fasting. The substance of this report will be presented under the three general phases into which the investigation resolved itself: (1) alterations of the various blood