As described in report 1, I obtained a vague idea about the relationship between the barbital concentration in blood and sleep in my previous experiments. In order to clarify this relationship still further, I used the δ-index of E. E. G. for the numerical representation of depth of sleep, and examined the possibility of a parallel relationship between this index and the barbital concentration in blood. The results revealed that ths two fluctuate in parallel up to one to two hours after oral administration of barbital but beyond two hours the δ-index decreased rather independently of the barbital concentration in blood. From this result, the brain seems to posess a resistant (a familiarized phenomenon) at this stage to barlital, and therefore, the beginning of a habitual tendency had apparently been created in the brain.