Local cerebral glucose utilization (LCGU) was studied in 35 conscious rats after a fluid percussion injury (cerebral contusion). The experiments were divided into two parts. (I) Natural course: LCGU was quantitatively measured in 17 rats by the 14C-deoxyglucose autoradiographic method (Sokoloff et al. 1977) at 1, 2, 4 and 24 hours after injury. Sham operation was made in six rats. (II) DC-potential and LCGU: LCGU was studied with DC-potential and EEG monitoring in 12 rats. Results were as follows: (I) i) Glucose uptake was reduced at the center of the contusion in all cases. ii) Two different effects were observed in LCGU change two hours after injury: normal or slight increase in four of six rats (type A) and a remarkable increase in the cortex of the injured hemisphere in two of six rats (type B). iii) The pattern of increase in LCGU (type B) resembles that of cortical spreading depression. (II) i) Negative shift of DC-potential concomitant with EEG suppression in the injured hemisphere was observed frequently one to two hours after injury. ii) The increased LCGU pattern during DC-potential negative shift was identical with that of type B. iii) LCGU pattern without DC-potential change resembles that of type A. We concluded that the hypermetabolism occurring in the damaged cortex was due to a spreading depression. The findings obtained here should yield very important information concerning pathogenesis and treatment of human head injury.