Fetal and neonatal hypoxia is suggested as a noticeable risk factor related to the occurrence of subependymal hemorrhages. To investigate the cause of newborn subependymal hemorrhages, a 0 day old newborn rabbit was exposed to either 5 or 10 minutes hypoxia by N2 box. After exposure, the brain was observed by optical and electric microscope using horseradish peroxidase (HRP) as the protein tracer to study the change in the capillary permeability. The result obtained were as follows: 1) In the newborn rabbit blood gas analysis, PO2 (mmHg) decreased more in the 5 and 10 min hypoxia groups than the control group which used the N2 box. pH and BE changes were similar to PO2 (mmHg). On the other hand, PCO2 (mmHg) increased proportionately. These results indicated that this hypoxic method changed blood gas and caused acidosis. 2) In the optical microscopic examination, the 0 day old newborn rabbit brain was found to have a thin subependymal layer. We could not find obvious subependymal hemorrhage using optical microscope. 3) In the electric microscopic examination, HRP was found in the cavity of the brain capillary at the subependymal layer and was slightly incorporated into the pinocytic vesicle of the lumen on the control. Since the tight junction filled the roll out, HRP was not found outside the capillary. 4) The 5 min hypoxia group caused astrocyte foot swelling, edema and slight opening of the tight opening of the tight junction. HRP was passed through the tight junction, but not outside the capillary. 5) The 10 min hypoxia group caused destruction of astrocyte and edema around the capillary. The tight junction was more opened than in the 5 min hypoxia group and HRP leaked out of the capillary through the tight junction. From the findings mentioned above, it is suggested that our hypoxic model increased permeability of the capillary in subependymal layers and that this change might be the first ultrastructural change before the onset of subependymal hemorrhages.