A series of 150 pneumoencephalograms which included 24 h films was reviewed to evaluate previous contradictory reports concerning changes in ventricular size at 24 h and to establish whether it is useful to obtain these follow-up films. In normal patients the changes in the ventricular index one day following the pneumoencephalogram ranged between -4-1 to +6-4% of the initial index. Statistically significant dilatation of the frontal horns of the lateral ventricles was observed in both patients who had dementia and central atrophy, and in those with normal pressure hydrocephalus. This ventricular enlargement was associated with deterioration in neurological status in some cases. Since ventricular enlargement at 24 h occurs in patients both with normal pressure hydrocephalus and with cerebral atrophy with hydrocephalus, it cannot be used to differentiate between these conditions although the most marked changes have been seen in the former. Significant dilatation did not occur in patients without intracranial disease, with cortical atrophy without hydrocephalus, or in a miscellaneous group. In patients with a history of trauma, enlargement at 24 h occurred only when there was some other encephalographic abnormality. The changes in size of the ventricles were not related to the age of the patient, the volume of the injected air, or to the presnece of subdural air. There were other findings on the follow=up radiographs. Cortical atrophy was more clearly demonstrated than on the filling films. Dilatation of the temporal horns, of questionable diagnostic significance, was not infrequently observed 24 h after air injection. The mechanisms of these changes are discussed; no single theory provides a complete explanation.