Our first interest in the study of the sleep-state electroencephalographic recordings in brain tumors was stimulated by the hope of improving the surface localization found in the waking-state EEG and of discovering a hidden focus when the waking-state EEG was indefinite. These hopes, except in an occasional case, were not realized.1 However, as noted by Grossman et al.,2 the capricious behavior of the slow-wave focus in the waking state when the brain tumor patient went to sleep became intriguing and demanded an explanation. Grossman et al. found that the delta focus tended to persist in sleep when the patient had seizure phenomena; although this finding was in accord with the data previously published,1 there were too many exceptions for this to be the only crucial factor. Hence we were moved to study the problem more closely and to attempt a correlation between the actual operative findings and