Many studies involving computerized EEG analyses are based on the assumption that normal subjects constitute a homogeneous population, so that under specified conditions across-subject averaging is justified. Quite often EEG signals are processed with little or no concern for the state of the particular person from whom they are obtained. It is the purpose of this discussion to draw attention to the fact that only can normals have different EEG's one from another, but also that the same ones can display markedly different EEGs from recording session to recording session. A consequence of this situation is that there is always a large between-subject variability, and that the source of the signals needs to be carefully documented before generalizations can be made.