It has been a nearly universal practice among sleep researchers not to use the first night or two of recordings because of “adaptational artifact.” Since this custom is costly and might result in the inadvertent loss of valuable data, this study examines the records of the first two nights and compares them with records obtained in subsequent recordings (third and fourth nights). Thirty-five psychiatric inpatients were studied for four consecutive nights. Several diagnostic groups were represented: psychotic (N = 7) and non psychotic (N = 12) unipolar depressives and non psychotic bipolar (N = 5) depressives, schizophrenia (N = 7) and others (N = 4). Results indicate a striking constancy in nearly all sleep parameters when nights 1 and 2 were compared to nights 3 and 4 for the entire sample. Of the 26 sleep parameters investigated, only sleep latency differed significantly on the latter two nights. Similarly, no significant differences were found in comparing nights 1 and 2 to nights 3 and 4 in any of the diagnostic subgroups. The implications of these findings to inpatient and outpatient sleep research are discussed.