Summary:The aim of our work was to investigate the interrelationships among rapid eye movement (REM) percent, latency, and density and their variations during different nights in an adequate REM-deprivation study. Twenty male subjects slept for 7 consecutive nights in the laboratory: 2 adaptation, 1 base line, 2 deprivation, and 2 recovery. Wilcoxon's test showed a significant increase in REM percent and a decrease in REM latency from base line to recovery nights. REM density on the contrary showed a decrease, possibly due to dilution caused by increased time spent in stage REM. Eye movement activity appears to remain stable after REM deprivation. Furthermore, REM density showed no correlation with other REM sleep measures, while its highly significant correlations across nights make it a reliable and stable descriptor of an individual's sleep pattern. The independence of eye movement activity with respect to other parameters of REM sleep is suggested.