The hypothesis that sleep deprivation depresses immune function was tested in 20 adults, selected on the basis of their normal blood chemistry, monitored in a laboratory for 7 d, and kept awake for 64 h. At 2200 h each day measurements were taken of total leukocytes (WBC), monocytes, granulocytes, lymphocytes, eosinophils, erythrocytes (RBC), B and T lymphocyte subsets, activated T cells, and natural killer (NK) subpopulations (CD56/CD8 dual-positive cells, CD16-positive cells, CD57-positive cells). Functional tests included NK cytotoxicity, lymphocyte stimulation with mitogens, and DNA analysis of cell cycle. Sleep loss was associated with leukocytosis and increased NK cell activity. At the maximum sleep deprivation, increases were observed in counts of WBC, granulocytes, monocytes, NK activity, and the proportion of lymphocytes in the S phase of the cell cycle. Changes in monocyte counts correlated with changes in other immune parameters. Counts of CD4, CD16, CD56, and CD57 lymphocytes declined after one night without sleep, whereas CD56 and CD57 counts increased after two nights. No changes were observed in other lymphocyte counts, in proliferative responses to mitogens, or in plasma levels of cortisol or adrenocorticotropin hormone. The physiologic leukocytosis and NK activity increases during deprivation were eliminated by recovery sleep in a manner parallel to neurobehavioral function, suggesting that the immune alterations may be associated with biological pressure for sleep.