This study examined N2 and CO2 components of narcosis by comparing the effects of three levels of PETCO2 [low = 29 mmHg (SD = 4 mmHg), medium = 47 mmHg (SD = 1 mmHg), high = 57 mmHg (SD = 2 mmHg)] at 1 and 6 atm abs in 12 male volunteers. Cognitive and psychomotor performances were examined using a variety of tasks, including a modified Stroop test, an arithmetic test, number comparison, a figure copying test, and the Purdue pegboard test. Performance on all tasks demonstrated significant (P less than 0.05) decrements at 6 atm abs. High CO2 tensions significantly impaired cognitive and psychomotor performance at 1 atm abs and caused further decrements at 6 atm abs (P less than 0.05). However, no significant N2-CO2 interaction (P greater than 0.05) or global threshold for the onset of CO2 narcosis was indicated by the test scores. The pattern of intratest results were different for N2 and CO2. At high PETCO2, performance deficits were due to a slowing of performance rather than a disruption of the accuracy of processing. Nitrogen narcosis, conversely, produced significant impairment through both decreases in the speed and accuracy of processing on the majority of performance tests. It was concluded that within the PETCO2 ranges studied: a) PETCO2 and PIN2 are additive in their effects on impaired cognitive and psychomotor performance at depth; b) high PETCO2 and PIN2 induce distinctly different strategic responses on the speed accuracy trade-off function of the performance tests; c) decrements in cognitive and psychomotor performance under high PETCO2 do not conform to the predicted narcotic potency of CO2 according to the lipid solubility theory of narcosis.