Fourteen healthy subjects (8 women, 6 men, aged 22-35 yr) were recruited. Each subject was exposed, in a counterbalanced order, to bright white light (BWL: 3000 lux) and to dim red light (DRL: <15 lux) at a 1-week interval. Light treatments were administered from 00:30 to 04:30 h during sleep deprivation. Salivary melatonin and urinary cortisol concentrations were measured as was core body temperature. Vigilance levels were evaluated by subjective estimates, maintenance of wakefulness tests (MWT), waking EEG recordings, and three performance tests. Under BWL melatonin secretion was suppressed and core body temperature was significantly higher than under DRL. The BWL and DRL conditions produced no difference in cortisol secretion. Significant effects of BWL treatment were found for the MWT and theta-alpha and beta-1 frequency bands of the waking EEG. There was no significant effect of BWL on subjective alertness and performance. Vigilance measures were similar under the two conditions for the tests performed 1.5 h after the end of light treatments. Overall, the findings suggest that bright light (BL) exposure in the first half of the night decreases EEG-defined sleep propensity but has only modest effects on other aspects of vigilance.