The pupillary light reflex (PLR) has been considered by some researchers to be responsive to changes in levels of sleepiness. However, no previous studies have tested this hypothesis with the dramatic variation of sleepiness across a complete circadian cycle. In this experiment, 20 normal individuals (age: 18-48 years) underwent a 27-hr constant routine, during which they were kept awake except for a 15-min nap opportunity every hour. Sleepiness was assessed both subjectively, by the Stanford Sleepiness Scale, and objectively, by a modified version of the Multiple Sleep Latency Test. The PLR in response to a flash of light was recorded every 2 hr, immediately before a nap period. Results showed that baseline pupil diameter became smaller with progressive sleep restriction, but there were no changes in any of the parameters of the PLR despite significant fluctuations in sleepiness. Some of the changes in the PLR were significantly related to changes in baseline pupil diameter. When baseline diameter was partialled out, there was still no effect of sleepiness on the PLR. The results suggest that the PLR cannot be used as a measure of sleepiness.