The "post-lunch dip" is a common behavioral phenomenon, though perhaps a misnomer. Biphasic models of the human sleep tendency rhythm suggest an alternative explanation for the afternoon decline in alertness. Sleep tendency was measured with the Multiple Sleep Latency Test (MSLT) at 2-hour intervals in 16 volunteers from three age groups (ages 10-12, ages 16-17, ages 62-74 years) during a constant routine in which small meals were given each hour. Baseline scores showed no significant Time of Day effect, although a trend for an afternoon dip was present in the eldest group. During the constant routine, a significant Time of Day effect was found for the two older groups and not for the prepubertal group. The results indicate a midday increase in sleep tendency that is unrelated to food intake but that may be related to developmental or maturational processes.