Results are presented for body temperature (A), plasma melatonin (B), wakefulness within scheduled sleep episodes (C), stage 1 sleep (D), stage 2 sleep (E), stage 3 sleep (F), stage 4 sleep (G) and REM sleep (H). Circadian waveforms are all double plotted with a resolution of 45 deg, except core body temperature and plasma melatonin waveforms, which are plotted with a resolution of 15 deg. Sleep-dependent main effects are plotted at quarters of the sleep episode (i.e. a resolution of 140 min). Sleep-dependent main effects of temperature and plasma melatonin are plotted with a resolution of 70 min. All parameters are plotted at the centre of the bins. Body temperature data (deviation from mean, °C) represent the fitted circadian component (left-hand panels) and the fitted sleep-dependent component (right-hand panels) as assessed by non-orthogonal spectral analysis. Plasma melatonin data are expressed as z-scores. Note that the first data point for the sleep-dependent modulation of body temperature and plasma melatonin represents the mean of these values during the first 35 min before lights out and the first 35 min thereafter. The s.e.m. is represented by the vertical bars and represents the between-subject standard error of the mean in each age group. Vigilance state parameters are expressed as percentage of total recording time (TRT) (wakefulness) or percentage of sleep time (stage 1, 2, 3, 4, REM).