Heart rate (HR), blood pressure (BP) and autonomic nervous system (ANS) activity vary diurnally, with a reduction in HR and BP, and a shift to vagal dominance during the dark phase. However, the cause of these changes, particularly the relative influence of sleep and circadian mechanisms, remains uncertain. The present study assessed the effect of sleep onset on HR, BP, high frequency (HF) component of heart rate variability (HRV), low frequency/high frequency (LF/HF) ratio and pre-ejection period (PEP). Sleep onset was dissociated from circadian influences by having subjects go to sleep at two different circadian phases, their normal time of sleep onset (normal sleep onset, NSO), and after a delay of 3 h (delayed sleep onset, DSO). The assumption was that changes caused by sleep onset would occur in association with sleep onset, irrespective of its timing, while circadian effects would have a consistent circadian phase and be independent of when sleep onset occurred. Thirteen and 17 subjects were run in the NSO and DSO conditions, respectively. Following a 1-h adaptation period, data collection began 2 h before subjects' normal time of sleep onset and continued until morning awakening. The lights were turned out after 2 h in the NSO condition and 5 h in the DSO condition. Subjects were required to maintain a supine position throughout the experimental sessions. The night-time decrease in HR was found to be due to both sleep onset and a circadian influence, with the circadian component being more prominent. In contrast, the fall in BP was largely due to a sleep onset effect. Increased vagal activity, as reflected in the HF component and a shift to vagal dominance in the LF/HF ratio, appeared to be primarily a function of the sleep system, while sympathetic activity, as assessed by PEP, reflected a circadian influence.