This study was undertaken to determine whether the internal clock contributes to the hormone and metabolic responses following food, in an experiment designed to dissociate internal clock effects from other factors. Nine female subjects participated. They lived indoors for 31 days with normal time cues, including the natural light: darkness cycle. For 7 days they retired to bed from 0000 h to 0800 h. They then underwent a 26-h 'constant routine' (CR) starting at 0800 h, being seated awake in dim light with hourly 88 Kcal drinks. They then lived on an imposed 27-h day (18 h of wakefulness, 9 h allowed for sleep), for a total of 27 days. A second 26-h CR, starting at 2200 h, was completed. During each CR salivary melatonin and plasma glucose, triacylglycerol (TAG), non-essential fatty acids (NEFA), insulin, gastric inhibitory peptide (GIP) and glucagon-like peptide-1 (GLP-1) were measured hourly. Melatonin and body temperature data indicated no shift in the endogenous clock during the 27-h imposed schedule. Postprandial NEFA, GIP and GLP-1 showed no consistent effects. Glucose, TAG and insulin increased during the night in the first CR. There was a significant effect of both the endogenous clock and sleep for glucose and TAG, but not for insulin. These findings may be relevant to the known increased risk of cardiovascular disease amongst shift workers.