Two groups of healthy subjects were studied indoors, first while living normally for 8 days (control section) and then for 18 x 27h 'days' (experimental section). This schedule forces the endogenous (body clock- driven) and exogenous (lifestyle-driven) components of circadian rhythms to run independently. Rectal temperature and wrist movement were measured throughout and used as markers of the amplitude of the circadian rhythm, with the rectal temperature also 'purified' by means of the activity record to give information about the endogenous oscillator. Results showed that, during the experimental days, there were changes in the amplitude of the overt temperature rhythm and in the relative amounts of out-of-bed and in-bed activity, both of which indicated an interaction between endogenous and exogenous components of the rhythm. However, the amplitude and the amount of overlap were not significantly different on the control days (when endogenous and exogenous components remained synchronized) and those experimental days when endogenous and exogenous components were only transiently synchronized; also, the amplitudes of purified temperature rhythms did not change significantly during the experimental days in spite of changes in the relationship between the endogenous and exogenous components. Neither result offers support for the view that the exogenous rhythm alters the amplitude of oscillation of the endogenous circadian oscillator in humans.