A shorter phase angle between habitual wake time and underlying circadian rhythms has been reported in evening types (E types) compared to morning-types (M types). In this study, phase angles were compared between 12 E types and 12 M types to verify if this difference was observed when the sleep schedule was relatively free from external social constraints. Subjects were selected according to their Morningness-Eveningness Questionnaire score (MEQ score). There were 6 men and 6 women in each group (ages 19-34 years), and all had a habitual sleep duration between 7 and 9 h. Sleep schedule was recorded by actigraphy and averaged over 7 days. Circadian phase was estimated by the hour of temperature minimum (T(min)) in a 26-h recording and by the timing of the onset of melatonin secretion (dim-light melatonin onset [DLMO]) measured in saliva samples. Phase angles were defined as the interval between phase markers and averaged wake time. Results showed that, in the present experimental conditions, phase angles were very similar in the 2 groups of subjects. However, results confirmed the previously reported correlation between phase and phase angle, showing that a later circadian phase was associated with a shorter phase angle. Gender comparisons showed that for a same MEQ score, women had an earlier DLMO and a longer phase angle between DLMO and wake time. Despite a significant difference in the averaged circadian phases between E-type and M-type groups, there was an overlap in the circadian phases of the subjects of the 2 groups. Further comparisons were made between the 2 circadian types, separately for the subgroups with overlapping or nonoverlapping circadian phases. In both subgroups, the significant difference between MEQ scores, bedtimes, and wake times were maintained in the expected direction. In the subgroup with nonoverlapping circadian phases, phase angles were shorter in E-type subjects, in accordance with previous studies. However, in the overlapping subgroup, phase angles were significantly longer in E types compared to M types. Results suggest that the morningness-eveningness preference identified by the MEQ score refers to 2 distinct mechanisms, 1 associated with a difference in circadian period and phase of entrainment and the other associated with chronobiological aspects of sleep regulation.