Heart rate (HR) was continuously monitored during successive 24-hr periods in 19 healthy subjects and 26 major depressed patients (DSM III-R). Recordings were performed after a 2-week wash-out period and the morningness or eveningness typology of each subject was determined. The chronobiological parameters and rhythm percentage (RP) were calculated by the single cosinor method from the smoothed HR curves of each subject. In normal subjects, HR follows a circadian rhythm (RP greater than 65%) with the lowest values at night. Morning type subjects have an earlier peak time (13:30) than evening type subjects (17:30). Major depressive patients were split into two groups; in the first one HR circadian rhythm was still present (RP greater than 63%) with a decrease in amplitude (24%) while in the second group, no circadian rhythm of HR could be detected (RP less than 25%, decrease in amplitude greater than 70%). In the group of patients with a persisting HR circadian rhythm, no veritable phase advance was observed. Our results suggest that circadian HR rhythm, which can be easily studied with non-invasive methods, might represent a chronobiological marker of some depressions. Given the lag that exists between the rhythms of morning type and evening type subjects, our study also stresses the importance of taking into account this behavioural trait in chronobiological studies.