Résumé : We have shown that clonidine, infused i.v. during the second non-REM period, was significantly less REM sleep suppressant in depressed patients than in control subjects. We have named this procedure the 'clonidine REM suppression test (CREST)'. In this report, we compare in the same sample (15 patients with primary major affective illness, 10 normal controls, 15 patients with minor depression and 15 patients with generalized anxiety) the efficacy of the CREST to separate the major depressed patients from the control subjects with the efficacy of three currently proposed biological markers of depression, i.e., the latency of REM sleep, the dexamethasone suppression test and the clonidine growth hormone stimulation test. We found that the CREST had the highest efficacy and suggest that further studies with independent and larger samples of patients and controls are needed to confirm those preliminary results and establish if the CREST could provide a new biological marker of major affective disorders. © 1995.