This study was designed to examine the basal cyclic AMP levels and the $10^{-5}mol/L$ isoproterenol-stimulated cyclic AMP levels of lymphocytes, by which ${eta}$-adrenoceptor function was shown, between to normal controls and 17 drug free patients(8 major depresive patients and 9 dysthymic patients), who were diagnosed by DSM-III-R. The severity of depression was assessed by Hamilton Rating Scale for Depression (HDRS). Cyclic AMP levels were measured by radioimmunoassay(double antibody). The results were as follows ; 1) HDRS score was significantly higher in major depressive patients($41.8{pm}4.6$) than in dysthymic patients($24.0{pm}4.2$)(p<005). 2) There was no Significant difference in basal cyclic AMP levels among normal controls($3.9{pm}1.7pmol/10^6cells/10min$), major depressive patients($2.1{pm}0.5pmol/10^6cells/10min$), and dysthymic patients($3.9{pm}1.8pmol/10^6cells/10min$). 3) There was significant difference in net cyclic AMP levels($10^{-5}mol/L$ isoproterenol-stimulated cyclic AMP levels minus basal cyclic AMP levels) among normal controls($16.5{pm}6.0pmol/10^6cells/10min$), major depressive patients($3.0{pm}1.4pmol/10^6cells/10min$), dysthymic patients($10.9{pm}4.4pmol/10^6cells/10min$)(p <005). 4) The net cyclic AMP levels were significantly correlated with HDRS scores in major depressive patients(${gamma}=-0.8^6$, p<0.05), but not in dysthymic patients(${gamma}=0.43$, p=0.25). In conclusion, we suggested that the dysthymic disorder might differ from the molar depressive disorder not only in the severity of depressive symptoms but also in ${eta}$-adrenergic responsiveness of lymphocytes, which was regarded as a biological marker of depressive disorder.