In depressions (especially endogenous depressions) the function of cerebral norepinephrine may be disturbed. This working hypothesis was repeatedly proposed and maintained. The authors relied chiefly upon the facts that reserpine and methyldopa can evoke depressive reactions in man and that antidepressive drugs of the imipramine type and inhibitors of monoamine oxidase increase the concentration of free norepinephrine in the neighborhood of amine receptors. Yet there are additional pharmacologic and clinical arguments in favor of the catecholamine hypothesis of the manic depressive psychosis; amphetamine and cocaine induce in nonaddicts a manic state and interfere with the function or metabolism of norepinephrine. The hypothetic underlying disturbance in depression must be in agreement with clinical observations: it must be effective only in the central nervous system and here only in one region; it must alternate and sometimes change from a minus into a plus variant (change from depression to mania); and it must connect manic depressive psychoses in some way with schizophrenia. The catecholamine hypothesis fulfills these conditions. The connection of manic depressive psychosis and schizophrenia is represented by amphetamine and cocaine psychoses. These drugs induce in a toxic dosage a paranoid hallucinatory syndrome. This syndrome is the best model of an acute paranoid schizophrenia. The catecholamine hypothesis is valid for depressions and manias and for some forms of schizophrenia.