The recent renewed interest in respiratory analeptic action in chronic respiratory failure raises the problem of the site of action and nature of these drugs. They lead to a direct or reflex increase in central nervous activity, which in consequence increases respiratory muscle activity. Indeed, patients with respiratory failure develop pressures two or three times greater than normal subjects to achieve adequate ventilation. During acute exacerbations the pressures generated are five or six times as large. This increase in pressure may risk precipitating fatigue. In general the anomalies of blood gases seem much more related to changes in ventilation-perfusion relationships than to global hypoventilation. Finally, a large increase in the force of muscular contraction may only have a minor effect on ventilation, given the alinear nature of the dynamic compliance curve. An increase of central respiratory activity may thus be ineffective and possibly even dangerous. In the overall assessment of drug considered as an analeptic one always ought to consider the possible non-analeptic actions on the bronchial calibre, the force of the respiratory muscles and the ventilation-perfusion relationships. Thus, it is possible that the therapeutic effects of certain analeptic drugs may be due to actions other than those of respiratory stimulation.