Tidal volume can be measured by recording the chang-in impedance of the thorax caused by breathing. Thus breathing can be recorded quantitatively from two electrodes placed on each side of the chest. The method is especially useful in breathless patients and studies of breathing patterns during sleep. A patient with a long circulation time due to left ventricular failure had periodic breathing both awake and asleep. When asleep the respiratory minute volume was considerably reduced, and there were long apnoeic periods. This was associated with a rise in mean right atrial pressure until the patient woke with an attack of paroxysmal dyspnoea. The right atrial pressure dropped to normal when the patient breathed more regularly. Anoxia probably caused the rise in right atrial pressure. Periodic breathing disappeared when the circulation time shortened after treatment of the heart failure. The possible mechanism of these changes is discussed.Until recently it has been difficult to study breathing quantitatively without influencing its pattern by the method of recording. Face The potential difference developed across the chest. measured at a second pair of electrodes, is amplified by a high frequency amplifier, and rectified and smoothed to give a direct voltage. This voltage is made equal to a reference voltage at the end of expiration. Inspiration increases the resistivity of the tissues, causing the direct voltage to exceed the reference voltage. This difference is amplified and displayed through the pen recorder, and is proportional to the inspired volume. The electrodes used were brass strips, 6-25 x 1-25 cm., with a 0-2 cm. diameter stainless steel tube soldered longitudinally to the centre of the strip to make a ridge. These are similar to the electrodes described by Hamilton, Beard, and Kory (1965). The position of the electrodes is important (Baker, Geddes, and Hoff, 1966), and the best position in each patient is found by experiment. In this patient, the two electrodes were placed longitudinally on either side of the mid-axillary line with the lower poles at the level of the ziphisternum. They were secured with adhesive plaster. The system is calibrated at the start and finish of each period of observation with a low resistance spirometer recording through a potentiometer mounted on the pulley. The calibration is affected by the position of the patient, which must remain the same during recording. The change in impedance is related in a curvilinear way to the tidal volume (Fig. 1). With optimum placing of the electrodes this curve approximates to a straight line. In this patient the correlation was approximately r= 0-97. 550