A barometric method for measurement of ventilation of newborn infants is described. Experiments with cats are reported to show the degree of accuracy obtainable at present. The variation from standard methods was found to average±1 per cent. Minute volume, frequency, and tidal volume of premature infants measured by the barometric method are within the range of such measurements made by plethysmographic methods. The barometric principle deserves further study since it offers the possibility of securing ventilation data with a minimum disturbance to the infant.