Ventilatory measurements especially in preterm infants are hampered by the additional apparatus dead space (face mask, pneumotach, shutter, ...). The apparatus dead space can be higher than the physiological dead space and affects gas exchange and blood gases significantly. For lung function testing in premature or term infants a background flow (flow-through technique) is frequently used to eliminate the apparatus dead space. This background flow must be higher than the maximum inspiratory flow (maxVin). For its optimization ventilatory measurements were performed in 62 sleeping infants aged 2 to 28 days (birth weight 760-5000 g). maxVin correlated strongly with minute ventilation VE (r = 0.90; p < 0.001). The ratio maxVin/VE decreased significantly (p < 0.05) with increasing frequency: from 4.4 +/- 0.8 (f < 40/min); 4.1 +/- 0.7 (f = 40-59/min); 3.8 +/- 0.5 (f = 60-79/min) to 3.4 +/- 0.4 (f > 80/min). With increasing frequency the flow signal becomes sinusoidal and the ratio maxVin/VE tends to pi. Therefore, during tidal breathing the background flow should be at least three but not more than the six times VE to avoid an impairment of measurements.