Measurement of air flow with variable orifice type flowmeters in laboratory experiments revealed discrepancies that could not be explained as errors arising from faulty calibration, leaks in the system, or failure to reduce readings to standard conditions of temperature and pressure. They were found to arise because of the occurrence of rapid pulsations in the air flow. A theoretical analysis of the effect showed that these instruments indicate root mean square rather than average flow rates. The results of experiments designed to test the theory were in reasonable agreement with the predicted values. It was found that errors arising from the occurrence of pulsations could be minimized by insertion of a suitable smoothing device between the pump and flowmeter. For example, an orifice with a pressure drop of 40 cm. of water reduced the flowmeter error to 5% of the true flow rate.