Precise measurement of gas-liquid interfacial surface area is essential to reactor design and operation. Mass transfer from the gas phase to the liquid phase is often a key feature that controls the overall process. Measurement of gas-liquid interfacial area is often made through a separate measurement of the gas holdup and bubble size with complex and/or sophisticated methods. In this work, an inexpensive method is presented for the simultaneous determination of both local gas holdup and bubble diameter. The method is based on the withdrawal of the air-liquid dispersion under non-isokinetic conditions and on bubble counting via a simple optical device. The method was calibrated in a bubble column with several withdrawal pressures using coalescing and non-coalescing media. During the same calibration experiment, gas holdup was also measured manometrically and individual bubble diameters were measured by a photographic method. With a vacuum pressure of 3 kPa, local interfacial area measured with the withdrawal method produced a relative error below 13%, compared to the manometric/photographic method. The method was then used to characterize local specific interfacial area in a bubble column under several operating conditions with coalescing and non-coalescing media. In coalescing media and with superficial gas velocities (v(g)) from 0.25 to 3.5 cm/s, the average interfacial area ranged from 17 to 197 m(-1). With non-coalescing media the average interfacial area ranged from 40 to 560 m(-1). Under the test condition it was observed that gas holdup is a parameter that has a greater distribution (standard deviation from 30% to 70%) than the volume-mean bubble diameter (standard deviation from 6% to 12%). It is shown that a model previously developed for characterizing gas holdup homogeneity is also suitable for characterizing interfacial area homogeneity. (c) 2007 Elsevier Ltd. All rights reserved.