An optical measurement system was used to investigate the effect of microchannel length and inlet geometery on adiabatic gas-liquid two-phase flow. Experiments were conducted with 146-mm- and 1571-mm-long, circular microchannels of 100 mu m diameter. Void fraction and gas and liquid plug/slug lengths and their velocities were measured for two inlet configurations for gas-liquid mixing: (a) reducer and (b) T-junction. The superficial gas velocity was varied from 0.03 to 14 m/s, and superficial liquid velocity from 0.04 to 0.7 m/s. The test section length was found to have a significant effect on the two-phase flow characteristics measured at the same axial location (37 mm from the inlet) in both microchannels. The mean void fraction data for the short (146 mm) microchannel with the reducer inlet agreed well with the equation previously proposed by Kawahara et al. (2002). On the other hand, the mean void fraction data for the long (1571 mm) microchannel obeyed the homogeneous flow model and Armand's equation for both the reducer and T-junction inlet configurations. Many long and rapidly moving gas plugs/slugs and long, slowly moving liquid plugs/slugs were observed in the short microchannel compared to the long microchannel, leading to the differences in the time-averaged void fraction data. The mean velocity of liquid plugs/slugs generally agreed well with Hughmark's equation and the homogeneous flow model predictions, regardless of the inlet configurations and microchannel lengths. Thus, both the microchannel length and inlet geometry were found to significantly affect the two-phase flow characteristics in a microchannel.