A silicon microreactor has been developed to investigate gas-liquid-solid catalytic reactions. The reactor employs a three-channel "catalyst-trap" design, whereby solid catalyst is suspended in the liquid channel by an arrangement of posts. Such a device has advantages in that commercial catalysts are supported, and that pressure drop across the bed can be reduced by engineering the packing density. A model incorporating the transport and kinetic effects is developed to design the reactor. The reaction chosen is one in a family of reactions relevant to the pharmaceutical industry, the liquid-phase hydrogenation of o-nitroanisole to o-anisidine. The reaction is carried out across a range of gas and liquid flow rates that encompasses three distinct flow regimes, termed gas-dominated, liquid-dominated, and transitional. Experiments seek to assign a reaction conversion and selectivity to each point in this two-phase "flow map", then reconcile differences in performance with the characteristics of the respective flow regime. We observe the highest reaction conversion in the transitional flow regime, where competition between the two phases results in the generation of a large amount of gas-liquid interfacial area. The experimental conversion is greater than that predicted by the model, an effect attributable to the mass transfer enhancement induced by transitional flow. Reaction time within the microchannel is sufficiently small so that selectivity towards production of o-anisidine is nearly 100% across all flow regimes. In this work we relate our map of conversion to the flow behavior and reactor geometry, and we discuss steps for further exploring the mass transfer characteristics of the transitional flow regime. This reactor architecture may be useful for catalyst evaluation through rapid screening, or in large numbers as an alternative to macro-scale production reactors.