The inactivation data for Escherichia coli are recorded for the three reactor geometries of Taylor-Couette flow and flow between either concentric cylinders or a square channel. All of the data are shown to be correlated with the assumption of plug flow. In particular, the effects of nonuniform radiation levels are accounted for by integration across the fluid channel as done previously. However, a new correction factor is introduced that is shown to be inversely proportional to the laminar, velocity boundary thickness to account for the effects of a concentration boundary layer of surviving pathogen. It has also been demonstrated that the common problems of nonuniform radiation levels and concentration boundary layer effects in UV reactors are largely eliminated with the use of Taylor-Couette flow. Moreover, the repetitive exposure of fluid parcels to a small number of lamps in the rotating Taylor-Couette flow decreases maintainance requirements compared to the hydrodynamic equivalent of cross-flow over a tube bank or lamp array. Over a 3-log reduction in the inactivation of E. coli was demonstrated compared to a conventional channel with the same radiation dosage. Moreover, greater than a 2-log reduction was evident compared to flow through concentric cylinders.