A characteristic of oscillatory flow mixing in a baffled tube is that the residence time distribution performance can be affected independently of net flow conditions. The effects of both the oscillatory velocity and the throughput velocity on the residence time distribution performance has been investigated in a 24 mm diameter, 2.8 m long, baffled tube oscillatory flow reactor. The experiments were performed by applying the tanks-in-series model to 'perfect pulse' tracer experiments over a wide range of oscillatory conditions and flow rates. An optimum set of oscillatory and net flow conditions was found which resulted in near plug flow behaviour, usually 50 tanks-in-series (N) or greater. For a throughput velocity, N was found to be sensitive to both the amplitude and frequency of oscillation, and this dependence could be expressed by means of the oscillatory flow Reynolds number (Re-0), which combines both parameters. A unique value for Re-0 for each value of the net flow Reynolds number (Re-n) gave the closest approach to a plug flow RTD. To relate the oscillatory and net flows, a dimensionless velocity ratio, psi, was used, defined as Re-0/Re-n. In order to relate the RTD performance for different flow rates, the tanks-in-series model was non-dimensionalized by the use of a stage-wise efficiency term, eta, defined as the ratio of N to the theoretical number of stages. Over the range of oscillatory and net flow conditions studied, it was found that the range 2 less than or equal to psi less than or equal to 4 corresponded to the optimum RTD conditions, where efficiencies close to 1 were achieved. It was concluded that these dimensionless parameters were sufficient to select, a priori, the oscillatory parameters necessary to obtain the optimum RTD in an oscillatory flow reactor based on a desired throughput specification.