A water quality model of a distribution system provides an alternative to extensive grab sampling of chlorine residuals and disinfection byproduct concentrations, but the accuracy of such models depends on the input of a chlorine-decay rate relationship and the specific rate constant for reaction with pipe-wall material. This has proven difficult to quantify in currently available distribution system modeling software. Unlike small coupons of pipe material that are used in other bench-scale tests of chlorine-decay rate, the pipe section reactor (PSR) proposed in this article uses an undisturbed section of the actual pipe material filled with test water and a variable-speed stirrer blade to simulate a range of water velocities. Results are given for chlorine-decay rate tests on old cast-iron pipe and new ductile-iron pipe. With further testing of the PSR, rate models and rate constants will be available for use in water quality models to replace empirical approaches. This should improve the accuracy of chlorine residual predictions, with less reliance on field calibration. Accurate predictions of chlorine residual throughout the distribution system are essential to determine where low residuals exist and thus to comply with the Stage 2 Disinfectants/Disinfection Byproducts Rule. Includes 19 references, tables, figures.