A subcritical assembly (29 × 38 × 29 cm) built of TRIGA-type fuel elements was pulsed by coupling it with the Illinois TRIGA reactor through a graphite thermal column (2 ft square by 4 ft long). Flux measurements were made at seven locations in four different fuel loadings—9, 16, 25, and 49 fuel elements—with keff varying from ∼0.4 to 0.92. A polynomial expansion method is used to provide a continuous representation of pulse shapes. Derivatives appearing in a diffusion-theory model, evaluated using this expansion, are then used to determine the propagation velocity and the neutronic parameters. The maximum “asymptotic” velocity (removed from the boundaries) varied from ∼4 × 104 cm/sec at keff = 0.60 to 2.54 × 104 cm/sec at keff = 0.92. The theoretical model involves an expansion which, depending on the number of terms retained, bounds the experimental data. However, differences of as much as 40% in absolute values are observed and they are attributed to inadequacies in the model for this small heterogeneous assembly. Uncertainties in the neutronic parameters, as well as nonlinearities in the instrumentation, may also contribute.