Several methods are available for the determination of thermal diffusion length. In general, those based on one-group diffusion theory are subject to the assumption of a particular source boundary condition. Errors introduced by the assumption of such boundary conditions usually result in incorrect prediction of the relative harmonic content of the thermal flux at different elevations in the pile. The effect of these errors on diffusion-length determination is to cause inconsistency in the results as additional data points taken close to the source are included. A method is presented whereby the constants Amn in the one-group thermal-flux equation are determined experimentally and used in the determination of diffusion length. This method is then compared with other methods using one-group and age-diffusion theory with respect to the consistency of results obtained.