Comparisons of calculation and experiment have been performed to test the adequacy of fewgroup subcritical diffusion theory in predicting neutron detector response induced from material changes in a subcritical water-moderated reactor. In many operations involving material changes in a subcritical reactor, it is desired to monitor changes in the multiplication factor (Keff) of the system to ensure the safety of an operation by avoiding an accidental close approach to criticality. This monitoring procedure is accomplished by the introduction of artificial neutron sources to the system and by the proper interpretation of changes in neutron detector readings in terms of Keff. Because of the rather complicated involvement of the source-core-detector system, proper interpretation of detector response observed during these operations can only be achieved by the availability to predict detector response obtained from an accurate calculational model. Comparisons of calculation to experiment show that diffusion theory may be used successfully for these purposes; however, certain limitations of the model must be recognized and avoided. The breakdown of the calculational model in certain cases can be related ultimately to the inability of few-group diffusion theory to predict the absolute magnitude of detector flux for large distances through a water (or metal-water) shield. This inability can result in inaccuracies in predicted count rate response when applied to a specific source-core-detector arrangement with the characteristic that a given material change results in gross changes in the axial flux distribution. These effects can be overcome by the suitable positioning of the neutron source and detector relative to the subcritical assembly.