In the design of irradiation experiments, prediction of the perturbation of the thermal-neutron flux by the experiment test specimens is frequently a problem. In order to obtain a model for calculating these perturbation effects, an experimental study was performed. A cylindrical shape was selected as being most typical of irradiation test specimens, and measurements of perturbation effects were made for several cylinders having different dimensions and made from different materials.Regression analysis was used to obtain polynomials from these measurements. These polynomials can be used to predict the flux perturbation, depression, and self-shielding factors as functions of cylinder materials and dimensions and control-rod position. The polynomials include a wide range of sizes and materials so that almost any cylindrical specimen that might be found in a typical irradiation capsule can be evaluated. In general, the error at the fitted points was only a few percent over the range of variables corresponding to most common materials and dimensions.In many cases, the polynomials have several distinct advantages over numerical models or mockup measurements. They are general (insofar as a cylinder is a typical shape for reactor test specimens), and they are simple to use since they require no computer calculations or reactor time. Also, their uncertainty can be established quantitatively (excluding the uncertainty due to reactor differences).These measurements were made in the Plum Brook test reactor. Because this is a typical reactor configuration (light-water moderated, MTR-type fuel elements with metal-to-water ratio of 0.75) much like many other reactors that can be found in industry and government today, these results should be generally useful. For reactors differing from this typical configuration, the results can be used to estimate flux perturbation effects after consideration is given to basic reactor differences. Also, we feel that the success we have experienced with this approach to the perturbation problem will be of interest to others confronted with this problem and having the facilities to repeat these measurements for their reactor.