AbstractMethods of calculating the“Snell Experiment”(the exponential experiment in natural uranium) are examined. It is found that integral transport theory is required for accurate predictions. The effect of spatial transients upon measured quantities is studied and it is found that experiments have not been done in a large enough mass of uranium to achieve an asymptotic neutron distribution. However, deviations from the asymptotic values of integral quantities are not large, and corrections are calculated and applied to recent experiments. It is shown that the use of recent cross-section data improves the agreement between theory and experiment. The relaxation length and all spectral indices are in fairly good agreement except for Np237-to-U238average fission cross section ratio.