[en] A correction is described for multiplication and moderation when doing passive thermal neutron coincidence counting nondestructive assay measurements on powder samples of PuO₂ mixed arbitrarily with MgO, SiO₂, and moderating material. The multiplication correction expression is shown to be approximately separable into the product of two independent terms; F/sub Pu/ which depends on the mass of ²⁴⁰Pu, and F/sub αn/ which depends on properties of the matrix material. Necessary assumptions for separability are (1) isotopic abundances are constant, and (2) fission cross sections are independent of incident neutron energy: both of which are reasonable for the 8% ²⁴⁰Pu powder samples considered here. Furthermore since all prompt fission neutrons are expected to have nearly the same energy distributions, variations among different samples can be due only to the moderating properties of the samples. Relative energy distributions are provided by a thermal neutron well counter having two concentric rings of ³He proportional counters placed symmetrically about the well. Measured outer-to-inner ring ratios raised to an empirically determined power for coincidences, (N/sup I//N/sup O/)/sup Z/, and singles, (T/sup O//T/sup I/)/sup delta/, provide corrections for moderation and F/sub αn/ respectively, and F/sub Pu/ is approximated by M₂₄₀/sup X//M₂₄₀. The exponents are calibration constants determined by a least squares fitting procedure using standards' data. System calibration is greatly simplified using the separability principle. Once appropriate models are established for F/sub Pu/ and F/sub αn/, only a few standards are necessary to determine the calibration constants associated with these terms. Since F/sub Pu/ is expressed as a function of M₂₄₀, correction for multiplication in a subsequent assay demands only a measurement of F/sub αn/