Water diffusion in polymers can often be approximated by a Fickian description, but a 2- phase model was proposed some years ago by Carter and Kibler (C&K), often referred to as “Langmuirtype” diffusion, by analogy with the Langmuir theory of adsorption. The two phases in question correspond to “mobile” and “bound” diffusant molecules. In this study, we have considered water uptake in an epoxy resin (an adhesive), employing gravimetry. A good, overall, empirical agreement with the C&K mathematical description of total mass increase with time has been obtained. In many applications of the C&K theory when used to quantify diffusion of water in polymers, only total water uptake is considered as a datum. However, a simple mathematical treatment of the theory enables the separate mobile and bound contributions to be isolated. These supplementary data have been used to try to get a better understanding of the meaning of the terms “mobile” and “bound” phases. Deuterium NMR analysis has been employed to study the mobility of the absorbed water. Decomposition of spectra has permitted us to assign two signals to the fractions of “mobile” and “bound” water. Analysis of peak evolution and a comparison with gravimetric data lead us to suggest that the “mobile” phase corresponds to diffusing molecules, whereas the “bound” phase corresponds to “clusters”.