The diffusivity of Cu2+, as determined by previous authors from analysis of experimental data in terms of the shrinking core (SCM) and linear absorption (LAM) models, is examined in light of the ability of the models to curve fit all the data. It is concluded from this further analysis that previous conclusions depicting the LAM to have an advantage over the SCM for predictive value are not justified. It is also shown that equally good curve fits can be obtained with a recent absorption/desorption model of diffusion which considers directly, through distribution theory, the effect of heterogeneity of material properties on the rate of diffusion.