Volume magnetostriction is considered on the basis of the molecular field theory. Instead of the assumption of a saturation magnetization independent of volume at absolute zero (${I}_{0}$), made in the original interpretation by Becker and Kornetzki, another procedure is adopted. The dependence of saturation magnetization ($I$) on temperature for coupled ($j=1$) and uncoupled ($j=\frac{1}{2}$) electrons is assumed to be given by a Brillouin function. With the experimental data on the magnetostriction of iron this method allows us to obtain values for (a) the molecular field "constant" $N$, (b) its dependence on the relative change of volume, $(\frac{1}{N})(\frac{\ensuremath{\partial}N}{\ensuremath{\partial}\ensuremath{\omega}})$, and (c) the dependence on volume of the saturation magnetization at $T=0$, $(\frac{1}{{I}_{0}})(\frac{\ensuremath{\partial}{I}_{0}}{\ensuremath{\partial}\ensuremath{\omega}})$. The value of $N$ and $(\frac{1}{N})(\frac{\ensuremath{\partial}N}{\ensuremath{\partial}\ensuremath{\omega}})$ agrees with that obtained from the magneto-caloric measurements. The dependence of ${I}_{0}$ on volume agrees with the value expected from the change of density. Another check on the value of $(\frac{1}{N})(\frac{\ensuremath{\partial}N}{\ensuremath{\partial}\ensuremath{\omega}})$ is made by considering the "magnetic" specific heat, a part of which is shown, by thermodynamical considerations, to be proportional to this quantity. All quantitative checks of our theory, including the probable change of the Curie point under pressure are more satisfactory if electrons coupled in pairs ($j=1$) are assumed than if uncoupled ($j=\frac{1}{2}$) electrons are considered.