A macroscopic model has been recently proposed which permits the calculation of the formation and activation volumes of defects. It is shown that this model can adequately explain the unusual negative activation volume ${v}^{\mathrm{act}}$ observed recently in $B\ensuremath{-}4$ AgI (and earlier in cerium). Furthermore, it can quantitatively explain the large (\ensuremath{\sim} 300%) temperature variation of ${v}^{\mathrm{act}}$ from room temperature up to 400\ifmmode^\circ\else\textdegree\fi{}K. The latter justifies an earlier proposal of the authors that the thermal-expansion coefficient of the formation (or activation) volume of a defect can exceed the bulk coefficient by an order (or orders) of magnitude.