Abstract For several inorganic glasses predictions have been made of the homogeneous crystal nucleation rate using classical nucleation theory. However, in none of these cases were comparisons made with experiment, due primarily to the inability of being able to guarantee homogeneous conditions. Evidence has been provided that crystalline formation in Li 2O · 2 SiO 2 glass may initiate by a homogeneous mechanism. Thus, we have computed the nucleation rate curve of lithium disilicate crystals in this glass. It is found that not only do all forms of the classical theory predict nucleation rates many orders of magnitude smaller than those observed, but also that the temperature dependence of the theoretical rate is quite different from that observed.