Anatase-to-rutile transformation in titania system is a metastable-to-stable irreversible transformation [I, 2]. Transformations of this type are very important both from a scientific as well as from a technological point of view. It has been recently shown that sintering of titania during the anatase-to-rutile transformation is an effective way to achieve near theoretical densities at temperatures as low as 600 C [3]. On the other hand, if the objective is to use titania for catalyst support or membrane applications this transformation should be prevented or at least retarded to stabilize the initial pore-structure [2, 4-6]. This enhancement in sintering (porosity reduction) is an example of the well known Hedvall effect [7]. During almost all solid state transformations, atoms are more mobile than usual and this will enhance the mass transport rate and enhances solid state sintering. Many technologically interesting systems are metastable in the as prepared form when prepared through solution-sol-gel routes. Zirconia (tetragonal-to-monoclinic transformation) and alumina (transition alumina-to-a-alumina) are some other examples. It has been generally observed that [2, 8] the rutile crystallites are much larger than their precursor anatase crystallites in the initial stage of transformation. However, no possible explanation for this behaviour is given in these studies. In this paper, an explanation based on the thermodynamics and kinetics of nucleation-growth type of transformations is given for the observed difference between the crystallite sizes of anatase and rutile phases during the initial stage of thse phase transformation. Two situations are considered here, pure titania and the titania phase in the titania-alumina nanocomposites. It is postulated that the critical nuclei size of Wile phase may be much larger than the transforming anatase particles.