Annealing in oxidizing and reducing atmospheres results in different color and conductivity changes of the material as discussed in the next section. However, both show an increase in /3 at the expense of a. This indicates that the surface-rich f3 component does not arise from any particular oxidation state, but rather from anneal-induced relaxation. This view is supported by the fact that thin film specimens, deposited from liquid or vapor phase and cooled slowly (and therefore allowed to partially relax, show a higher fraction of (3 component than mechanically polished ones [Fig. l(c)].