IT is well known that the relaxation of stress in glass during annealing does not follow the Maxwellian relaxation formula. Indeed, annealing schedules for glass are always based on Adams and Williamson‘s1empirical law that the reciprocal of the stress is proportional to the time. Littleton2 and Lillie3 have attempted to account for the discrepancy on the basis of the viscosity being a function of the time, and many workers have attributed it to the dependence of the stress relaxation on some ‘molecular viscosity' or upon changes in the equilibrium state of the glass.