Water vapor corrosion of a simple soda-lime glass has been studied in regard to its effect on static fatigue of the same glass. A mechanism of dissolution has been proposed in which alkali ion self-diffusion controls the initial steps of water corrosion and leads to breakdown of the glass network. Since experiments show that an expansion of a glass network enhances corrosion rate, it is postulated that asymmetrical contitions of expansion around a surface flaw, brought about by applied stress, could lead to growth of the flaw in a preferential direction to bring about delayed failure.

An analytical model, which is applicable to static fatigue of lime glass, has been extended to account for dynamic fatigue of the same glass. The model successfully predicts the room temperature strain rate sensitivity of the failure process in lime glass and indicates a method by which the stress concentration relationship, applicable to microscopic flaws on glass surfaces, may be obtained by experiment. Resulting experiments showed that the stress concentration relationship proposed by Inglis is valid.

Static fatigue of a simple soda-lime glass has been investigated in relation to the sensitivity of this glass to atmospheric corrosion. An analysis of the failure process has been given which is based on the concept that inherent surface flaws grow by corrosive mechanisms to critical dimensions by virtue of a reaction between water vapor in the atmosphere and components of the glass. The rate of this reaction is determined by the stress conditions around local areas and the temperature, pressure, and composition of the surrounding atmosphere. Since the experimental work shows a close relationship between the temperature dependence of the failure process and that of the self diffusion of sodium ion in bulk glass it is concluded that alkali content is responsible for the very low long time strengths of most inorganic glasses.

Opalescence and clearing techniques were used to determine the metastable immiscibility surface for sodium borosilicate solutions. These results indicate that a three-liquid region, which may or may not be metastable to two-liquid regions, underlies the immiscibility surface.

Using the phase diagram and a simple solution model, a subliquidus miscibility gap was estimated for the B2O3‐SiO2 system. The predicted coexistence boundary, showing a consolute temperature of 520°C, was flat and symmetrical and extended across the complete binary. Gradient furnace heat treatments of selected compositions in this system resulted in phase separation which corresponded closely to the coexistence boundary initially predicted. Calculations and preliminary experimental results indicate that temperatures and compositions exist wherein metastable three‐liquid immiscibility occurs in R20–B203‐SiO2systems.