In a surprising variety of cases, interfaces in normally ductile materials can undergo timedependent brittle cracking under the influence of a tensile stress, either applied externally or existing as an internal residual stress. The connecting feature in all these cases is the presence of a surface-adsorbed element that is highly mobile in comparison to the constituent elements of the material. As in the phenomena of diffusion creep and diffusive growth of cavities at high temperatures, the driving force for this cracking is the work done by the tensile stress when a surface atom enters the solid. At temperatures below about 0.5 Tm of the solid, this occurs mainly along grain boundaries. Examples of systems that have been studied in some detail include cracking of alloy steels by sulfur, Cu-Sn alloys by tin, and nickel-based alloys by oxygen. Because the cracking involves diffusive penetration along grain boundaries, the rate of cracking is highly sensitive to grain-boundary structure and composition, and these variables offer opportunities to control the problem. We are aiming at a quantitative understanding of the effects of grain-boundary structure, stress, and temperature on this phenomenon by crack-growth experiments on bicrystals, by atomistic modeling of the stress-driven diffusion, and by micromechanical modeling of the events occurring at the tip of a growing crack.