An increasingly important metal joining procedure is based on “solid-state bonding”, which employs the application of temperature and stress to a metal-metal interface. Under ideal conditions[1], a sequence of metallurgical processes produces a bond whose mechanical properties are identical to those of the host material. This sequence involves the deformation of the interfaces to achieve intimate mechanical contact, atomic diffusion to achieve the “solid-state” bonding, and grain growth to remove the microstructural memory of the original interface. Standard NDE techniques can be used to detect discrete flaws in the bonds. However, even in the absence of discrete flaws, strength and fatigue resistance may be degraded by subtle differences between the microstructures of the interface region and the bulk material. This paper is motivated by a particular bonding process (diffusion bonding as applied to aircraft engine rotor components), and considers only one metallurgical cause of weak bonding (distributions of microscopic pores in the bond plane). Figure 1 presents a model problem which has been chosen to simulate the inspection of such bonds as they occur when turbine blades are joined to hubs. The model geometry consists of two rectangular plates of similar material which have been diffusion bonded in the xy plane.