Abstract
Both stereological and direct experimental procedures of estimating the distribution of interface normal orientations in a polycrystalline material are discussed in connection with the problem of intergranular damage heterogeneity. Two stereological procedures of obtaining the interface normal orientation distribution are compared. The mathematical analysis required in obtaining this distribution from stereological measures is briefly described. It is shown that the stereology incorporating the line length per unit area of interfaces, as seen on a test section, is superior to that which counts intersections per unit length of scan line. A discussion of the advantages and limitations of these techniques is included.