It was recalled that, since the seminal work of Read and Shockley, studies of grain boundary structure had assumed that hypothetical boundaries of infinite extent were adequate representations of real interfacial structures. It was noted that, while this was adequate for many interfaces, the assumption still required some investigation; especially with regard to nanocrystalline materials. Here, the circumstances under which so-called infinite boundary models were likely to produce misleading predictions were identified. These included the cases of very short boundaries that were terminated by triple lines, and of short boundaries which were terminated by free surfaces. In both cases, end effects were found to be potentially important and, in extreme cases, could even be the predominant feature of the boundary structure.

A.H.King, L.Balasubramanian: Materials Science Forum, 1995, 189-190, 143-8