The traditional model of electromigration damage resulting from grain boundary migration has been modified to include surface diffusion effects. The model is based on detailed TEM and SEM observations taken before, during and after current stressing. Significant effects were caused by the thin oxide overlayer that grows naturally on Al and Al-rich alloys during air exposure. The stresses applied by a typical thick passivation layer would surely modify the experimental results in ways that depend crucially on the exact experimental situation. Examples are also presented that show that where grain boundary migration was hindered or absent, surface diffusion phenomena became the dominant damage mechanism. The observations presented or referred to in this review and the interpretation of them should prove useful in helping our understanding of more complex electromigration phenomena.