Recent years have seen unprecedented outbreaks of avian influenza A viruses. In particular, highly pathogenic H5N1 viruses have not only resulted in widespread outbreaks in domestic poultry, but have been transmitted to humans, resulting in numerous fatalities. The rapid expansion in their geographic distribution and the possibility that these viruses could acquire the ability to spread from person to person raises the risk that such a virus could cause a global pandemic with high morbidity and mortality. An effective influenza vaccine represents the best approach to prevent and control such an emerging pandemic. However, current influenza vaccines are directed at existing seasonal influenza viruses, which have little or no antigenic relationship to the highly pathogenic H5N1 strains. Concerns about pandemic preparedness have greatly stimulated research activities to develop effective vaccines for pandemic influenza viruses, and to overcome the limitations inherent in current approaches to vaccine production and distribution. These limitations include the use of embryonated chicken eggs as the substrate for vaccine production, which is time-consuming and could involve potential biohazards in growth of new virus strains. Other limitations include the requirement that the current inactivated influenza vaccines be administered using needles and syringes, requiring trained personnel, which could be a bottleneck when attempting to vaccinate large populations in mass campaigns. In addition, the current inactivated vaccines that are delivered by injection elicit limited protective immunity in the upper respiratory tract where the infection process is initiated. Most of these limitations of the current vaccines are being addressed by research on novel approaches to vaccine development that are described in many of the chapters in this volume. As an introduction to the topic, H.L. Yen and R.G. Webster describe the reservoir of influenza viruses with pandemic potential present in aquatic birds, particularly focusing on the evolution of highly pathogenic H5N1 viruses in Asia. As these viruses have continued to spread geographically, they also continue to diversify genetically, raising a strain selection problem for vaccine development. However, A.C.M. Boon and R.J. Webby review recent studies that show that substantial levels of antigenic cross reactivity are exhibited among the surface antigens of H5N1 strains, and that they can elicit cross-protective immune responses. A better definition of antigenic epitopes involved in cross-protection will be an important advance in enabling the design of effective vaccines.