A growing body of evidence supports the nucleation hypothesis of fibrillar amyloid formation. In this article, it is hypothesized that the fibrils formed with human A beta, rodent A beta, and a mixture of the two peptides may form nearly identical physical structures with clearly different biological activities. Data is reported supporting the concept that a specific "strain" of nucleation seed could impart a new structure on a growing amyloid fibril, thereby changing its biological activity. The data that the biological activities of specific prion strains have a basis in strain-specific structure have been supported experimentally.