In the creation of surprise tsunami, submarine landslides head the suspect list. Moreover, improving technologies for seafloor mapping continue to sway perceptions on the number and size of surprises that may lie in wait offshore. At best, an entirely new distribution and magnitude of tsunami hazards has yet to be fully appreciated. At worst, landslides may pose serious tsunami hazards to coastlines worldwide, including those regarded as immune. To raise the proper degree of awareness, without needless alarm, the potential and frequency of landslide tsunami have to be assessed quantitatively. This assessment requires gaining a solid understanding of tsunami generation by landsides and undertaking a census of the locations and extent of historical and potential submarine slides. This paper begins the process by offering models of landslide tsunami production, propagation, and shoaling and by exercising the theory on several real and hypothetical landslides offshore Hawaii, Norway, and the United States eastern seaboard. I finish by broaching a line of attack for the hazard assessment by building on previous work that computed probabilistic tsunami hazard from asteroid impacts. ¿ 2001 American Geophysical Union