The procedure, first used in an earlier work (Wilson, et al , 1962; 

 Wilson, 196U), depends on visually identifying the underlying wave sys- 

 tems, and on separating them from each other and the background of the 

 astronomical tide. As an example. Figure UUa, the marigram for Juneau, 

 may be separated into curves (b), (c), and (d), of which (b) may be 

 considered the characteristic signature of the tsunami and (d) a minor 

 oscillation related to the locality. The signature shows a beat system 

 of waves with periods close to 1.7 hours. The first beat has five waves, 

 the largest of which is a negative wave near the front of the beat. 



Referring back to Figure 27, the main flanking attack of the tsunami 

 on the North American coastline took place between Yakutat and Crescent 

 City. Analyses covering stations along this coast (Figures U3 to ^+9) 

 show the first five waves in a rather pear-shaped beat with the largest 

 modulation near the front. The shape of the beat modulation undoubtedly 

 relates to the shape of the source, and generally confirms the interpre- 

 tations of seabed disturbance in Figures 19, 3^, 36, and 37. Figure 

 27 shows that the waves striking Yakutat originated from a generating 

 area off Hinchinbrook Island, where the seabed deformation had the same 

 dumbbell shape as the beat of waves reaching Yakutat (Figure 9). 



The waves reaching Sitka came from the generating area just south- 

 west of Montague Island (Figure 27) where highest known uplift occurred 

 (Figure 12). Figure U5c shows an apparently corresponding prominence of 

 the forefront of the beat of waves of about 1.7 hours period. The maximum 

 envelope height of waves for Sitka was 12 feet. 



Further south, Prince Rupert, Canada, was hit by 7-foot waves 

 (maximum envelope height), Figure i46c . Figure 27 shows that these waves 

 originated off the southern end of Kenai Peninsula (Figure 36b). 



Port Alberni is of special interest. According to the analysis of 

 Figure ii7c , which attempts to infer what the tide gage would have recorded 

 had it not failed during the first waves , the maximum envelope wave height 

 may have been as much as 27 feet. Port Alberni suffered severe damage 

 from the tsunami (cf. Wigen and White, 196^+; White, 1966; Abernethy, et 

 al, I96U). Water levels for the first three waves were established by 

 observation, thus permitting some definition of the wave system by inter- 

 polation between these observations and the later marigram (Figure h^a) . 

 The waves approximate 1.72 hours in period and clearly must have gained 

 their extraordinary height through some near-resonance local phenomenon 

 since the corresponding maximiim waves at Victoria (Figure U8d), only a 

 short distance southeast along Vancouver Island, were only 3 feet high. 



Port Alberni lies at the head of a long inlet, about kO miles from 

 the mouth of Barkley Sound on the west coast of Vancouver Island. (Figure 

 68). The inlet, a narrow channel, varies in width from 1/2 to 1 mile 

 and in depth from about 100 fathoms at the mouth to less than 30 fathoms 

 at the head (White, 1966). The approximate natural period of oscillation 

 of such a canal may be calculated by assuming that the depth profile along 

 the length is parabolic. Thus from Lamb (1932) (see also Wilson, I966), 



91 



