Middleton Island on the edge of the Continental Shelf (see Figure 

 33 for bathymetric details) occupies a unique position in relation to 

 the tsunami-generating area. According to Brown (196^+), the initial 

 regression of the sea there was observed to last about 15 to 20 minutes, 

 and to drop about 2 to 3 feet (Chance, I968). A wave arrived some 20 

 minutes after the quake but rose no higher than the tectonically ele- 

 vated higher high-water line (Chance, I968). St. Amand's observation, 

 on the other hand, found swash marks at an elevation 10 feet above the 

 latter level (Brown, 196^+). 



In Figure 3^, we place these facts within a visual framework which 

 shows a cross-section of the Continental Shelf between Middleton Island 

 and Montague Island, normal to the hinge line of earth movement (section 

 AA' in Figures 8 and 33). The vertical uplift of the land and seabed in 

 this region, as inferred from Figures 8 to 11, is shown, together with an 

 estimated initial elevation of the sea surface, whose crest is displaced 

 somewhat to the right (southeast toward Middleton Island) because of the 

 horizontal thrust of the land during the earthquake. Although the water 

 would have risen at once about 1.7 meters (ab) at Middleton Island, the 

 greater uplift of about 2.5 meters (ac) of Middleton Island itself caused 

 the water actually to drain from around the Island. This regression 

 would have lasted about 20 minutes until the point f on the tsunami front 

 advanced to the elevation c (Figure 3^). Although St. Amand's observa- 

 tion e might suggest extremely high waves, it is probable that the swash 

 marks are the result of the runup in a concave embayment and thus re- 

 flect an amplification of the true wave height. Brown (l96i+) records 

 that two consecutive waves approached Middleton Island about an hour 

 after the earthquake; these were about three feet high and separated by 

 about half a mile of distance and one minute in time. Undoubtedly these 

 waves were secondary to the main tsunami and were probably parasitic. 



Figure 3^ suggests that the crest elevation (above normal sea level) 

 of the tsunami at its source near Montague Island, may have been about 5 

 meters (l5-l6 feet). This crest elevation implies a nominal wave height 

 of 30 feet. Southwest of Montague Island the larger vertical uplift 

 (>50 feet) and horizontal land thrust (>60 feet) probably would have 

 created waves about twice as high. 



At Cape Yakataga, Charles Bilderback recorded the arrival of the 

 waves and noted their levels with reference to landmarks, and so was able 

 to assess their runup height (cf. Berg, et al , I96U; Pararas-Carayannis , 

 1965; Chance, I968). Figure 35, a plot of Bilderback 's observations, 

 suggests that the underlying dominant wave had a period of about 1.5 

 hours and a height of about 8 meters (26 feet). This plot shows that in 

 the period range from 5 to 20 minutes, local oscillations increased the 

 height by almost 2 meters. As we shall see later, these and other wave 

 periods are a common feature of tide gage records of the tsunami. 



At Puget Bay and Whidbey Bay on Kenai Peninsula (Figure l), close 

 to the hinge line of zero vertical earth movement, waves up to 35 feet 



50 



