The fact emerges (assuming the record itself to be accurate as to 

 time) that the vibrations start at intervals of about 50 minutes, with 

 the last and biggest group of evanescent oscillations starting almost 

 exactly 25 minutes after the Alaskan earthquake. These oscillations 

 are so obviously unlike any that would be ascribed to normal wave action 

 (being of impulsive, evanescent type) that some association with the 

 imminent earthquake mechanism seems implied. 



At Freeport , Texas, there is an equivalent interval of about 25 min- 

 utes between the time of occurrence of the earthquake and the development 

 of the water disturbance in the canal. Donn (196U) has remarked on this 

 case and indicated that the natural periods of transverse oscillation 

 across the channel are of the same order as the periods of Love (Q) and 

 Rayleigh (r) waves. McGarr (1965) bas investigated this case more closely 

 and evolved a very suggestive mechanism for explaining the disturbances 

 by the transverse vibration of the channel bed at the frequency of the 

 Love and Rayleigh waves. deBremaecker (Donn, I96U) had pointed out that 

 the Rayleigh waves would have had a period of about 16 seconds and double 

 amplitude of 15 centimeters. Donn remarks that the arrival time of these 

 R-waves corresponded to the development of the channel water oscillation. 

 On close examination of the Freeport record we find, however, that there 

 is evidence for underlying 2.25, 1.75, and 1.5 minute oscillations of 

 impulsive type which are not fully accounted for in McGarr 's otherwise 

 admirable analysis. The explanation for these longer oscillations, if 

 real, is not immediately apparent. 



Even at Freeport there is a suggestion of an impulse at almost 

 exactly 50 minutes prior to these disturbances, or 25 minutes prior to 

 the earthquake. Though we have diligently searched other readily avail- 

 able records of seismic seiches, (cf. Spaeth and Berkman, 19^7; McGarr 

 and Vorhis, I967) no other similar antecedent effect can be detected. 



The 25-minute interval between disturbances at Freeport, Texas and 

 Pensacola, Florida, and the curious occurrence of these oscillations 

 earlier than the earthquake suggests strongly that the earth was in its 

 second (football) mode of spheroidal oscillation (natural period 5^ 

 minutes - Press, 196^) with respect to the lunisolar axis, forced by 

 the lunisolar earth tide of 12.^4 hours, whose 15th harmonic would be 

 about 50 minutes. At the time (~01.30 GMT; March 28, 196U) that the 

 water oscillations start at Pensacola, the position of Pensacola in 

 relation to such a spheroidal earth oscillation would have been close to 

 the nodal circle on the moon side of the earth (Figure 30). Consequently, 

 if the spheroidal forced oscillation actually existed, the ocean water 

 in Pensacola Bay would have been subject to the maximum horizontal 

 accelerations across the node. 



This raises the point, that, had the earth been in such a spheroidal 

 (football) mode of oscillation with respect to the lunisolar axis, then 

 at the time the earthquake occurred, the hypocenter, moving with the 

 rotation of the earth on its axis, would have just reached the antinodal 

 position of maximum vertical acceleration (Figure 30). Further, since 



46 



