although the two records are separated by an interval of only 1.5 

 hours, Figure 5 contains distinctive large amplitude temperature 

 variations while Figure 6 reveals little activity and no apparent 

 correlation between stations. Figure 7 is an example of very well 

 defined temperature variations at two levels, clearly signifying the 

 passage of a regular wave form whereas the temperature changes 

 at three levels shown in Figure 8 are small and show little coherence 

 between measuring stations. 



The case illustrated in Figure 9 is particularly interesting because 

 there seems to be a temperature "node" at the 70-foot level. Inspection 

 of the mean thermal structure reveals a dual thermocline and the 

 measuring levels (50, 70, and 95 feet) are located such that if the 

 entire thermal structure moved down about 15 feet there would be no 

 temperature change at mid-depth while the upper and lower levels 

 would change as observed. The fact that the changes are in phase 

 significantly supports the argument for vertical migration of the 

 water column as a unit. In many other cases, however, this phenomenon 

 is by no means discernable. Lee (I960) has indicated the occurrence 

 of similar cases at the U. S. Navy Electronics Laboratory's tower in 

 56-foot depth off Mission Beach, California. 



Figure 10 is an example of close coherence between different levels 

 in a constant thermal gradient — another indication of synchronous 

 vertical movement of the water column. Two almost identical single 

 temperature crests occurring at the same level on two different days 

 are shown in Figures 11 and 12. Note that the mean thermal structure 

 has the same form on both days. The computed travel directions of 

 the two cases are somewhat different; viz., 335° for Figure 11 and 300° 

 for Figure 12. 



Comparison of individual signature forms recorded at each of 

 the three stations raises the question of uniformity. Temperature 

 amplitudes and "hump lengths" are not usually identical. Three 

 possible explanations are offered: (1) the internal waves are short- 

 crested relative to the 400- and 500-foot station spacings, (2) the 

 preset elevations of sensing elements at each station were not uniform 

 enough to initially be or to subsequently remain in horizontal planes, 

 and (3) elevations along individual internal wave crests were not always 

 constant. With regard to the second explanation, it is certainly unlikely 

 that the prefabricated buoys would have been exactly uniform or the 

 bottom perfectly flat but vertical tolerances of two feet in sensor 

 planes seem quite reasonable and any error of this kind, other than 



