Clark and Yarnall 



1) The sinusoidal character of the signatures in the acous- 

 tic records. (Fig. 16) . The signatures are similar to 

 the surface wave effects in that they appear, for most 

 occurrences, in both acoustic phase and amplitude. 



2) The spectral distribution of the signatures lies 

 within the range predicted for stability oscillations. 

 Typical curves of the stability frequency (Vaisala- 

 Brunt frequency) , calculated for local conditions, 

 appear in Figure 17. 



A study of short period internal wave signatures has been 

 undertaken for H3 hydrophone data only; we are thus concerned with 

 processes which occur in the first 3 MM of the Fowey Rocks to Bimini 

 acoustic path. The data are from the recent long continuous test 

 of September 1966 through January 1967 (LCT-1) . 



All data records from LCT-1, as in Figure 16, were produced 

 at a chart speed of 0.025 mm/sec, much too slow to resolve the high 

 frequency surface wave signatures. At this chart speed, periodicities 

 of a few seconds merely produce a widening of the trace. As a rule, 

 surface wave effects were less prominent at H3 than at H43 and on 

 many occasions were close to the resolution limits of the instrumen- 

 tation. But many of the short period internal wave signatures were 

 sufficiently well defined to permit determination of occurrences and 

 an estimate of duration and characteristic period by visual inspection. 

 These wave trains are not as persistent as those due to surface wave 

 effects and it was necessary to establish the minimum wave train to 

 be classified as a signature. Four clearly marked cycles of a given 

 period appearing in either acoustic phase or amplitude were arbitrar- 

 ily selected as this "minimum signature" . In practice the minimum 

 signature was not of significant occurrence. The mean "length" of the 

 wave- trains determined by an average over all of the signatures was 

 12 cycles. 



Distributions of the number of occurrences and the persis- 

 tence in time or total duration, of observed internal wave periods 

 are presented in Figure 18, Due to limitations of techniques employed, 

 each designated period in these figures should be understood to con- 

 tain a spread of perhaps 10% of the enumerated values. The similar- 

 ity of the two distributions suggests that the persistence of a signa- 

 ture is more closely related to its probability of occurrence than to 

 the length of its period. The dashed curves in Figure 18 illustrate 

 an interesting suggestion of symmetry that appears in these data. It 

 can be seen that minor peaks in the distribution seem to occur in the 

 vicinity of multiples of the large peak near 3 minutes. 



The distributions of Figure 18 are the result of a rather 

 crude technique of visual analysis. At best, these are very approx- 

 imate spectral descriptions of the short period internal wave signa- 

 tures observed during LCT-1. The data-handling problems and the 



323 



