A spectral analysis was made of the eight records of the 20 -isotherm 

 depth, one for each of the eight courses. The records were, of course, in 

 time but they have been interpreted in terms of distance on the reasonable 

 assumption that the ship's speed was constant. The wave-number range was 

 from zero to 6 cycles/km. 



The general form of the spectrum is such that the (amplitude)^ per unit 

 wave-number decreased with increasing wave-number, approximately as 

 (wave-number)^ . There is some indication that at the lowest wave-numbers, 

 below about 1 cycle/km, the falloff was less rapid. 



This spectral form can perhaps be interpreted in terms of a recent pro- 

 posal by 0. M. Phillips, based on the hypothesis that the amplitude of 

 internal waves is limited by shearing instability at their troughs and 

 crests. 



No clear directional property of the spectra was found - those from 

 all eight courses are virtually indistinguishable. This is consistent with 

 their monotonic form and provides an interesting negative result. 



If internal waves are generated by the nonlinear interaction of 

 surface waves then their directional spectra must be broad. No quantita- 

 tive Estimate of the directional spread seems to be possible. Equally, 

 since we have no clear idea of the damping and decay of internal waves, 

 they may be interpreted as a residue or continuum of internal waves 

 generated at other places and at other times. 



It must always be borne in mind that the analysis of records such as 

 these in terms of internal waves may not provide the most useful interpret- 

 ation. It may be more useful to consider them as turbulent fluctviation in 

 an hydro statically stable layer. 



Further analysis planned will attempt to assess the vertical structure 

 of the temperature fluctuations so as to provide information on the best 

 interpretation. This will be used to design ftirther observations using a 

 British thermistor chain and to make a quantitative estimate of the effect 

 of the thermal inhomogeneity on sound transmission in the ocean. 



A further accoxont of the investigation is being prepared for publication 

 elsewhere. 



178 



