based on data taken i'rom one location; bherefore, they are 

 valid only in that area. 



3' The continuous density model 



So far the investigation has centered around a model 

 in which the wave occurs along the interface between two homo- 

 geneous layers of water. If more than one interface are present, 

 then several internal waves caa result. Concerning this event- 

 uality, the Oceans (1932) says "...the greater the number of 

 boundary surfaces the greater the number of internal waves. On 

 the basis of this reasoning, when the density varies contin- 

 uously with depth one should expect an unlimited nuraber of 

 possible internal waves." The effect of this concept upon an 

 investigation of periodic variations in the oceans frequently 

 would be to give results different in nature from those given 

 by the two-layer system. With many infinitesimal vertical 

 changes in density occurring simultaneously, it does not 

 necessarily follow that the largest amplitude will be found at 

 the thermocline, A complete statistical analysis of this 

 problem has been presented by Selwell (1914-2). This consists of 

 a coiriplete harmonic analysis of the oscillations at several depths 

 at "Atlantis" anchor station 321-1-5. One of the main conclusions 

 of this work was that the smaller the vertical density gradients 

 at a given depth, the larger will be the amplitudes of the 

 oscillations, and the smaller will be the temperature changes 

 with time .at that depth. In the case of the particular data 

 that were analyzed in this study, it was found that the greatest 

 amplitudes occurred at a depth som.ewhat greater than that of 



57 



