as a guide in the absence of better evidence. The consistency of 

 the data definitely indicates a clockwise rotary current ellipse with 

 the major axis oriented approximately NW-SE. The observations were 

 made during a period of comparatively isothermal conditions so that 

 maximum correlation of currents to tidal motion might be expected. 



C. ANALYTICAL COMPARISONS OF SIGNATURE SPEED 



The recorded internal fluctuations of temperature throughout this 

 field study characteristically appear in groups of one to four temperature 

 "humps" at irregular intervals. This strongly suggests the application 

 of solitary internal wave theory after Long (1956) and Keulegan (1953) 

 in order to compare observed speeds; both of these theoretical 

 developments assume a two-layer system. Long further imposes a 

 rigid boundary at the water surface as well as at the bottom. Keulegan' s 

 work was therefore selected as being most appropriate. His equations 

 are summarized below and then applied to several observed cases 

 of temperature- salinity structures for comparison with measured 

 signature speeds. 



Consider a two-layer fluid medium of thickness hj and density 

 P, between the free surface and the internal density discontinuity 

 overlying a layer of thickness h? and density h^ between the discontinu- 

 ity and a rigid bottom. It is shown that a solitary wave of infinitesimal 

 height occurring at the internal discontinuity will travel at a phase 

 speed, c , given 



gy 2 



V h 2 



Pir 1 ) 



1/2 



(2) 



where g is acceleration due to gravity. As a matter of interest this 

 is the same equation as derived by Defant (1950) and Haurwitz (1950) 

 for the oscillatory internal longwave neglecting the effect of earth 

 rotation. Ekman (1904) also applied this equation to explain the "dead 

 water" effect on vessels moving slowly in a shallow, low-density 

 surface layer. 



The phase speed, c, of the internal wave of finite amplitude tf 

 is found to be 



h x -h 2 



1 + 71 



h l h 2 



1/2 



(3) 



