structure are small, the magnitude of the horizontal gradient is 

 nearly zero. The smoothed temperature structure of Sample 

 Area 9 exhibits some localized areas essentially free from 

 vertical fluctuations of the isothermal surfaces. Examination of 

 the vertical- and horizontal-gradient fields for Sample Area 9 

 shows that, when the vertical-temperature gradient is about 

 2 x 10' 1 °C/ft, the corresponding horizontal-temperature gradient 

 is of the order 10" 6 °C/ft and decreases to zero. Therefore in 

 regions of strong vertical-temperature gradient (high vertical 

 stability), vertical fluctuations in the temperature structure are 

 small and time -dependent horizontal-temperature gradients 

 nearly disappear. 



The regularity with which the horizontal-temperature 

 gradient differs from the vertical-temperature gradient suggests 

 that the slope of the isothermal surfaces is the regulating factor. 

 The vertical- and horizontal-gradient interrelationship may be 

 written as the identity 



8T/Sx = (ST/Sz) (dz/dx) (2) 



where ST ox is the horizontal-temperature gradient in °C/ft, 

 8T 8z is the vertical-temperature gradient in °C/ft, and dz. dx 

 is the slope of the isothermal surface. The ratio of ST, 8x to 

 8T/Sz taken over the samples shows that the mean value of dz dx 

 is 1.12 x 10 ' 2 . 



In an independent investigation (also using thermistor chain 

 data) of the thermal structure around the tip of Baja California 

 (LaFond and LaFond, 1966), 18 two isotherms were selected for 

 analysis, one in the main thermocline and one below it. The 

 location of the data sections is shown in figure 7. The depth 

 differences from point to point along the isotherms were deter- 

 mined from the equation 



Yk = *i ~ X, + l (3) 



where 1 < k < N ; X, and X- +1 are depths in feet of a given isotherm 

 at the beginning and end of the z' th distance interval along the 

 track; and Y k is the computed depth difference in feet for the /c th 

 interval. The ship's speed was 6 knots and the sampling interval 

 was 30 seconds, corresponding to a distance interval of 304 feet. 



20 



