shapes are similar the magnitude of the sound velocity shows differences 

 up to 2 "^/sec at various depths in the thermocline. Differences of this 

 order of magnitude are not surprising. A/Piip (personal communication) 

 of the Bermuda Sofar Station has found a semidiurnal variation in the 

 sound velocity at a depth of 1150 m. of a little over 1 ™/sec in this 

 area. Figure 3 shows Profiles 34 and 47 which were made 4 days apart 



about half way between Bermuda and Puerto Rico. The differences between 

 them in the magnitude of the sound velocity are small. However, one 

 significant difference occurs at 300 m. In Profile 34 the sound velocity 

 gradient becomes positive for a short distance below 300 m. causing a 

 secondary sound channel while in Profile 47 the gradient is negative 

 throughout the thermocline and there is no secondary sound channel. We 

 shall return to this point later. Figure 4 shows profiles 5 and 49 made 

 34 days apart. Differences in magnitude of the sound velocity of up to 4 

 to 5 ^^/sec are found in the main thermocline with somewhat larger 

 fluctuations near the surface. Again the general shape has not changed 

 and both profiles have a decidedly positive sound gradient near 300 m. 



LONG TERM VARIATIONS 



Comparison of the data taken in 1962 with that acquired in 1964 gives 

 us an idea of some changes which have taken place during this two year 

 period. The 196 2 data were spread over a greater area but the 1964 

 profiles were closer together. From his data Reitzel drew isovelocity 

 contour curves from 200, 300, 400, 500, 600, 800, and 1000 m. Contours 

 of course, are an interpretation of results, but the 1962 data have been 

 contoured independently by three people all skeptical of the others' 

 interpretations. The contours of all three are essentially the same. The 

 1964 profiles do not cover a broad enough area to be contoured but we can 

 determine whether they agree even generally with the 1962 data. Figure 5 

 shows the 1962 contours and the 1964 sound velocities at a depth of 800 m. 

 Contour plots at all depths between 300 m. and 1000 m. are qualitatively 

 similar to that at 800 m. except as noted below. Sound velocities have 

 been rounded off to the nearest meter per second and are close to 1500 

 "^/sec. 1400 or 1500 must be added as appropriate to give the true sound 

 velocity. The minimum and maximum sound velocities at a given depth 

 may differ slightly from the two years but the range is strikingly similar 

 at all depths. The next most general feature is a regional decrease in 



ouuna velocity from NW. to SE. near the surface which gradually rotates 

 around to N. to S. as one progresses down to 1000 m. If we chose to call 

 this the zero order mode of change then the first order change manifests 

 itself as a maximum of the sound velocity in the SW. and a minimum in 

 the NE. in the 196 2 contours. The numbers from 1964 show a reversal 



438 



