1080 THE BELL SYSTEM TECHNICAL JOURNAL, SEPTEMBER 1957 



1000 



2000 



a: 



'^ 3000 



UJ 



5 



4000 



Q. 



a 



5000 



6000 



012345678 

 TEMPERATURE IN DEGREES CENTIGRADE 



Fig. 18 — Deep-sea temperature gradients. Mean gradient in the sediments is 

 indicated by the dashed line. 



category and most of these were either between Greenland and Labrador 

 or near the Equator. The number of actual bottom observations is 

 limited by inability to determine when the bottom was reached and a 

 reluctance on the part of observers to risk losing expensive equipment 

 by having it snagged on the bottom. 



At the present time there are insufficient data or knowledge of the | 

 mechanism involved to permit reliable extrapolation of bottom temper- i 

 atures from a series of mid-depth observations. Fig. 18 illustrates the- 

 problem. Here are six different near-bottom gradients observed in dif- 

 ferent parts of the Atlantic. Assuming these gradients terminated 500 . 

 meters above the bottom (as do many of the observed data), it is ap-j 

 parent that extrapolating such data to the bottom is not feasible. ]\Ieas- '< 

 urements of gradients to the bottom at stations for which near-bottom 

 data exist, coupled with a knowledge of the processes causing the gra- j 

 client, may make it possible in the future to make use of manj^ of the 

 old mid-depth observations in studying bottom temperature. 



From the compilation of available data, the three profiles (Figs. 20- 

 22) whose locations are shown in Fig. 19 were prepared. More recent 

 studies indicating that deep water temperatiu'es may vary a few tenths 

 of a degree Centigrade with time make it probable that some of the 

 ripples in the isotherms are not real but instead reflect the fact that the 

 data were taken at widely different times. The data for Profile 1 were 



