a downward perturbation, lateral diffusion of heat from the descending 

 column results in an increase in the negative buoyancy of the finger. 

 Therefore it is able to overcome the stabilizing influence of the temperature 

 gradient and continue its downward motion. Whether the process is self- 

 limiting (i. e., whether convection cells thus generated are limited to a 

 vertical length scale) is not known. However, a self- limiting situation 

 must operate if layering is to occur due to the unstable density flux mixing 

 both salt and heat downward. 



Turner hypothesized that in the ocean some relation (as yet unknown) 

 between the temperature gradient and the salt flux specifies the criteria for 

 the formation of a new layer. There is observational evidence supporting 

 his hypothesis. Tait and Howe (1968) measured very distinct layers under 

 Case II conditions in a region of the Eastern Atlantic where warm, salty 

 Mediterranean water overlies colder less salty water. Their data show 

 extremely sharp interfaces in both salinity and temperature. The column 

 is marginally stable at the outset and is maintained so by continuous hori- 

 zontal advection of Mediterranean water which is the source of energy to 

 resupply the energy lost by vertical molecular and convective diffusion 

 processes. 



Layering in Case III 



Turner also experimented with Case III conditions. Using a dimensional 

 argument, he showed that the ratio of turbulent transfer coefficients for salt 

 and heat, Kg/KT, is some function of the density difference (stability) ratio, 

 P AS/a AT, where (3 is the coefficient for density change due to salt,and a is 

 the coefficient of volume expansion doe to temperature. For stability ratios 

 between 2 and 7 the ratio of turbulent transfer coefficients decreases linearly. 

 For the same stability ratio range the potential energy ratio is constant, with- 

 in measurement error, at a mean value of about 0. 15. 



Layering in the Arctic 



In March 1969, we obtained a unique set of temperature profiles from 

 the Arctic ice island, T-3. T-3 was then at 84<'34.1'N 128° 17'W where the 

 water is about 2650 meters deep, a favorable location for the study of thermal 

 microstructure. Previous reports of temperatures measured in that portion 

 of the Arctic gave mean temperature drops of 0. 01° to 0. 02° C over a depth 

 interval of about 200 meters, centered at the top of the relatively warm 

 Atlantic water (Case III). 



460 



