that this column may be the result of the inverse penetration. 

 Attempts to reconstruct such an occurrence, however, were uncon- 

 vincing and the question is left open. 



Instability of Transitional Profiles 



The inversions in the velocity profiles may be assumed to 

 dissipate rapidly, despite the fact that they were gravitationally 

 stable. The upper interface- of one intrusion is profiled in 

 Figure 8 and is seen to be a sharp boundary between colder 

 fresher water above and warmer more saline water below. In a 

 series of tank experiments, Turner and Stommel (1964) and Turner 

 (1965) investigated the mechanism and transfer rates across this 

 class of interface. When the gravitational stability provided 

 by the salinity was large compared to the inverse tendency to 

 overturn associated with the temperature gradient considered 

 alone, vertical transfer rates were low. Reducing the relative 

 stability by considering smaller salinity differences caused the 

 interface to undulate, with breaking waves providing an eddy 

 mechanism for an accelerated transfer of both heat and salt to 

 the upper layer. In the present observations the gravitational 

 stability parameter was low (density differences due to salinity 

 were only 507o greater than the opposing differences due to tem- 

 perature) and large effective diffusion coefficients should be 

 expected, at least in the early stages after penetration. 

 Furthermore, the possibility of "salt-fingering" at the lower 

 interface of the intrusion will tend to decrease the stability of 

 the upper interface and enhance transfer rates there. 



Summation and Possible Extension 



It has been shown that the Mid-Atlantic Ridge marks a change 

 in the sound velocity profile in the region studied. Although 

 this change is not overly significant it was seen that, in the 

 process of changing, unusual distortions occur in the region of 

 the sound channel axis. The distorted profiles were not found to 

 be permanently located, and it is supposed that their creation at 

 any given time will depend on the local water movement at a depth 

 of around 1,000 meters. The primary effect of the presence of the 

 ridge is thought to be that it sufficiently controls these water 

 movements so that a quasi-stationary transitional zone is estab- 

 lished at the ridge axis. 



Two types of transitional profiles were reported. One was 

 characterized by an inversion layer above the sound channel axis, 

 and it was shown how it may arise from the penetration of the 

 local manifestation of Mediterranean water into a fresher water 

 column. It was further suggested that the inversion layers would 

 rapidly be eroded away by vertical mixing. The other type of 



418 



