VERTICAL AND HORIZONTAL MOVEMENTS IN THE OCEAN 613 



from others a name from Celtic folklore, "the Merlin Complex," 

 is appropriate. 



Between 5° and 8° the theta/S diagram (Fig. 2), confirming 

 work in earlier years, suggests another water mass in neutral 

 equihbrium (sigma-theta = 27.800), but there may be a slight 

 change in slope of the theta/02 curve at about 6°. 



Discussion 



The usual concept of the deeper water in the eastern North 

 Atlantic is that, in the water below 2000 m and the bottom, there 

 is a small but steady increase in density downwards. This investi- 

 gation suggests that this view should be reexamined. 



It seems that we have a multiple sandwich structure, consisting 

 of a series of layers, each in completely neutral adiabatic equi- 

 librium. For these layers it is possible, by the method of least 

 squares, to calculate straight line relationships between potential 

 temperature and salinity and between potential temperature and 

 oxygen. The changes in temperature and in salinity compensate 

 each other precisely to give potential temperature /salinity curves 

 parallel with the isopycnals. Working up from the bottom we may 

 have adiabatic water masses with potential densities 27.903, 

 27.892, 27.862, 27.831, 27.800, and conceivably two others at about 

 27.818 and 27.803 (Fig. 8). If the broad principle presented here 

 should be confirmed, it is not anticipated that these figures will be 

 precisely reproduced. Particularly, may this be so for data obtained 

 with a salinometer. 



The curves for neighboring water masses in adiabatic equilibrium 

 do not intersect. Consequently, they may never have been in direct 

 contact, and their areas of formation may need to be sought in quite 

 different places. 



These labile well-ordered strata seem to be separated by what 

 were first called zones of conflict. In these zones the density in- 

 creases with depth, and the oxygen contents are chaotic but usually 

 high. It has not been possible to bring any kind of order into the 

 results. No curve can express them accurately. Our close sampling 

 has not been close enough to reveal any fine structure that the 

 zones of conflict may have. It may not be chance that large dis- 



