The neutrally-buoyant float measures deep 

 currents. Each free-floating tube is 

 weighted so that it will stabilize itself 

 at a chosen depth . As it is carried along 

 by a current, a battery-powered transmitter 

 sends out sound pulses. By observing 

 these "pings" with ship-borne hydrophones, 

 the movement of deep water can be traced. 



than half a mile, where it finds its own density level and spreads 

 out across the North Atlantic, easily recognized by its relatively 

 high temperature and high salinity. 



Although water sampling gives us a clear picture of the spreading 

 of different water masses over the oceans, it is more difficult to 

 assign speeds to these movements. Calculations from the density 

 distributions which, in a few cases, can be supported by direct 

 measurements, indicate speeds of about a mile a day in some of the 

 deep-water masses ; but these speeds are variable and uncertain, and 

 the average movements of these deep-water currents may be much 

 slower. 



One way in which the average rate of movement of deep water 

 masses can be found is by measuring radiocarbon concentration. 

 When the surface waters sink in high latitudes to form new deep 

 water, they carry down traces of radiocarbon from the atmos- 

 phere. Thereafter this radioactive isotope of carbon decays at a 

 known rate. Despite the difficulties in applying this method, 

 more and more measurements suggest that the deep waters spend, 

 on an average, several hundreds of years below the surface. Assum- 

 ing that, in this time, a typical particle of North Atlantic Deep water 

 travels southward through the whole length of the Atlantic, its 

 average speed must be about one mile per month — much less 

 than the currents observed directly over a period of a few days or 

 weeks. 



How certain can we be of any particular pattern of move- 

 ment we may infer for the deep waters? The real pattern may, in 

 fact, be quite different from the simplest pattern suggested by our 

 readings of temperatures, salinities, and other properties. Recent 

 theoretical work on the dynamics of the deep-water circulation 

 suggests that the flow may be more complicated. If we start with 

 the two main sources of deep water (the extreme North Atlantic 

 and the Weddell Sea, where the water is sinking) and suppose 

 that everywhere else the deep water is very slowly moving up- 

 ward, it can be shown that the horizontal movements of the deep 

 water should be everywhere directed slowly toward the poles - 

 except at the western boundaries of the oceans where strong deep 

 currents should be found. This is, of course, quite a different pic- 

 ture from the slow southward drift of the North Atlantic Deep 

 water that can be inferred from observations of its temperature 

 and salinity. There is some evidence for the existence of the pre- 

 dicted southward deep current along the western side of the Atlan- 

 tic, but other details of this theoretical model still await confirmation. 



In two recent discoveries, though, the observations were ahead 

 of the theoretical predictions. Only a few years ago a remarkable 

 undercurrent was found flowing eastward along the Equator under 

 the westward-flowing South Equatorial Current in the Pacific. At 

 a depth of only two or three hundred feet below the surface, this 

 undercurrent flows at a speed of two or three knots for at least 

 3 5 oo miles along the Equator. 



Even more recently a similar undercurrent has been found 

 flowing under the South Equatorial Current in the central Atlantic. 

 To date we do not know what kind of current flows under the 

 equatorial current in the Indian Ocean. With the changing surface 

 currents brought on by the monsoons, the pattern may be quite 

 different there. 



213 



