turbulent motions can be computed with an acceptable significance 
level. 
In order to elucidate why it is crucial to obtain quantitative informa- 
tion on the transport of momentum and vorticity by large-scale eddy 
processes in oceans, we call attention to an analogous position in de- 
velopment of the theory of the atmosphere’s general circulation about 
10 years ago, when the fundamental quantitative studies of statistics 
of upper-air data carried out by Starr and his collaborators over- 
threw the classical picture of a predominantly meridional circulation, 
in which it was thought that the observed large-scale fluctuations 
played a more or less passive, dissipative role. Starr’s investigations 
of observations demonstrated that fluctuations actually drive the 
mean circulation, and present-day theoretical studies of atmospheric 
circulation allow the fluctuations to play this more important role. 
We are in a similar position in oceanography. The fundamental 
concept, about which all theoretical investigations from 1947 to 1962 
are pivoted, is the basic Sverdrup relation between local cur] of the 
mean wind stress and vertical integral of the meridional velocity com- 
ponent. The theory of the thermocline, thermohaline, and wind- 
driven circulation all depend upon this simple idea: That large-scale, 
quasi-geostrophic eddy processes do not play an important dynamical 
role in vorticity balance in the interior of the oceans. 
During the past few years serious doubts about the neglecting of 
eddy processes have begun to arise: 
(a) Aries measurements in the Atlantic, originally planned by 
Swallow, Crease, and Stommel to determine the mean velocity field 
at different depths, unexpectedly revealed the presence of large-scale, 
long-period eddies whose root mean square amplitudes were two 
orders of magnitude greater than the expected means, indeed so large 
that it is difficult to imagine that they can be decoupled from the mean 
fields as is implicit in the Sverdrup relation. At any rate irregular 
motions were so large that it was not possible to test the Sverdrup 
relationship in the simple way which the Avies measurements were 
originally intended to do. In order to obtain a statistical description 
of these eddy processes and to be able to map and describe them, it is 
evident that an effort at least an order of magnitude greater than the 
Aries measurements is necessary. 
(6) Calculation of the amplitude of the abyssal circulation from 
IGY and Norpac data—by the method of Stommel (1956)—yields 
abyssal circulation rates much too large to be compatible with water- 
mass analysis and radiocarbon data. The same lack of agreement ap- 
pears when the thermocline theory is semiquantitatively applied to 
actual density distribution in the ocean. These discrepancies also sug- 
gest that something important is omitted from the simple Sverdrup 
relation. 
115 
