Tsolated sets of short-duration observations have been made (O. V. 
Schubert, Ufford, Cox, Fofonoff). Some show evidence of internal 
tides, directional propagation, several vertical modes and inertial 
motions. Preliminary discussions of any of these results immediately 
lead to questions which can be resolved only by significantly more 
complicated deploying of buoys and sensors. As an example one can 
cite measurements of water temperature at Bermuda (1958: Haurwitz, 
Stommel, and Munk) in which long-duration measurements were ob- 
tained at the cost of mounting only two thermistors on the sloping 
bottom. More thermistors at various depths in the open oceans would 
have clearly been better but were beyond the technical resources and 
money available. It is clear that multiplication of sensors is the 
direction in which we must go; this means that oceanography must de- 
liberately attempt to establish an instrumented portion of the deep 
sea capable of obtaining refined measurements of internal gravity 
waves. 
Each set of measurements would be continued for long enough dura- 
tion to contain roughly 100 of the longest period waves of interest. 
According to Eckart’s analysis of internal gravity waves, we could 
anticipate that periods between 10 minutes and 1 day would be of 
primary interest, so that the duration of an individual set of measure- 
ments should be 100 days. This is a long time to keep a complicated 
array of many recording sensors operating at sea and would also lead 
to a formidable data-processing program. The array might consist 
of from 3 to 10 moored buoys with perhaps 10 to 20 vertically spaced 
sensors on each. The spacing horizontally between buoys would be 
varied: a working estimate to begin with could be 1 kilometer, the 
whole array being within a 4-minute square. 
Measurements of horizontal velocity components recently made at 
Woods Hole Oceanographic Institution actually do not seem to indi- 
cate coherent wave motion. This may be an inherent difference in 
the nature of velocity and temperature structure in the ocean. Above 
tidal frequencies horizontal velocity components appear to consist of 
horizontally isotropic, incoherent eddies resembling turbulence. Hori- 
zontal scales range from 3 to 5 kilometers at the lower frequencies (0.1 
cycle per hour) to perhaps 1 meter at the high-frequency limit of 
resolution of the current meters. Energy density decreases as the 
“minus 5/3” power of frequency. The motion is clearly not isotropic 
vertically and, because of vertical stratification, may have little vertical 
coherence. Vertical scales associated with this motion have not been 
determined. 
As these turbulentlike fluctuations may be capable of developing a 
high vertical shear and perhaps generating shear turbulence, they may 
bey a vital role in vertical mixing and transfer processes. Associated 
equency distribution and spatial scales for temperature and salinity 
220-659 O—66——_9 113 
