Microscopte Structures of Wind Waves 
contribution to mean-square slope from wave components having 
their wavenumbers greater thank, are about the same for all wind 
velocities in the smooth regime. This trend is along the lines of 
Phillips concept (Phillips 1958a) of the development of the wave 
spectrum, in which the higher-frequency waves reach the saturated 
state earlier. The shift toward the left of the clean-surface data or 
the shift toward the right of the slick data is also very interesting. 
This clearly indicates that the wind boundary-layer transition from 
smooth to rough regime is delayed by the presence of the slick. It 
is expected that the ocean surface becomes smoother when it is cove- 
red by a dense oil slick. 
The difference of mean-square slope between the clean and 
the contaminated sea surfaces at low wind velocity is about 0.0115. 
This difference is the contribution to the mean-square surface slope 
from wave components having their wavenumbers greater than k_. 
Accepting this argument we can estimate from the slope difference 
the cutoff wavenumber k, at low wind velocities provided that Kewis 
smaller than k, . From (13) and (14) we have 
0. 0046 Ln (k_ /k_) = 0.0115 k = 2.5¢m_ (15) 
This value is indeed smaller than k, which is about 3.6 Sen “whe 
closeness of these two values indicates that the straightline fitted 
through the clean surface data at low wind velocities may be the upper 
bound of the contribution from waves in the gravity range. In other 
words, the contribution to the mean-square slope at higher wind velo- 
cities shown in figure 17 above the extension of this straight line must 
come from waves in the capillary range. At first look, this conside- 
ration may seem rather arbitrary. Actually, since B is very small, 
the choice of a slightly different cutoff wavenumber k, has an insi- 
gnificant effect on the results. On the other hand, the coefficient B' 
to be shown later, is much greater than B, so that wave components 
in the capillary range contribute much more effectively to the mean- 
square sea surface slope than those in the gravity range. Consequent- 
ly, once the integration of the wave spectrum extends into the capil- 
lary range, a change in the trend of the data, such as that shown in 
figures 16 and 17, is expected. More studies are needed to see 
whether it is just a coincidence that this separation of slope behavior 
coincides with the change of the regimes of the wind boundary layer. 
An excellent correlation of data between s* and U, g 
is shown in the hydrodynamically rough regime of wind. This 1928 
indicates two possibilities : the contribution to mean-square slope 
1453 
