shown in figure 10.2. 



Figure 10.2 shows that no single theoretical curve for a parti- 

 cular wind velocity lies completely within the bounds of the 90 percent 

 confidence bands. In general the values for the observed spectrum 

 are too high for |jl = 2'n-(ll)/96 and 27r(12)/96 and too low near 

 (X = 27r{15)/96. 



However, it is also evident that at least one of the five points 

 plotted for the five different theoretical spectra falls within the 90 

 percent confidence bands on the observed spectrum for all values 

 of k between 10 and 30. A variation in wind velocity of J; 5 per- 

 cent about a value of 18o25 knots is more than sufficient to explain 

 the observed spectrum at each of these points. 



At values above k = 30, the observed spectrum is a little 

 above the theoretical spectrum. This may in part be due to a 

 small amount of white noise. 



An appeal to the meteorological turbulent variation of the 

 wind speed and to the theory of wave generation and propagation 

 must be made in order to clarify this point. The observations 

 of the ATLA.NTIS as plotted in 



138 



