{low frequencies) and spread, out over a wider range for short waves (high 

 frequencies) and (3) that the integrated spectrum, continues as predicted into 

 higher frequencies as determined by the wave pole data, The properties 

 should be expected to be the sa.me for other spectra obtained for other con- 

 ditions at other times. 



There are other properties of the particular spectrum studied which are 

 in part probably due to sampling variation and in part due to the particular 

 local wind .field which generated the waves. The values of 0, show that the 

 peak in the angular variation of the spectrum, shifts from what corresponds 

 to 180° in figure 11.29 to 140° as the frequency increases from 2it(11)/96 

 to 2th'2?)/96, Also a secondary peak a.t frequencies corresponding to 

 2Tr{16)/96. 2TT{i7)/96» 2'n-(18)/96.. and perhaps even for higher frequencies, 

 is indicated in figure 11,22, a.nd by the high values of c^ and the values of 

 ^2 ^^' figure 11 30, This secondary peak causes the graphs in figure 11.22 

 to have the property that they are not even functions about sonae central 

 value of the direction The change in Oi can be explained partly by sampl- 

 ing variation and partly by the fact that the local wind direction was reported 

 to be from 330° and the winds further to the north were from 360 v Pos- 

 sibly the winds to the north were the ones which generated the longer waves. 

 The skewness of the curves for the angular variation may or may not be real 

 in the sense that it would still show up in a spectrum with a larger number of 

 degrees of freedom However,, it should eilso be noted that there is a wind 

 shear present over the area of w9-ve generation with tlie property that the 



