Role of Domtnant Wave tn Spectrum of Wind-Generated Waves 
more useful. 
I think that G, Hidy and I were the first (Hidy and Plate 
(1965)) to suggest a method for avoiding the problem of determining 
the proper functional form of the scale parameters. To normalize 
the spectral density we used the fact that the integral over the spec- 
trum had to equal the variance o* of the sea surface, and we used 
the frequency “to scale the frequency. In this manner, a similari- 
ty spectrum of the form : 
oui )=S(w) — (3) 
is obtained, where the non-dimensional spectrum Sg is a universal 
function of w ore We did not specify the functional form of equation 
3. But we pointed out that the high frequency end could be represented 
by the - 5 - power law. Many later writers have adopted the same 
procedure and represented their spectra in the form of equation 3. 
Examples are shown in figure 3, in which three different results for 
the similarity spectrum are given. The solid curve is a ''best fit"' 
equation through non-dimensional data obtained on Lake Michigan by 
Liu (1971). - This curve has a maximum at w/w_= 1 of 1.5. Super- 
imposed are the curves of Hidy and Plate (1965) and of Mitsuyasu 
(1969). The curve of Hidy and Plate was derived from laboratory da- 
ta. It has been corrected here for a scale factor of ten by which the 
vertical scale had been distorted in the original paper. Thus, the 
maximum of the peak is found to be at about 5 rather than 0.5 (as 
has been used, for example, by Plate and Nath (1969)). Ata first 
glance,one mayattribute the difference in the shapes of this and Liu's 
curve to the difference in the conditions at which the data were ta- 
ken and one may conclude that there exist different spectral forms 
for laboratory and field. Older analytical spectra for sea waves are 
found to have dimensionless peaks close to the one given by Liu. 
Mitsuyasu (1969) has shown that the maximum values obtained from 
the spectra of Pierson and Maskovitz (1965) and Neumann (1952 ) 
are equal to 1.43 and, 1. -15,, respectively. 
But there is evidence that there must be a different reason 
for the difference in the peak values of Hidy and Plate (1965) and Liu 
(1971). There are the spectra for laboratory and field waves, as pre- 
sented by Mitsuyasu (1969). For both conditions he finds almost iden- 
tical spectral shapes, with a maximum dimensionless density at 
w Jw in = | which is equal to 2.74, One may therefore suspect that 
the difference in the three results may be due to the data analysis 
1375 
