of 600 points. Where gaps or irregularities occurred, the records 

 were smoothed by hand as accurately as possible. 



The time series of 6OO points was then analyzed on the 

 CDC 1604 so as to estimate the energy spectrum of the waves at 60 

 points over the frequency range from zero to 0.333 cycles per second 

 by means of the procedures given by Tukey (1949) as explained in 

 detail by Blackman and Tukey (1958). The smoothing operation that 

 was used to go from L to U in the equations of Blackman and Tukey 

 was 



(1) U, = 0.25L, , + 0.50L, + 0.25L, ^, 



h h- 1 h h+1 



with suitable corrections at the ends of the range. 



The spectral estimates so obtained still had to be cor - 

 rected for the response of the shipborne wave recorder (Tucker, 

 1956) and for the introduction of noise in both the original record 

 and in the digitization procedure. The calibration of the shipborne 

 recorder depends on the ship, and the calibration curves were pro- 

 vided by Mr. D. E. Cartwright for this purpose. The calibration curve 

 for the Weather Exp lore r is given by Table 1. The calibration curve 

 for the W eather R eporter is given by Table 2. 



As in another investigation (Bretschneider, Crutcher, et 

 al (in press)), it was found that the application of the above cali- 

 bration curves to the spectra that were originally computed resulted 

 first in a decrease and then a rapid increase in the spectra at high 

 frequencies due to the presence of noise and other irregularities 

 (possibly from nonlinear effects in the original wave records) at 

 the high frequency end of the spectrum. To eliminate this effect. 



