A grid of 30 feet was therefore chosen„ Smaller values of Ax w^ould, requsire 

 a much greater m. than that actually chosen and majiy more spot heights o 



The winds at the surface varied from 17 to 20 knots just prior to the 

 time of the observations and hence the values seemed consistento Spectral 

 periods a,s Mgh as 11 seconds mj.ght have been present in the waves due to 

 the 20 knot winds. This period, would correspond to a wavelength of about 

 600 feet. 



With a grid spacing of 30 feet, the area of the stereo ajialysis for one 

 pair of photographs was found to contain about 60 points on the short side 

 and more than 90 points on the long side. This would imply the determinationi. 

 of 5400 spot heights from each stereo pair„ 



Various lags were then tested and a value of m equaJ. to 20 wa.s chosen 

 for two reasons. The first was that there would be adequate resolution,, 

 and the second was that there would be enough statistical reliability. 



With respect to resolution, wavelengths greater than 2400 feet woxxld 

 then show up at the origin and since this corresponds to a period of over 20 

 seconds the energy at zero wave number should be entirely due to aliasing 

 and white noise reading error on the assumption that the Neumann spectrum 

 was roughly correct. The next wave number would cover a range in J.eni.gths( 

 from 2400 feet to 800 feet, and it would also not be expected to show any 

 appreciable wave energy. 



These values were also checked on the assumption that the peak of the 



spectrum would fall at an angle of 45 degrees to the coordinates of the spec^ 



55 



