than the results on which the original material was based. 



The spectrum computed from the wave pole data does not agree with 

 the spectrum computed from the stereo data nor with the corresponding 

 theoretical Neumann spectrum.. Variability in the spectrum due to wind 

 variation of the order of one knot would explain the discrepancy. However, 

 a more likely reason for the discrepancy appears to be in the calibration of 

 the wave pole. 



The numerical results which have been obtained provide valuable data 

 on a wind generated sea for a fairly low wind speed. It will be particularly 

 useful in studying the topography of the sea surface and in problems con- 

 nected with seaplanes and small vessels. 



Recommendations 



The use of stereo photographs to determine the directional spectrum 

 of a sea has proved feasible. Due to attrition, an originally desired 50 

 degrees of freedom per spectral estimate was reduced to only 19. The com- 

 putations were lengthy and difficult, but nevertheless results of consider- 

 able value were obtained. 



It is difficult to generalize the results obtained to higher wind speeds, 

 and one determination of a directional spectrum is not enough to provide 

 comprehensive details on fully generated seas for a range of wind speeds. 



It is therefore recommended that an experiment similar to the one de- 

 scribed in this report be repeated for a fully developed sea at at least one 

 higher wind speed. A wind of 24 knots, a fetch of 130 NM and a duration of 



254 



