Part 1. 

 INTRODUCTION 



On November 25, 1954, two aircraft rendezvoused with the R. V. ATLANTIS 

 at a point in the North Atlantic. The aircraft made a sequence of passes over 

 the ATLANTIS and flew back to base. Months of preparation went into the 

 flight, months of thought went into the problem of what to do with the data ob- 

 tained on the flight and by the ATLANTIS, and mionths of work went into the 

 numerical processing of the data. 



The results of the flight were stereo pairs of photographs of the sea sur- 

 face. The ATLANTIS provided base line calibration and wave pole and visual 

 observations. Two of the best pairs of photos were reduced to 5400 numbers 

 on a rectangular grid. The wave pole records were reduced to a time series 

 of discrete points of about 1,800 numbers each. The purpose was to take the 

 two sets of 5400 numbers, estimate the directional spectrum of the waves on 

 the sea surface and compare it with the frequency spectrunn as estimated for 

 the three sets of 1800 points read from the wave pole records as a check. 



To go from the 10,800 stereo nunmbers and the 5,400 wave pole nuinber s 

 to the desired spectra required a total of about 9,000,000 multiplications and 

 an equal number of additions. After the directional spectra were computed, 

 the results obtained were inconsistent with the theoretical models, and the 

 St ereo data had to be carefully re-analyzed witli the result that part of the 

 data had to be discarded. The 5400 numbers were reduced to about 3500 

 numbers, and the computations were done over again. 



