to 600 and m equal to 60 (assuming all the necessary trigonometric 
terms are available in a matrix), it should take about one eight 
hour day with a skilled operator. For punch cards on an IBill 
machine, about four hours are needed. High speed calculators re- 
quire two and a half minutes, plus the time required to code the 
machine. It might be well worth while to set up a high speed cal- 
culator for the permanent part of the process and process a great 
many observations in one single day. 
The determination of [a(n ,e)]° 
i se 
The function, [A,(# ,6)]°, is very difficult to evaluate. 
It is a function of two variables, and, over a complex sea surface, 
especially at the edge of a storm, the short crested waves imply 
a wide variation of the function. The needed measurements can be 
made and it is theoretically possible to determine the structure 
of the function. 
There are two instruments discussed in the literature which, 
along with an ordinary pressure (or spar) type wave recorder, make 
it possible to measure the appropriate functions. One instrument 
is the airborne wave recorder described by Deacon, Darbyshire, 
and Smith [1949]. It measures elevation of the sea surface along 
some chosen line over the sea surface at practically an instant 
of time. It will be assumed in this derivation that the speed of 
the aircraft is so great compared to the variation with time of 
the sea surface that the values obtained are essentially instant- 
aneous. Minor (in principle) modifications of the procedures which 
will follow could eliminate this assumption. The second instrument 
would have to be an extension of the stereoptican methods to high 
= 287° = 
