periods shorter than 2 At should be negligible in the record. 
Wave records obtained by pressure recorders are an excellent 
illustration of this point. If, for example, the depth of the 
water is 22.5 feet, then waves with wave lengths less than 45 
feet have very little effect upon the pressure at the bottom. 
Consequently components with periods less than three seconds 
will have very little effect on the pressure at the bottom and 
will not show up in either the record or the power spectrum. In 
this example, then, a At of 1.5 seconds would be sufficiently 
small to peste that there was no aliasing in the analysis of 
the pressure record. This point will be discussed in greater de- 
tail later. 
Resolution and the choice of m 
‘For a fixed value of At, the larger the value of m, the 
more points are determined for the power spectrum, and the greater 
the ability of the analysis to determine the finer details of the 
power spectrum. For example, if it was suspected that a wave re= 
cord contained a sharp peak near a certain frequency, say 21/5, 
then the sharp peak could be determined more accurately by taking 
a larger value of m. For A t equal to 1 and for m equal to 10, 
then frecuencies from 2173/20 to 275/20 are present in the band 
containing 27/5. For At equal to 1, and for m equal to 50, 
then frequencies from 2719/100 to 2721/100 would be present in 
the band containing 27r/5.* If the power were actually concentrated 
*Periods from 6.67 to 4 seconds in the first case, and periods 
from 5.26 to 4.76 seconds in the second case. 
- 280 - 
