Consequently, the Rhythmic Grunt sound is a prime suspect 
as the main cause of the elevated base ambient noise 
spectrum at 2300 hours in the deep water case. 
In addition to being lower in frequency, on the 
average, the Rhythmic Grunts at the deep hydrophone were 
characterized by a slower repetition rate. Checks on 
widely scattered samples showed the following means of 
the Grunt repetition rate: at the shallow hydrophone, one 
per 4.39 sec +0.056 (SE), one per 4.55 sec +0.055 (SE), and 
one per 4.12 sec +0.054 (SE); at the deep hydrophone, one 
per 5.04 sec +0.053 (SE) and one per 5.68 sec +0.13 (SE). * 
An additional sample checked for timing was part of a 15- 
minute towed-array sample recorded on SALUDA (IX 87) 
sailing at 5 knots in 6-fathom water off Black Warrior 
Lagoon, B. C. This sample, which was recorded on 
9 February 1965, seemed almost identical to some of the 
San Clemente shallow samples and also revealed a very 
similar repetition rate: one per 4.28 sec +0.060 (SE). 
The most mysterious of all the sounds, of course, 
were the 20 c/s Long Pulses. They were present in every 
sample. Although the mean maximum S/N per sample was 
18.7 dB deep and 23.4 dB shallow, many other Long Pulses 
were just audible. One explanation for the generally 
greater S/N in the shallow-hydrophone channel for 20 c/s 
Pulses both Long and Short is the rapid slope off of the 
deep hydrophone response below 40 c/s, which for the 
analysis band centered at 20 c/s could emphasize the noise 
background in the top half of the band more than the 20 c/s 
Pulses. 
* SE = standard error of the mean 
