A. B. Wood 167 
As a preliminary to the experiments to be outlined, a test was made of the 
"constancy" or "repeatability" under prescribed conditions, extending over a 
period of 24 hours. In these tests a "point" (omnidirectional) transmitter emit- 
ting acontinuous wave was arranged at a fixed depth below the water surface while 
a point receiver was slowly raised from bottomto surface. Records of the pres - 
sure-amplitude variations, shown in Fig. 10.7, withmaxima and minima of some- 
what irregular spacing and amplitude, were made at various intervals of time. 
For short periods of three or four hours, the repetitions of the bottom-to-surface 
records were reasonably consistent, but for longer periods extending overnight 
to the following day considerable variations were noted, maxima and minima 
positions being often interchanged. At first, temperature fluctuations were sus- 
pected, but temperature control and stirring to prevent gradients of temperature 
made no appreciable difference. Eventually the trouble was traced to a small 
change in the depth of the water due to evaporation. This effect, which may be 
very critical, will be dealt with later. When a directional transmitter was used, 
however, the effect of variation with time was much less noticeable. 
10.2.1. Directional Properties and Depth of Transmitter 
A comparison of directional and omnidirectional transmitters when bottom - 
to-surface records are made at various depths and ranges of the transmitter is 
very striking. Whena transmitter of semiangle 10° is used and the bottom is steel, 
there may be only 3 or 4 maxima between bottom and surface, whereas the num- 
ber is increased to 10 or 12 if a point transmitter is used. The positions and 
amplitudes of the maxima vary as the depth or the range of the transmitter is 
changed, but the general characteristics of the records are the same, consider- 
ably more maxima being observed when thetransmitter is a point than when it is 
directional. As the range is varied, unexpected variations in location of the 
maxima are liable to occur, however, and it is practically impossible to asso- 
ciate a particular maximum atone range witha "corresponding" maximum at any 
other range. If any order exists between maxima at different ranges, there is 
no obvious means of connecting them one with another by this point-by-point 
method of recording. 
10.2.2, Range from Transmitter to Receiver Varied 
In this case the receiver and transmitter are fixed in depth while the range 
between them is varied. Here again the directional properties of the transmitter 
are very important. There is a gradual decrease of amplitude as the range is 
increased from, say, 1 to 4.5 m, but superposed on this are irregular fluctua- 
tions in amplitude. With directional transmission these fluctuations are much 
less noticeable than in the case where the transmitter is a point. In the latter 
case, when the bottom is steel, the fluctuations are very numerous indeed, but 
are much less serious when the steel bottom is covered with sound-absorbent 
material. In all cases the difference between narrow-beam and wide-beam trans - 
mission is very striking. 
10.2.3. Effects of Varying Nature of Bottom and Frequency of Sound 
It might appear that a scale-model tank having a "sea bed" of ‘-in. steel 
sheet would have little resemblance to a real sea bed consisting of mud, sand, 
gravel, or rock. When it is considered, however, that the sea bed in shallow 
